Method and apparatus for range and coverage extension in a heterogeneous digital chaos cooperative network

10069522 ยท 2018-09-04

    Inventors

    Cpc classification

    International classification

    Abstract

    The present invention teaches a system and method for improved signal recovery for range and coverage extension in a heterogeneous cooperative network of digital chaos transmissions with OFDM component signal transmission. The invention improves upon the state of art in side channel information from the transmit side containing information on the clipped amplitude. In-band transmission of the side information is achieved by exploiting the sparsity of the resulting clip amplitude position with improved levels of compression over the prior art using Gabor Transform Multiple Symbol Encoding transmitter. The information rate of the clipped amplitude is sub-Nyquist relative to the original OFDM component signal transmission, which allows very low power spreading by a cooperative digital chaos sequences at a transmit side and recovery of the clipped amplitude at a receive side. Further, an improved noise resistance side channel performance is achieved by decoding Gabor Transform symbols for symbol recovery.

    Claims

    1. A method of compensating for in-band distortion using signal recovery of clipping amplitudes of a primary user in a cooperative heterogeneous network of secondary users, wherein the secondary users comprise digital chaos signals, wherein said primary and secondary users are wirelessly transmitted, the method comprising: measuring the instantaneous amplitude envelope of a primary user, the instantaneous amplitude envelope including multiple measured instantaneous amplitudes, wherein the primary user contains at least one group of Orthogonal Frequency Division Modulation (OFDM) signal components, wherein said instantaneous amplitude envelope is measured at a rate of at least four times the base rate of the OFDM signal components; clipping at least one of the multiple measured instantaneous amplitudes when the at least one of the multiple measured instantaneous amplitudes exceeds a predetermined threshold, said clipping to reduce said instantaneous amplitude envelope to a value no more than the predetermined threshold, wherein the predetermined threshold is derived from the maximum allowable power according to a peak-to-average ratio (PAPR) constraint specified at a transmit side; collecting multiple clipped instantaneous amplitudes values into a sparse vector, wherein the sparse vector has a nonzero value and a corresponding nonzero value position location, wherein said nonzero value and said nonzero value position location are aligned with instants of clipping for a frame data as defined by the international protocol standard for the primary user; processing the sparse vector into a smaller vector, the smaller vector containing a vector of non zero values, the processing of the sparse vector comprising Gabor Transform clipped amplitudes and position location to produce Gabor transformed compressed samples; transmitting the nonzero values as the payload of an extremely low powered digital chaos secondary user, wherein the payload of the digital chaos secondary user is transmitted in-band with a primary user payload, wherein the amplitude of the digital chaos secondary user is less than the amplitude primary user according to the error vector magnitude (EVM) requirements of the regulatory requirements; receiving said digital chaos secondary user payload and said primary user payload at a receive side; processing said digital chaos secondary user payload in the presence said primary user payload to produce the nonzero value and the nonzero value position; using said nonzero value and the nonzero value position to produce an estimate of the sparse vector; and using the estimated sparse vector to compensate for distortion created by clipping the at least one of the multiple measured instantaneous amplitude.

    2. A method of claim 1, wherein the step of clipping at least one of the multiple measured instantaneous amplitudes comprises polar clipping, and wherein the amplitude of instantaneous amplitude envelope is reduced to the predetermined threshold, but the phase of the instantaneous amplitude envelope is preserved.

    3. A method of claim 1, wherein the step of clipping at least one of the multiple measured instantaneous amplitudes comprises rectangular clipping, wherein the amplitude of the sample is reduced to the threshold, but the phase of the sample is not preserved.

    4. A method of claim 1, further including modulating said sparse vector with a digital chaos sequence as part of a secondary user transmission.

    5. A method of claim 1, wherein the digital chaos secondary user includes a power level, wherein the digital chaos secondary user power level is determined from an error vector magnitude (EVM) caused by the sparse vector compared to that which is allowable by the regulatory standard protocol to which the primary user signal is a compliant.

    6. A method of claim 1, further comprising decoding and demodulating of Gabor Transformed compressed samples to improve detection of the clipped amplitude.

    7. A method of claim 6, wherein decoding Gabor Transformed compressed samples comprises using a dynamic programming algorithm (DPA).

    8. A method of claim 1 further including mapping said sparse vector nonzero value and said sparse vector position location using a hash function having a load factor, said hash function for producing multiple hash values, wherein the number of hash values equals the number of samples contained in the baseband frame of the primary user and the load factor for the hash is an order magnitude greater than the Complementary Cumulative Distribution Function (CCDF) threshold requirement of the PAPR constraint for the primary user signal, wherein said baseband frame of the primary user includes said instantaneous amplitude envelope measured at a rate of at least four times the base rate of the OFDM signal components.

    9. A method of claim 8, wherein said digital chaos sequence comprises a maximum spreading factor and a frame length, wherein the maximum spreading factor is determined by floor operation of the ratio of the frame length to the number of hash values.

    10. A method of claim 5, further including the step of mapping the clipping amplitudes to a scaled version of digital chaos sequences stored on the device.

    11. A method of claim 8, further including a step of updating the digital chaos sequence for use in signal amplitude recovery, and further using beacon frames with an update clip and a reset clip field element in a beacon frame signal field.

    Description

    BRIEF DESCRIPTION OF DRAWINGS

    (1) A more complete understanding of the present invention may be derived by referring to the various embodiments of the invention described in the detailed descriptions and drawings and figures in which like numerals denote like elements, and in which:

    (2) FIG. 1 is an exemplary multiple-input-multiple output (MIMO) wireless transmission system that may be used with the various embodiments of the invention;

    (3) FIG. 2 is an exemplary single-input-multiple output (SIMO) wireless transmission system that may be used with the various embodiments of the invention;

    (4) FIG. 3 is another exemplary single-input-multiple output (SIMO) wireless transmission system that may be used with the various embodiments of the invention;

    (5) FIG. 4 is another exemplary MIMO wireless transmission system that may be used with the various embodiments of the invention;

    (6) FIG. 5 is another exemplary MIMO wireless transmission system that may be used with the various embodiments of the invention;

    (7) FIG. 6 is an exemplary wireless transmitter in accordance with various embodiments of the invention;

    (8) FIG. 7 is an exemplary wireless receiver in accordance with various embodiments of the invention;

    (9) FIG. 8 is a flowchart of an exemplary method for constructing of a digital chaos sequence according to various embodiments of the present invention;

    (10) FIG. 9 is an exemplary receiver synchronization process according to various embodiments of the invention;

    (11) FIG. 10 is an exemplary embodiment of packet formation according to various embodiments of the invention;

    (12) FIG. 11 is an exemplary embodiment of null space processor subsystem of the invention;

    (13) FIG. 12 depicts a typical cell or group arrangement 900 that may be used to describe exemplary embodiments of the present invention; and

    (14) FIG. 13 is an exemplary embodiment of a group arrangement 1300 having group A, group B, and group C;

    (15) FIG. 14 is an exemplary implementation of unique key management and distribution unit based on a device identification in accordance with various embodiments of the invention;

    (16) FIG. 15 is an exemplary implementation of dual in-band side channel and primary user transceiver, wherein in-band side channel information rate is much less than the primary user information rate;

    (17) FIG. 16 depicts the composite time-domain signals of the higher powered primary user embedded with a lower powered clipped amplitude of a sparse data signal;

    (18) FIG. 17 is a prior art example of hash function implementation; and

    (19) FIG. 18 is an exemplary implementation of sparse data signal for the present invention as a hash table.

    (20) FIG. 19 is an exemplary implementation of a sparse data signal recovery transceiver which includes the preferred implementation as part of the overall architecture for compressed (e.g. sub-Nyquist sampling) and clipping error recovery.

    (21) FIG. 20 is an exemplary implementation of a sparse data signal compression using (1) compressive sensing technology or (2) Finite rate of innovations (FRI) to compress sparse data signals or (3) address hashing scheme of preferred implementation

    (22) FIG. 21 is an exemplary implementation of a sparse data signal down-sampling compression using (1) compressive sensing technology to compress sparse data signals

    (23) FIG. 22 is an exemplary implementation of a sparse data signal recovery from spherically modulated compressed samples using compressive sensing recovery algorithms to recover sparse data signals

    (24) FIG. 23 is an exemplary implementation of the second stage of compression using a spherical modulator of first-state of samples which are sent to the chaos waveform spreading algorithm

    (25) FIG. 24 is an exemplary implementation of the spherical demodulator of chaos waveform de-spreading samples.

    (26) FIG. 25 is an exemplary implementation of the chaos waveform samples de-spreading and RAKE receiver

    (27) FIG. 26 is an exemplary implementation of the Gabor Transform modulator of first stage compressed samples whose output goes directly into the chaos waveform spreading algorithms

    (28) FIG. 27 is an exemplary implementation of the Gabor Transform demodulator of chaos waveform spreading algorithms whose output goes directly into the first stage of compressed sampling recovery.

    (29) FIG. 28 is an exemplary implementation of the Trellis for Dynamic Programming Algorithm (DPA) to decode soft measurements from joint sample sparse recovery

    DETAILED DESCRIPTION

    (30) The description of exemplary embodiments and best mode of the invention herein makes reference to the accompanying drawings and flowcharts. While these exemplary embodiments are described in sufficient detail to enable those skilled in the art to practice the invention, it should be understood that other embodiments may be realized and that logical and mechanical changes may be made without departing from the spirit and scope of the invention. Thus, the description herein is presented for purposes of illustration only and not of limitation. For example, the steps recited in any of the method or process descriptions may be executed in any order and are not limited to the order presented.

    (31) The present invention may be described herein in terms of functional block components and various processing steps. It should be appreciated that such functional blocks may be realized by any number of hardware and/or software components configured to perform the specified functions. For example, the present invention may employ various integrated circuit (IC) components (e.g., memory elements, processing elements, logic elements, look-up tables, and the like), which may carry out a variety of functions under the control of one or more microprocessors or other control devices. Similarly, the software elements of the present invention may be implemented with any programming or scripting language such as C, C++, java, COBOL, assembler, PERL, or the like, with the various algorithms being implemented with any combination of data structures, objects, processes, routines or other programming elements. Further, it should be noted that the present invention may employ any number of conventional techniques for data transmission, signaling, data processing, network control, and the like. Still further, the invention could be used to detect or prevent security issues with a scripting language, such as JavaScript, VBScript or the like. For a basic introduction of cryptography, please review a text written by Bruce Schneider which is entitled Applied Cryptography: Protocols Algorithms, And Source Code In C, published by john Wiley & Sons (second edition, 1996), which is hereby incorporated by reference.

    (32) It should be appreciated that the particular implementations shown and described herein are illustrative of the invention and its best mode and are not intended to otherwise limit the scope of the present invention in any way. Indeed, for the sake of brevity; conventional wireless data transmission, transmitter, receivers, modulators, base station, data transmission concepts and other functional aspects of the systems (and components of the individual operating components of the systems) may not be described in detail herein. Furthermore, the connecting lines shown in the various figures contained herein are intended to represent exemplary functional relationships and/or physical couplings between the various elements. It also should be noted that many alternative or additional functional relationships or physical connections may be present in a practical electronic transaction or file transmission system.

    (33) As will be appreciated by one of ordinary skill in the art, the present invention may be embodied as a method, a data processing system, a device for data processing, and/or a computer program product. Accordingly, the present invention may take the form of an entirely software embodiment, an entirely hardware embodiment, or an embodiment combining aspects of both software and hardware. Furthermore, the present invention may take the form of a computer program product on a computer-readable storage medium having computer-readable program code means embodied in the storage medium. Any suitable computer-readable storage medium may be utilized, including hard disks, CD-ROM, optical storage devices, magnetic storage devices, and/or the like.

    (34) To simplify the description of the exemplary embodiment, the invention is described as being able to be used with single-input-single-output (SISO) and multiple receive antenna systems, such as, single-input-multiple-output (SIMO), multiple-input-single-output (MISO), and multiple-input-multiple-output (MIMO) wireless transmission systems. For example, the invention may be used with a SISO DSSS systems and MIMO DSSS systems as well.

    (35) It will also be appreciated that many new applications of the present invention could be formulated. For example, the present invention could be used to facilitate any conventional wireless communication medium. Further, it should be appreciated that the network described herein may include any system for exchanging data or transacting business, such as the Internet, an intranet, an extranet, WAN, WLAN, WPAN, HAN, Ad hoc Networks, mobile ad hoc networks (MANET), satellite communications (SATCOM), and/or the like.

    (36) FIG. 1 is an exemplary block diagram of a MIMO wireless system 100 with which the invention may be used. Exemplary MIMO wireless system 100 may be comprised of transmitter 102 for receiving a data source 101 (information signal 101) and transmitting the data source 101 to a receiver 104 over a transmission channel 116. Transmitter 102 may include a signal splitter 104 for receiving data source 101 and splitting the data source 101 into distinct streams of data source 101. Each of the distinct streams of data source 101 may be independently received by, and encoded by, multiple encoders 106a-n. Transmitter 102 may include multiple chaos modulators 108a-n for receiving the encoded data source and chaos modulating the encoded data source to produce multiple chaos modulated data source signals. Each of the encoded data source signals is digitally modulated and spread using a digital chaos spreading code. In one embodiment, each encoded data source signal is digitally modulated and spread with a digital chaos spreading code retrieved from a digital chaos sequence memory as discussed more fully below. In another exemplary embodiment, transmitter 102 may not include a splitter 104. Alternatively, data source 101 may be provided to encoders 106a-n.

    (37) The multiple chaos modulated data source signals may then be spatially mapped (e.g., spread over multiple spatial channels) by spatial mapper 110. The spatially mapped multiple chaos modulated data source signals may then be provided to multiple radio frequency oscillators systems 112a-n prior to transmitting the spatially mapped multiple chaos modulated data source signals over the multiple spatial communication channels 116 via multiple antennas 114a-n.

    (38) The spatially mapped multiple chaos modulated data source signals may be received by receiver 104 at multiple antennas 118a-n. The spatially mapped multiple chaos modulated data source signals may be recovered from the channel 116 using multiple radio frequency receiving systems 120a-n. RF receiver system 120a-n may recover the summed chaos modulated data source signal from the signal transmitted over channel 116. For example, RF receiver system 120a-n may recover the summed chaos modulated data source signal from the signal transmitted over channel 116 using any conventional methods for recovering a data signal from a wireless channel as are found in the art. For example, RF receiver system 120a-n may recover the transmitted signal by down converting the transmitted signal to baseband analog format and converting the baseband analog signal to baseband discrete signal.

    (39) Receiver 104 may further include a MIMO equalizer 122 for separating the spatially mapped multiple chaos modulated data source signals produced by the channel. MIMO equalizer 122 may separate the channel signals according to estimates of each channel amplitudes and phases characteristics associated with each path traverse by the spatially mapped modulated data source signal to produce received baseband modulated signals. The received baseband modulated signals may then be chaos demodulated by multiple chaos demodulators 124a-n according to data source signal channel. The multiple chaos demodulated data source signals may then be decoded by multiple decoders 126a-n. The multiple decoded chaos demodulated data source signals may then be merged by a signal merger 128 for combining the multiple data source signals into a single merged signal. In one embodiment, the merged signal may be a copy of the data source 101. Receiver 104 may provide the merged signal to a data sink 130.

    (40) Splitter 104, encoders 106a-n, spatial mapper 110, MIMO equalizer 122, decoders 126a-n, signal merger 128, and RF oscillator systems 112a-n, RF receiving system 120a-n may be of conventional construction and operation as is found in the art. The operation and construction of chaos modulators 108 an and demodulators 124a-n are discussed more fully below.

    (41) FIG. 2 is an example of an embodiment of a SIMO wireless transmission system 200 with which the invention may be used. As shown in FIG. 2, receiver 104 communicates with a transmitter 202 having a single antenna 214. Transmitter 202 contains similar elements as are found in transmitter 102, which have similar operation as is described above. For example, splitter 204, encoders 206a-n, chaos modulators 208a-n, RF oscillator system 212, and antenna 214 have similar construction and operation as is described with respect to splitter 104, encoders 106a-n, chaos modulators 108a-n, RF oscillator systems 112a-n, RF receiving system 120a-n, and antennas 114a-n. Particularly, data source 201 (information signal 201) is split into multiple distinct streams of data source. The multiple distinct streams of data source are received by multiple encoders 206a-n to produce multiple encoded data source signals. Multiple chaos modulators 208a-n may receive and chaos modulate the encoded data source signals to produce multiple chaos modulated data source signals. Each of the encoded data source signals are digitally modulated and spread using a digital chaos spreading code. In one embodiment, each encoded data source signal is digitally modulated spread with a digital chaos spreading code retrieved from a digital chaos sequence memory as discussed more fully below.

    (42) The multiple chaos modulated data source signals may then be summed at a signal summer 210 prior to being provided to a RF oscillator system 212. Transmitter 202 may then transmit the summed chaos modulated data source signal via an antenna 214. Transmitter 202 may transmit the summed chaos modulated data source signal via a communication channel 216. The chaos modulated data source signal may be received by receiver 104 at antennas 118a-n. The summed chaos modulated data source signal may be received by multiple RF receiver system 120a-n. RF receiver system 120a-n may recover the summed chaos modulated data source signal from the signal transmitted over channel 216, in similar manner as discussed with respect to FIG. 1. Chaos equalizer 218 may receive the multiple summed chaos modulated data source signals and reassemble the data source signal according to the transmitted packets.

    (43) FIG. 3 is another example of another SIMO wireless transmission system 300 with which the invention may be used. According to SIMO wireless transmission system 300, transmitter 202 is of similar description and operation described with respect to FIG. 2, Receiver 304 is of similar operation and description as described with FIG. 2. Receiver 304 for FIG. 3 may provide the multiple decoded chaos demodulated data source signals to multiple data sinks 103a-n, which may be at distinct and separate locations.

    (44) In alternate embodiments, receiver 304 may comprise multiple independent receivers where each receiver may include a chaos demodulator 124. Similarly, transmitter 202 may comprise multiple independent transmitters, where each transmitter includes a chaos modulator 208a-n.

    (45) FIG. 4 depicts MIMO wireless transmission system 100 in further detail. As previously noted, chaos modulators 108a-n chaos modulates the data source signal to produce multiple chaos modulated data source signals. A shown in FIG. 4, chaos modulator 108a-n may include a symbol mapper 402a and a chaos spreader 404a. In some embodiments, the chaos modulator 108a-n may not include a symbol mapper 402a-n. Symbol mapper 206a-n may be a conventional symbol mapper including conventional transmitter components such as a scrambler, differential encoder, symbol generator or the like. Symbol mapper 206a-206n further transmits the respective signals to chaos spreader 404a-n. Chaos spreader 404a-n may modulate the data source signal using a digital chaos spreading code sequences formed using a method included in FIG. 8. The chaos modulator 108a may chaos modulate the data source signal before the data source signal is spatially mapped (spatial mapper 402) and upconverted (RF oscillator systems 112a-n) prior to being transmitted to receiver 104, via antennas 114a-n.

    (46) According to FIG. 4, receiver 104 receives the data source signal transmitted by transmitter 102. The data source signals may be downconverted at RF receiver systems 120a-n and provided to MIMO equalizer 122. MIMO equalizer 122 separates the data source signals according to spatial channels (i.e., channel 111) and recovers the transmitted symbols and provides the signals chaos demodulators 124a-n. In one embodiment of receiver 104, Chaos demodulators 124a-n may include a symbol demapper 46a-n and a chaos despreader 408a-n. In one embodiment, chaos demodulators 124a-n may not include a symbol demapper 402a-n. Chaos despreader 408a-n uses a digital chaos sequence code for despreading the data source signal as noted below. The chaos de-spread signal may then be symbol demapped by a symbol demapper 406a-n for recovering the transmitted distinct streams of data source signal.

    (47) FIG. 5 is another embodiment of transmitter 102, As shown, transmitter 102 may not include a splitter 104. In this embodiment, transmitter 102 may not include a spatial mapper 110. Alternatively, data source 101 may be provided to encoders 106a-n. In yet another embodiment, data source signal 101a-n may be chaos spread at a chaos spreader 404a-n and provided to RF oscillator systems 112a-n prior to being transmitted to receiver 104. Receiver 104 processes the received data source signal in similar manner as is described with respect to FIG. 4.

    (48) FIG. 6 depicts another exemplary embodiment of a transmitter 102 useful with the invention. According to FIG. 6, transmitter 102 includes a channel encoder 106, a symbol mapper 402, multiplexer 604, RF oscillator system 112 wherein multiplexer 604, channel encoder 106, symbol mapper 402, multiplexer 604 and RF oscillator system 112 that are traditional elements as are found in the prior art. As such, their construction and operation is not discussed in here for brevity.

    (49) Transmitter 102 receives data source signal 101 and channel encodes the sequence at channel encoder 106. Data source signal 101 may be bits, symbols, or sampled analog waveforms. A chaos spreading code sequence, the construction of which is described below with respect to FIG. 8, is stored in memory at chaos spreading code sequence memory 606. In alternate embodiments of the invention, the chaos spreading code sequence may be used as a chaos spreading sequence without encoding the data source signal. The chaos spreading may occur without channel coding. The chaos spreading transforms the bit, symbol, or sample analog waveform into a digital chaos waveform with information embedded (embedded data) in the amplitude and phase of the digital chaos waveform.

    (50) In accordance with the present invention, chaos modulator 108 uses the chaos spreading sequence in modulation to generate a data payload having pre-ambles and mid-ambles. The pre-ambles and mid-ambles may be constructed so that multiple embedded signals can be detected at one or more locations without interference from the native performance of each constituent signal. In one particular embodiment, the data payload may be comprised of at least one chaos modulated signal and at least one other signal (either chaos modulated or not) signal that is part of a cooperative network protocol. The pre-amble and mid-amble may also be constructed by repeating the digital chaos sequence of sign flipping a copy of the digital chaos sequence in the next extended symbol period.

    (51) In one exemplary embodiment, the data payload includes pre-ambles and mid-ambles that may be constructed so that the data payload may be augmented for the inclusion of a signal field and a symbol delimiter within each of aggregated digital signals. The augmented data payload may include digital information within multiple digital chaos waveforms so that the time of arrival of each constituent signal, part of the aggregated digital signals can be identified accurately and reliably. A signal field portion instructing the receiver of at least one length information of the digital signal and data rate scheme information for the remaining payload. Further, the signal field may contain parity information for protection against and detection errors of other information within the signal field.

    (52) During operation of chaos modulator 108, the data source signal is spread with the chaos spreading sequence stored in chaos sequence memory 606 using, for example, spreader 602. The chaos spreading sequence may be used in the generation of the pre-amble 608 and the mid-amble 610. The payload generated by chaos modulator 108 may be augmented to include the symbol delimiter 612 and signal field 614 as is described with respect to FIG. 10.

    (53) FIG. 6 depicts any one of the chaos modulators 108 (or chaos modulator 208) in further detail. According to FIG. 6, transmitter 108 may include a channel encoder 106 and a symbol mapper 402 of similar operation as described above. Chaos modulator 108 may include a mixer 602, 616, a multiplexer 604, a RF oscillator system 112, and antenna 114 that are traditional elements as are found in the prior art. As such, their construction and operation is not discussed in here for brevity.

    (54) During operation, transmitter 102 receives a data source signal at encoder 106 which channel encodes the data source signal. The data source signals may be any information bearing signals such as bits, symbols, or sampled analog waveforms.

    (55) A chaos spreading sequence, the construction of which is described below with respect to FIG. 8, is stored in memory at chaos spreading sequence memory 606. In alternate embodiments of the invention, the chaos spreading sequence may be used as a spreading sequence without encoding the information signal. The chaos spreading transforms the bit, symbol, or sample analog waveform into a digital chaos waveform with information embedded (embedded data) in the amplitude and phase of the digital chaos waveform.

    (56) The digital chaos sequences stored in chaos sequence memory 606 are constructed using, for example, the digital chaos sequence generation method 800 of FIG. 8. The digital chaos spreading code sequence may be constructed by recording native analog chaos circuit. Alternatively, the digital chaos sequence may be constructed by recording computer simulated non-linear dynamics of deterministic and segmenting the recorded signal. (Step 802) The recorded segments are sampled such that successive samples appear independent, and segments of a predefined length and variable quantity have low cross correlation. (Step 804) The samples may then be stored in memory. (Step 806) Sampling rate or sampling period may be varied or irregular, but the number of samples taken is fixed for a particular spreading factor and can be any number of samples for the spreading factor. In accordance with the invention, the segments are quantized. (Step 810) The mean value of the recorded segments is then subtracted and the recorded segments are normalized. (Step 812) Normalization of the sequence ensures that autocorrelation peak occurs at unity or near unit because of the quantization errors.

    (57) An irregular sampling interval according to the invention may be, for example, determined by modulo counting of a known sequence generator such as Fibonacci numbers, Lucas numbers, Perrin numbers or any pseudo random number generators. For implementation ease with semiconductor technologies for digital system, the amplitudes may be quantized to finite levels based on the maximum allow cross-correlation (.sup.L), where is L is the number of bits used to represent by each sample amplitude) between code sequences. Independent segments of the digital chaos sequences are grouped together to form a vector span for transmitting the information-bearing communication signals or training signals. It is well-known in mathematics that any signal in an n-dimensional subspace can uniquely represented an n-tuple of scalar corresponds to the projection of the signal onto the orthonormal bases of the n-dimensional. The final step of the digital chaos process is to convert the independent digital chaos segments into a group of orthonormal sequences spanning the same subspace as the original segment. This process may be performed using the Gram-Schmidt orthogonalization process.

    (58) The chaos sequence memory 606 (and the chaos replica memory 706 of FIG. 7) may be a volatile memory. Chaos memory 606/706 may be partitioned such that groups of digital chaos spreading codes are stored independently of each other. For example, the distinct groups may be organized according to the application for which it will be used. Typical applications include any wireless applications requiring voice over IP (VoIP) capability, video capability, and data capability for point-to-point operation and/or point-to-multi-point. Inside the groups, the volatile memory is further partitioned into slots for storing a digital chaos sequence code. The slot may be further partitioned into a plurality of sub-slots for storing subsets of the of the digital chaos sequence.

    (59) Once the chaos sequence memory 606 is fully populated with digital chaos spreading sequences, the entire memory 606 may be subjected to Gram-Schmidt processing. The entire memory 606 may be subjected to an orthonormalization process. In alternate embodiments, independent digital chaos segments may be converted into a group of orthonormal sequences spanning the same subspace as the original segment.

    (60) A preferred embodiment of the invention for the packet formation is shown in FIG. 10 In this exemplary embodiment the sample rate at the receiver is targeted at 20 MHz and the chipping rate is proposed at 4 Mcps at the transmitter. The minimum center frequency spacing between adjacent systems will be 5 MHz. The framing structure may be a radio frame of 10 ms divided into 5 sub-frames of equal duration 2 milliseconds (ms). These sub-frames may be configured as transmit or receive slot for any user.

    (61) A super-frame consists of several frames transmitted in succession with 2 ms gap spacing between frames. Each frame to be transmitted consists of a preamble training sequence, mid-amble training sequence, and data payload. The flexibility of frame structure can accommodate many other embodiments to specific applications. In this embodiment, sufficient training information is included to securely and reliably.

    (62) As is well known, the key to a successful wireless design is to incorporate sufficient training information to recognize the arrival of packets, align symbol boundaries, estimate channel characteristic and correct for frequency offset. In one embodiment of the invention utilizes a header field. The header field comprises a ten symbol preamble and 48 symbol signal field that defines the configuration state for the receiver. The training sequences are modulated using differential chaos shift keying (DCSK) and repeated predetermine number of times; nine times is used in FIG. 6. Each repetition is modulated with either a 1 or 1 according to normal DCSK techniques. The modulation input can be an alternating sequence of positive and negative ones, embedded with control information for the rest of the packet. The preamble and mid-ambles can have their powers significantly higher than the data to aid in the synchronization at the receiver. For example, one embodiment uses a 3 dB boosted in relative power to the data samples. This will permit the high probability of detection without an overly burdensome overhead for the frame. If total overhead is 10% or less in duration for the frame, significant improvement in detection and synchronization at the receiver is achievable while sacrificing only 0.79 dB is signal power compared to no power boost. Each symbol is comprised of a chaos sequence of predetermined length that can range from 16 chips to 4000 chips, depending on the application requirements for throughput and covertness. The signal field is comprised on a 6 bit scrambling seed, which is used to initialize the pseudorandom number (pn) generator for sequence pattern. The state of the registers of the pn determines which of 2.sup.6 stored sequence is selected or, optionally, which sequence in the chaos family should be transmitted for the current symbol.

    (63) As previously noted, the present invention addressed problems in traditional MIMO WLAN transmission. Namely, prior art systems such 802.11x compliant system are more susceptible to interference, wireless collisions, and interception by unintended parties. The present invention addresses these problems by providing a system and method for aggregating and embedding multiple information-bearing communication signals within digital chaos communication waveforms occupying the same frequency channel bandwidth transmitted with a multiple antenna system. Digital chaos may be a waveform generated by sampling a chaos signal, where chaos signals are determined by deterministic nonlinear dynamics. Digital chaos sequences generated according to the invention as described below, are used as a spreading sequence in accordance with various embodiments of the invention.

    (64) The signal transmitted by transmitter is received by a receiver 104 of FIG. 7. Receiver 104 recovers the data embedded data in the transmitted data source signal. In accordance with one aspect of the invention, the digital chaos sequence stored in chaos sequence memory 606 used as a spreading code for spreading the data source signal at transmitter 102 is compared to the digital chaos sequence stored in chaos replica memory 306. As noted before, chaos demodulator 124 uses a replica of the digital chaos sequence stored in chaos replica memory 706 as a despreading code.

    (65) FIG. 7 is an exemplary embodiment of receiver 104 for receiving the digital chaos waveform. Receiver 104 includes an antenna 118 for receiving the transmitted signal, channel filter 702 to reject signals not in the band of interest, analog-to-digital (A/D) converter 704 is used to sample and quantization the analog signal for digital processing. A chaos replica memory 706 provides an exact replica of the digital chaos waveform stored in chaos sequence memory 606 at the transmitter 102. Packet detection 708 operation is performed to determine when at least one packet arrives. A matched filter 710 is used to recover symbol timing for at least one signal. A channel estimator 712 may be used to estimate and compensate the distortions to the waveform due to multipath fading. Doppler Correction 714 may be used to estimate and correct frequency offsets to due oscillator drift and mobility. Receiver 104 may also include a symbol detection device 716 to estimate the mapping symbol sent by the transmitter, a symbol D-map look-up table 718 to recover informational symbol, and a channel decoder 720 to recover the original transmitted bits 722. The recovered information bits 722 may be provided to one or more data sinks (not shown).

    (66) Receiver 104 receives the transmitted signal to recover the embedded data from the signal. It should be noted that there may be two common receiver modes in preferred embodiments. In the first mode, the high-speed multiplication with chaos replica memory 706 occurs directly after operation of A/D 704. This embodiment is preferred when a sampled analog waveform is the information-bearing signal. In the second mode, the high-speed multiplication with chaos replica memory 706 occurs prior symbol detect 716 and after Doppler Correction 714 and channel estimation 712 operations. This embodiment is best suited when the information-bearing signal, bits or symbols. Either configuration works for the information-bearing signals in the form of bits or symbol. However, the second mode has the best performance and the first mode has the lower power consumptions. After despreading the high-speed digital chaos sequence, the receiver operations are typical of those performed by commercially standard receivers for 802.11x, WCDMA, or CDMA 2000, the description of which is omitted for the sake of brevity.

    (67) The chaos modulator 108 and demodulator 124 may be implemented as part of a wireless local area network (LAN), wireless personal area network (PAN), wireless home area network (HAN) or metropolitan area network (MAN) system, a cellular telephone system, or another type of radio or microwave frequency system incorporating one-way or two-way communications over a range of distances. The invention may employ various signal modulation and demodulation techniques, such as single-carrier frequency domain equalization (SCFDE), direct sequence spread spectrum (DSSS) or orthogonal frequency division multiplexing (OFDM), for example. However, throughout this description, references are made with respect to a SIMO and MIMO communication systems or a system including a transmitter and receiver merely to facilitate the description of the invention. All the similar components of the wireless channels 711 will also have similar descriptions to each other.

    (68) The transmitters of the present invention may transmit different signals from each antenna in transmit antenna array so that each signal is received by the corresponding antenna in a receiving antenna array at the receive side. Various transmitters described herein may transmitted the data source signal as an aggregate signal and received as an aggregation of all the transmit signals, or an aggregation of parts of the signal. All signals are transmitted once and the receiver demodulates the aggregate signal using a replica of the chaos spreading sequence spreading code stored in chaos replica memory at the receiver.

    (69) FIG. 9 illustrates a method 900 for recovering the data embedded in the transmitted data source signal. In recovering the data, receiver 104 receives the transmitted signal and recovers the data signal by the following steps depicted in FIG. 9: The packets are continually searched until the receiver detects the arrival of a valid packet (step 902). The detection of the packet is based on the output of a free-running correlation (packet detected 708) that exploits the preamble structure. The validity of the packet is determined from the cyclic redundancy check (CRC) of the signal field shown in FIG. 10. After the packet has been declared valid, the preamble is used to perform two synchronization processes: symbol timing estimation & correction (step 904) and frequency estimation & correction (step 906). A match filter or bank of matched filters 710 is used to estimate the timing error and the appropriate correction is made in the receiver timing. A separate correlator is used to estimate the frequency error, such as for example Doppler Correction 714 and the appropriate correction is applied to the baseband received signal. The channel estimate is computed using the pre-computed convolution matrix based on the training symbols from the preamble. The pseudo inverse of this matrix, which can be also computed off-line since it doesn't change unless the preamble changes, is used to compute the minimum mean square estimate of the channel taps (channel estimator 712) (step 908). Averaging is possible for each of process steps 902, 904, 906, and 908 based on the repetition of the training symbols in both the preamble and mid-amble. The payload may then be processed (step 910). For example, processing the payload may include detecting a symbol (symbol detection 716), mapping the symbol (Symbol D-Map 718), channel coding or decoding the payload (channel decoder 720) and recovery of information bits 722 contained in the payload.

    (70) FIG. 11 is an embodiment of an exemplary null-space processor subsystem which may be useful with the communication systems described in this invention. In accordance with this exemplary subsystem, the signal to be recovered (the Selected i.sup.th User Data) and the remaining signals (the Remaining User Data) are multiplied in the null space processor (Null Space for i.sup.th Selected User corresponding to the Selected i.sup.th User Data producing a signal containing the Remaining User Data signals. The Remaining User Data signals are then subtracted from the signal containing the Selected i.sup.th User Data and the Remaining User Data such that Selected i.sup.th User Data is output. In some instances, the output Selected i.sup.th User Data may appear with residual signals from the Remaining User Data. The Selected i.sup.th User Data may then be recovered by using the Selected i.sup.th User Data to identify the i.sup.th User Chaos Code for recovering the i.sup.th User Data as described above.

    (71) The digital chaos systems and methods of the present invention are suitable for operation in wireless transmissions desiring coordinated transmissions to enhance network capacity. Such systems often require multiple transactions between tightly coordinated nodes or access points. By coordinated, what is meant is that the transmission protocol of each node (receiver) in a network is organized into a relationship with a second receiving node in the network to ensure that subsequent transmissions are efficient according to the requirements of the network. By efficient, what is meant is that the node to node transfer is optimized according to the requirements of the node or the requirements of the transmission medium. In one instance, improving efficiency may include improving the throughput of the network. Coordinated nodes may be such that one node, or a group of nodes may include a transmission protocol that depends on the transmission received from one node in the network or the group of nodes. In the instance where multiple nodes depend on a transmission received, and the transmission protocol of a multiple of nodes may cooperatively communicate to ensure optimization of the network or transmission medium.

    (72) One transmission protocol that can require coordination is adaptive changing of channel by a group of nodes. By channel, what is meant is an available time slots, hopping frequency, independent spatial path, or distinct digital chaos sequence at a transmitting side. By time slot, what is meant is the next transmission opportunity (Txop) after a channel reservation request expires, the channel is idle for a specified period of the time including time requirement for a delay and disruption tolerant protocol known at the receiving and transmitting side, or the channel is busy and the number of digital chaos signals are less than the multi-user limit for simultaneous transmission. By hopping frequency, what is meant is any of the tunable center frequencies of capable by the equipment that is permitted in a hopping sequence; whereby the minimum spacing between center frequencies is equal to the chipping rate of a generated digital chaos sequences at the transmitting side. By independent spatial path, what is meant is that separate paths arriving at the receive side wherein the cross-correlation between paths are sufficiently small such that low to ensure the transmission of the information at the rate and with the quality required under specified conditions. And distinct digital chaos sequence at a transmitting side may originate from the same source or separate sources. In the case wherein they originate from the same antenna, the distinct digital chaos sequences are orthogonal by construction and hence meet the requirement to be considered separate channels. In the case wherein distinct digital chaos sequence at a transmitting side originate from difference antennas, results in sufficient differential signal attenuation for different arriving paths at the receive side coupled with the cross-correlation between distinct digital chaos sequences to ensure each transmission of the information at the rate and with the quality required under specified conditions. Determination that the channel of the wireless medium is busy or currently in use is achieved by sensing the physical channel using one of several method techniques taught in the prior art (referred to as detectors) and comparing against a predetermined threshold. Some of these detectors can be applied directly to our circumstances; others have to be adapted to exploit the unique properties of the digital chaos. If the channel metric measured exceeds the predetermined threshold, clear channel assessment (CCA) is set false. For instance, the prior art teaches energy detection as a means of determining assessing CCA for many wireless protocol as a failsafe. An energy detector does not exploit a signal structure or property of the incoming signal at the receive side thus can be applied to any signature; however, some signals such as the digital chaos signals are constructed to reduce the probability of detection by these energy detectors. Other physical channel sensing methods exploits the construction of the signal such as cyclical detectors. Most communication systems exhibit some known periodicity to its structure such as periods of the symbol clock, sample rate clock for intermediate frequency (IF) carrier or RF carrier or even repetition of sequences for training purposes. These known periodicities can be incorporated in a cyclical detector by correlating different segments separated by the known periods and computing the energy over the window of data for which cyclostationary property is valid. For instance, differential chaos shift keying (DCSK) shown in FIG. 6 illustrates the said cyclostationary property by repeating copies of digital chaos sequences a predetermined number of times9 in this case. A cyclic detector configured to the proper delay and detect the arrival of the preamble and mid-amble sequences if it also knows and the data wiping technique to the incoming sequences. In another prior art, a method of physical channel sensing entails using a match filter. The matched filter is the optimal linear filter for maximizing the signal to noise ratio (SNR) in the presence of additive stochastic noise. A match filter is a linear filter whose coefficients are equal to time reverse, complex conjugate of the sequence is attempting to detect. This method requires duplicate replica of the digital chaos sequences at the receive side to detect the presence of a digital chaos sequence. In yet another prior art, a method of physical channel sensing using Bayesian statistics is employed. In Bayesian statistics, a maximum a posteriori probability (MAP) estimate is a mode of the posterior distribution. The MAP can be used to obtain a point estimate of an unobserved quantity based on empirical data. In the present invention, the unobserved quantity would be one of the allowed digital chaos sequences for the user. Based on the observation at the receive side, one would like to determine the likelihood that a specific digital chaos sequence was sent based on the collected samples at the receiver. In other words, one wishes to maximize the likelihood function of sample distribution at the receive side. Using direct application of Bayer's theorem, we can compute and maximize the likelihood of a digital chaos sequence was sent from the marginal likelihoods of hypothesized likelihoods conditioned on each allowable digital chaos sequence and no sequences sent.

    (73) In the instance of multiple transmissions, as noted above, multiple transmissions may create increased opportunities for compromised data transmission or collisions of data transmission. In one embodiment of the invention, the coordinated nodes may include knowledge of the transmission protocol one or more of the other nodes in the network. Alternative, the coordination of the transmissions from one node, or a group of nodes, may depend on the transmission received from a node outside the network or group. In another embodiment, coordinated nodes or coordinated transmission over a wireless medium may mean that transmission from one node coordinated with another node may occur at the next transmission opportunity (Txop) or within the time specified by a delay and disruption tolerant protocol known at the receiving side.

    (74) The digital chaos waveform described herein may be used to secure the data transmissions while improving network throughput. For example, coordinated multi-point transmission and reception over heterogeneous wireless networks comprise a set of disparate transmission points, access points or nodes operating in the same cell (e.g., group), overlaps cells, or mutually exclusive cells, simultaneously or in a coordinated fashion. Coordinated multi-point transmission may be used to utilize to increase throughput and service quality in wireless networks, particularly at or near the edge of a given cell in a cellular network or group of nodes, access points or users.

    (75) A typical cooperative network that may be used with this invention is the Internet of Things (IoT). The IoT refers to interconnection and autonomous exchange of data among devices which are machines or parts of machines. IoT may typically be used to support, for example, Machine-to-Machine (M2M) communication. M2M is defined as data communication among devices without the need for human interaction. This may be data communication between devices and a server, or device-to-device either directly or over a network. Examples of M2M services include security, tracking, payment, smart grid and remote maintenance/monitoring. Consequently, a coordinated network according to the invention may include the autonomous exchange of data among devices nodes or members of the coordinated network.

    (76) As used herein, nodes belonging to a single cell may be described as members of a single group. In some instances, to facilitate the coordination of wireless transmission, members may be described as members belonging to one group, or to more than one group. Signals received by a specific member may be further processed according to the signal preamble or mid-amble information. Membership to a group requires first a request to join a group by node. The process of joining is characterized by three distinct states: {unauthenticated, unassociated}, {authenticated, unassociated}, and {authenticated, associated}. In one embodiment of this invention, authentication is performed by unique preloaded authentication digital chaos sequences associated with unique identifiers for communicating device. The coordinator for the group has access to a repository of all complementary authentication digital chaos sequences for other authorized devices. These complementary authentication digital chaos sequences are used in the handshaking exchange for response queries between the requesting node and the coordinating node. Analogous to the cryptography exchange procedure wherein a node A desiring to communicate to a node B, send it message encrypted with node B public key. Node B uses it private key and the incoming message to decrypt the package. In this case, Node B demodulates the digital chaos sequences using its complementary authentication digital chaos sequences along with its private unique authentication digital chaos sequences. Node B uses its private unique authentication digital chaos sequences or derivative of to encapsulate any response frame including an acknowledgement frame. In yet another embodiment, after a node has been authenticated it transitions to state two within the coordinating node and an association to a specific group is started. Association to group includes but not limited to exchanges of capability information to the coordinator and verification of the coordinator of a set of rates, digital chaos sequences for data frames, beacon frames, request-to-send (RTS) frames, clear-to-send (CTS) frames and group acknowledgement frames. Completion of this step transitions the state of the node to membership granted status as the coordinator node. The coordinator node provides in one of its data payload of beacon frame information containing a temporary local ID for subsequent communications as well as informs other members of the group of the new member credentials for communicating. Membership to a group is not permanent and maybe voluntarily terminated by the any member group. In additional, the network manager or coordinator may terminate the membership for any member deem detrimental to operation of the network. For example, a rogue node may temporarily gain access to the network but based on its traffic pattern but later kick out of group membership whereby not member will engage in future transmission with that node.

    (77) In one embodiment of this invention, authentication digital chaos sequences can share the same level of difficulty to securely update, particularly without exposure to unauthorized listeners or distribute to a large number of users based on a similar process central authority called the certificate authority (CA) in key management for public-key cryptography. Strong cryptography designs strictly adhere to Kerckoffs's Principle in design good encryption scheme; that is, the security of the encryption scheme must depend only on the secrecy of the key and not on the secrecy of the algorithm. The rationale behind the rule is that algorithms are hard to change since they are normally built into software or hardware, which can be difficult to update. In the present invention, the network coordinator that provides some CA functionality for private unique authentication digital chaos sequences.

    (78) As used herein, the coordinated multi-point system may be a MIMO system, wherein the members may use multiple antennas at both the transmitter and receiver. The present invention may be also useful multi-user multiple-input-multiple-output or MU-MIMO systems. As used herein MU-MIMO systems are wireless communication systems in which available antennas are spread over a multitude of independent group members, access points and independent radio terminals, wherein each member has one or multiple antennas. The present invention may also be used with conventional SISO (single input-single output), SIMO (single input-multiple output), MISO (multiple input-single output) systems, or other similar systems as is known in the art.

    (79) FIG. 12 depicts a typical cell or group arrangement 900 that may be used to describe exemplary embodiments of the present invention. Group arrangement 900 may include a group A having members A1-An. Similarly, group B may include member B1-Bn. In the example shown, group member B3 may also be a member of group B and also group A. To facilitate description of the invention, when a member belongs to more than one group, the member may be designated with both group designators such as A/B3 depicted in group arrangement 1200.

    (80) FIG. 13 is an exemplary embodiment of a group arrangement 1300 having group A, group B, and group C. While the multiple groups of group arrangement 1300 are depicted as being imbricate, the groups may be mutually exclusive. Where a member belongs to more than one overlapping group, the member may receive and process digital chaos signals addressed to either one of the overlapping groups. In such instance, the member may be called a multi-group member, herein. In the arrangement shown, members A/B3, indicating that the member B3 also belongs to group A.

    (81) Group arrangement 1300 further depicts the wireless transmissions that occur when one digital chaos signal is transmitted between members. For example, member A6 may transmit digital chaos signals to member A2, A5 or An; member C1 may transmit digital chaos signals to C8; B3 may transmit digital chaos signals to B1, B4, or B9 depending on how the digital chaos signal is addressed in the digital chaos preamble. In the instance where the digital chaos is received by a multi-group member, the receiving member may transmit the digital chaos signal to the corresponding group member to which the receiving member belongs. This may be true even when the intended group member belongs to a separate overlaps member. By overlaps, what is meant is that more than one group shares at least one group member. In the group arrangement 1300 shown, group A overlaps with group C, and group A overlaps with group B.

    (82) It should be noted that the digital chaos signals discussed with respect to the embodiments in FIG. 12 and FIG. 13, may be received from outside any one or all the groups depicted. For example, digital signal S1 is depicted as being generated outside of group A, but received by group member A2. Similarly, digital chaos signal S2 is depicted as being generated outside group C, but being received by group member C8. Contrarily, digital chaos signal S3 is depicted as being generated in group B by group member B7 and received by group member Bn.

    (83) In a typical coordinated transmission according to the present invention, group members operate in a coordinated fashion to improve the overall network capacity for all members sharing the wireless medium. By coordinated fashion what may be meant is that signals are processed together to combat the distortive effects of the wireless medium and mutual-interference to improve the detection for all active distinct user data signals. The group members process the extracted distinct user data signals according to membership or non-membership to a specific group. The group member may process the extracted distinct user data signals according to membership or non-membership to a particular group wherein said extracted distinct user data signals not addressed to the receiving side (e.g., receiving member) are aggregated, time aligned, and retransmitted over the wireless medium at the next transmission opportunity (Txop) or time specified by a delay and disruption tolerant protocol known at the receiving side. For example, the delay and disruption tolerant protocol may be such that the network operates effectively over extreme distances such as those encountered in space communications or on an interplanetary scale. On the other hand, where a digital chaos signal is received by a receiving group member to which it is not addressed, the receiving group member may terminate the signal and not forwarded it at all.

    (84) In a typical example, using FIG. 13, data signal S4 may be addressed to group member B9, but received by multi-group member A3 (e.g., A/B3). In this instance, group member A3 may extract the distinct user data signal indicating that the data signal is addressed to group member B9. Since group member A3 is also a member of group B (e.g., B3), then group member A3 may transmit the signal to group B. More particularly, A3 may transmit the signal to group member B9.

    (85) In some instances, where group members of different groups are in proximity to each other, a receiving group member may receive a first fragment of the received signal, and time delay transmission of the received signal until a second fragment of the received signal is received by the receiving group member.

    (86) It should be appreciated by one skilled in art, that the present invention may be utilized in any device that implements the DSSS encoding scheme. The foregoing description has been directed to specific embodiments of this invention. It will be apparent; however, that other variations and modifications may be made to the described embodiments, with the attainment of some or all their advantages. Therefore, it is the object of the appended claims to cover all such variations and modifications as come within the true spirit and scope of the invention.

    (87) In another example, the keys for dual level network access is shown using FIG. 14. Private keys are preload onto devices and public keys are managed by each local network manager or coordinator for a group. Public and private keys are paired and generated by the same chaos function. The chaos function may be any chaos function implemented on the device that is known at both the transmitter and the receiver.

    (88) A preferred embodiment of an exemplary diagram of a heterogeneous transceiver 1500 according to the present invention is shown in FIG. 15. As shown, a high PAPR baseband signal 1501, such as, for example, LTE or Wi-Fi compliant signal undergoes a clipping function 1502 to produce a low PAPR baseband signal 1503 and a sparse clipping baseband clipping signal 1505 for transmission. The low PAPR baseband signal 1503 is fed directly to an Analog Front End (AFE) 1506b of a digital chaos Transmit chain. AFE 1506 may be comprised of more than one digital upconversion (DUC) and digital-to-analog converter (DAC) (DUC, DAC) 1506a and (DUC,DAC) 1506b. In one embodiment, the sparse baseband clipping signal 1505 is fed to a digital chaos modulator system 1504a. Digital Chaos Modulator System 1504a may include a chaotic sequence generator for generating chaotic spreading sequences in accordance with the present invention. Digital chaos modulator system 1504a may further include a chaotic spreader for receiving the sparse baseband clipping signal 1505, and spreading the sparse baseband clipping signal 1505 with the chaotic spreading sequences generated by the chaotic sequence spreader to produce a chaos spreaded sparse baseband clipping signal. The digital chaos modulator system 1504a may further include a digital chaos modulator for receiving and modulating the chaos spreaded sparse baseband clipping signal. The digital chaos modulator system 1504a may output the resulting modulated signal to AFE 1506b of Digital Chaos transmit chain. The outputs of (DUC, DAC) 1506a and (DUC, DAC) 1506b may then be summed together 1506 prior to being modulated onto a RF carrier and transmitted by an antenna 1508.

    (89) Since both the signal produced by (DUC, DAC) 1506a and the signal produced by (DUC, DAC) 1506b are modulated on the same RF carrier, the two signals appear spectrally in-band relative to each other. The resulting RF modulated signal may be transmitted out at least one antenna 1508.

    (90) Ata receiving side, at least one antenna 1510 may receive the RF modulated signal transmitted by antenna 1508. The RF modulated signal received at antenna 1510 may be provided to a downconverter system 1512 including an analog-to-digital converter (ADC) for converting the signal to a digital signal, and a downconverter (DDC) for down-converting the digital signal. The down-converted signal at downconverter system 1512 may be split into two paths by a splitter (not shown). A first signal path may be provided to a digital chaos demodulator system 1504b for producing an estimate of the sparse data info samples. For example, the first signal path may be provided to a digital chaos demodulator for demodulating the downconverter system 1512 signal. The demodulated signal may then be despreaded by a chaos despreader to recover the sparse baseband clipping signal.

    (91) The signal recovered by digital chaos demodulator system 1504b may be sampled to produce the sparse data information samples 1513 of the sparse clipping baseband clipping signal 1505. In another embodiment, the signal recovered by digital chaos demodulator system 1504b may be sampled to produce the recovered clipped amplitudes of the sparse clipping baseband clipping signal 1505 (i.e., a sparse clipped amplitude vector 1513), in which case the recovered signal is sparsely distributed in the estimated time positions of the clipping event within the original frame of the LTE or Wi-Fi baseband receive signal.

    (92) A second signal path provided by down-converter system 1512 is signal conditioned to produce low PAPR soft data samples 1511. In one example, the low PAPR soft data samples 1511 and the sparse clipping baseband clipping signal 1513 may be subjected to a signal recovery algorithm 1514. The signal recovery algorithm 1514 processes the low PAPR soft data samples 1511 and recovered sparse clipped amplitudes vector 1513 to produce a received baseband signal substantially void of clipping distortion for final processing at the primary user demodulator 1516. For example, such baseband signals substantially void of clipping distortion are typically required to decode LTE or Wi-Fi compliant signals.

    (93) The present invention describes an improvement in wireless side-channel feedback information using specialized data compression and extraction processing. In the prior art, the wireless side-channel feedback information is either sent on a separate signaling channel or in-band signaling in lieu of some data samples. Both situations lead to increase overhead for the transport of the intended information. Others have sought to quantize side-channel information and restrict the number of bits per side-channel sample (e.g. 3 bit quantization) to reduce the overhead burden but that leads to performance degradation due to the higher quantization error resulting from lower bit resolution of the side-channel sample. When the side-channel information is the result of computed errors resulting from transmitter function, not only are the value of the errors required but the location of those errors are needed as well. In the present invention, we exploit the properties of the Gabor Algorithm (GA) to compress the error values and locations and zero position in the error signal into a lower rate signal comprising all non-zero values. This reduces the amount of information that is need to be sent over channel without the associated degradation in performance experienced in the prior art. In a preferred embodiment of the present invention, these lower rates side-channel information is transmitted cooperatively using Digital Chaos in-band signaling with the original undistorted signal.

    (94) FIG. 19 describes an alternative embodiment of a transceiver 1900 for transmitting clipped error samples (e.sub.n), associated with GA, and recovering the clipped error samples over the same in-band channel. The transceiver 1900 may include a transmitter 1902 for compressing, encoding, and modulating the clipped error samples. Transmitter 1902 additionally spreads the modulated clipped error samples using the chaos spreading sequences as described in accordance with various embodiments of the invention (i.e., chaos waveform spreader 1910).

    (95) As discussed below, the exemplary transceiver 1900 and the methods described according to the invention minimizes PAPR. The PAPR may be minimized at the transmitter and received (i.e., measured) at a receiver. Transceiver 1900 does not necessarily require knowledge of the location and amplitudes of a sparse signal such as described with respect to FIG. 15 since the location and amplitude of the sparse signal are extracted directly from compressed sparse signals. However, information such as the total number of sparse signals without knowledge of location and amplitudes improves the performance of the compressive sampling algorithms.

    (96) Exemplary transmitter 1902 may include a compressive sampling unit 1906 for compressing and sampling the clipped error samples. Compressive sampling unit 1906 may compresses the clipped error samples for data aggregation. Compressive sampling unit 1906 may comprise, for example, address hashing, a finite rate innovations (FRI) encoder, or a compressive sensing encoder.

    (97) Where compressive sampling unit 1900 use an FRI, the method compresses sparse signals to a small number of compressed samples that can be reconstructed at the receive side. Where the compressive sampling unit 1900 uses a compressive sensing encoder, the method used is a non-parametric technique to improve robustness. Additionally, where compressive sampling unit 1900 uses address hashing, the transceiver may perform further processing of the clipped error samples with a second stage of sample compression using spherical modulation, which may improve noise immunity at the receiver.

    (98) Transmitter 1902 may further include a spherical encoder and modulation unit 1908 for receiving the compressed and sampled clipped error samples from compressive sampling unit 1906. Spherical encoder and modulation unit 1908 may provide additional compression of the already compressed and sampled clipped error samples. In one instance the spherical encoder and modulation unit may 1908 may map the amplitude of the compressed and sampled clipped error samples to the phase of the compressed and sampled clipped error samples to improve receiver signal to noise ratio (SNR) performance for reconstructing compressed samples.

    (99) Transmitter 1902 may further include a chaos waveform spreader 1910 for receiving and the spreading the spherically encoded and modulated clipped error samples using the chaos spreading sequences as described in accordance with various embodiments of the invention. In preferred embodiments of the invention, chaos waveform spreading unit 1910 performs spreading for improved multiple access interference resistance. Digital chaos waveforms and waveform classes are configurable to many spreading factors and lengths. As shown, of both clipped IDFT signals (x.sub.n) and clipped error (c.sub.n) are summed at summer 1912, preferably at the sampling rate of clipped signals. In which case, both signals may be transmitted in-band multiple access transmission.

    (100) FIG. 19 shows an exemplary receiver for use in receiving the signal transmitted by transmitter 1902. Receiver 1904 may be configured to receive the signal, despread the signal using the chaos sequencing as described herein, demodulate and decode the despread signal, sparse reconstructed before recovering the clipped error samples processed and transmitted according to as is described with respect to exemplary transmitter 1902.

    (101) For example, receiver 1904 may include a chaos waveform despreader and rake receiver unit 1916 for receiving the signal transmitted by transmitter 1902. Chaos waveform despreader and rake receiver unit 1916 may despread the received signal for example, to restore the spherically modulated sample through processing gain. The rake receiver of unity 1916 may be used for combining multipath signals in a fading wireless channel, thereby enhancing the signal decoding process, such as performed by a spherical demodulator and decoder unit 1918. Spherical demodulator and decoder unit 1918 may use the same compression sensing algorithms used by spherical encoder and modulator unit 1908 to demodulate and decode the despreaded signal and thereby recover the compressed sparse signals discussed with respect to the output of compressive sampling unit 1906.

    (102) Sparse clipped error signals may be reconstructed to restore sparse clipped error signals using the said compression sensing algorithms. For example, a sparse reconstruction unit 1920 able to perform one of re-hashing, FRI, or BPDN on the signal from spherical demodulator and decoder 1918.

    (103) The original transmitted signal received by receiver 1904 may then be recovered by summing the estimates of the clipped signal to the estimates of the clipped error.

    (104) FIG. 20 is an illustration of the steps which may occur at for example, compressive sampling unit 1906, for compressive sensing down-sampling of sparse clipped error input signals. For multiple block processing, a contiguous block of (N) sparse data samples is down-sampled to a compressed sample size of (M), where M<<N, by using a matrix operation () on the input sparse samples, where [] is a matrix of (M) rows and (N) columns of random noise (e.g. samples of a standard normal probability density function or can be a deterministic matrix). For example, if each time epoch for a sparse signal occurs at (nT) time epochs, then the output samples occur at each (knT) time epoch, where (k>>1, or (N/M=k)). This means a much lower sampling rate than the original sample rate of the input sparse samples. For multiple block processing, this step is repeated for all input samples to encode the entire sparse input signal.

    (105) In a typical example, M<<N for compressive sensing reconstruction may ordinarily be accomplished with M>2*m, where m is the number of sparse samples in a sparse signal. Consequently, M>2*m/N of 10%-20% of N.

    (106) FIG. 20 shows that the output of compressive sampling unit 1906 may be a time series of compressed samples at a lower sampling rate. Equation 1 shows the processing in matrix-vector format to used get compressed samples in accordance with various embodiments of the invention, where
    [].sub.MN=[.sub.1.sub.2.sub.3 . . . .sub.N1.sub.N],.sub.i=[.sub.i,1.sub.i,2.sub.i,3 . . . .sub.i,M].sup.T [g].sub.N1=[].sub.MNe,e=[e.sub.1e.sub.2e.sub.3. . . e.sub.N1.sub.N].sup.T, Equation 1

    (107) FIG. 21 shows is an illustration of the steps which may occur at for example, sparse reconstruction unit 1920 for producing the output of Compressive Sensing sparse signal recovery for input received from spherical demodulator and decoder 1918. The spherical demodulated and decoded samples are estimates of the compressed sparse signals produced by compressive sampling unit 1906. Any number of sparse signal recovery algorithms can be used to estimate sparse signals (e) such as L1-norm.

    (108) FIG. 22 shows an exemplary method for generating an output of spherical modulation encoder unit 1908. The output of unit 1908 may be produced, for example, using inputs from compressive sampling unit 1906. According to the spherical modulation and encoding method shown, the compressed samples provided by compressive sampling unit 1906 are quantized by A/D converter (step 2202). For example, the signal may be 6 bit A/D quantized for each In phase (I) and quadrature components of the signal received. In a typical example, Each complex (I,Q) of the compressed and sampled signal is quantized by A/D converter to a unique position element in a sparse vector which represents the level of the input sample. Each in-phase (I) and quadrature component (Q), are sparse vectors, with one element assigned to a 1, are jointly encoded into one sparse vector (step 2204) by, for example, summing the (I,Q) component vectors. The summation results in the 2 elements of the sparse vector that are non-zero. However by summing the (I,Q) components, the actual assignment of (I,Q) in this vector sum, becomes ambiguous because of the element ordering in the individual vectors. Equation 2 shows the encoding step can be performed using a sparse joint sample encoder unit (step 2204). Hochwald and et al. taught that sparse dictionary matrix =[.sub.1 .sub.2 . . . .sub.N] can be algebraically constructed. In another exemplary embodiment, a random dictionary matrix may be constructed. However, in a preferred embodiment, spherical modulation is used because of its excellent compressive sensing signal reconstruction properties and they generate PM waveforms which improves channel noise immunity.

    (109) Let u.sub.i=[0 . . . u.sub.i,k.sub.i 0 0].sup.T, u.sub.i,k.sub.i=1, i=1,2, k.sub.iZ.sup.+, sparse address vector of A/D sample level

    (110) u.sub.i,I=[0 . . . u.sub.I,58 0 0].sup.T, u.sub.i,Q=[0 . . . u.sub.Q,16 0 0].sup.T for I,Q samples at time (i), for 6-bit quantizer N=64.

    (111) For I,Q samples each sparse vector is stacked to form NM1, where each N1

    (112) [ b 1 b 2 .Math. b N - 1 b M ] M 1 , = [ 1 2 3 .Math. N - 2 N - 1 N ] [ u _ i , I + u _ i , Q ] N 1 , k is M 1 atom of an algebraically constructed dictionary where M << N. Equation 2

    (113) FIG. 22 shows and additional encoding step (Step 2204), which assigns values to the 2 sparse samples (i.e., sparse I and sparse Q components provided by A/D quantizer at step 2202) in the vector sum according to whether the 2 values need to be swapped at the receiver after decoding. The permutation groups define values used to determine when swapping occurs. The properties of permutation groups are used to assign values to the 2 sparse elements in the vector. An examination of Equation 2, and Equation 3 taken together, teaches that (I,Q) ordering is not preserved in the vector sum, but new ordering is an element of a permutation group, with 2-elements in the group. For example, 2 sparse vectors with 2-sparsity elements can be binary coded to an element of the permutation group (step 2208).

    (114) In one examplary embodiment, the permutation table used in premutation group product mapper at Step 2208 may be defined for 2 elements as such:

    (115) Let permutation table be defined for 2 elements

    (116) u.sub.i,I=[0 . . . u.sub.i,58 0 0].sup.T, u.sub.i,Q=[0 . . . u.sub.i,16 0 0].sup.T

    (117) then v.sub.i=u.sub.i,I+u.sub.i,Q=[0 . . . u.sub.i,16 0 . . . 0 u.sub.i,58 . . . 0].sup.T

    (118) .sub.k is an atom of the constructed dictionary and now 2-vectors are selected from the dictionary. For 2-elements, a permuation group multiplication table using standard cycle construction can be defined as

    (119) ( 1 ) ( 12 ) binary encoding of sparse samples ( 1 ) ( 1 ) ( 12 ) ( + 1 , + 1 ) = > No swap ( 12 ) ( 12 ) ( 1 ) ( - 1 , - 1 ) = > Swap Equation 3

    (120) One skilled in the art will recognize this as a swap wherein the vector has (I,Q) swapped in the vector sum, and both sparse samples are multiplied by (1) in the sparse vectors u.sub.i,I,u.sub.i,Q. Consequently, when the vector has unswapped (I,Q) in the vector sum, then vector sum elements u.sub.i,I,u.sub.i,Q are multiplied by (+1) or they are unchanged.

    (121) These encoded samples are then used in joint I, Q sample spherical modulation in accordance with Equation 2 to encode samples sent to chaos waveform spreader 1910. Joint I, Q sample spherical modulation may be performed, for example by a spherical encoder/modulation unit 1908.

    (122) Equation 4 shows the additional encoding step, such as joint I,Q Sample Spherical Modulation 2220 for multiple (I,Q) Spherical Modulation Encoding. The Spherical Modulation Encoding assigns values to 3 sparse samples in the vector sum according to whether the 3 values need to be swapped. The more likely case is 4 sparse, (I,Q) samples but this example indicates how the procedure works for more samples, without burdensome algebra.

    (123) For example, multiple symbols can be encoded for multiple (I,Q) samples when the (I,Q) ordering is not preserved but new ordering is an element of a permutation group, with 6-elements in the permutation group S.sub.n, n=3!=6. In this particular example, the 3 sparse vectors with 3-sparsity elements can be binary coded to an element of the permutation group. Permutation element (13) represents, the swapping of first and third vector elements of u.sub.1,u.sub.3 while u.sub.2 is not swapped. Each vector u.sub.1,u.sub.2,u.sub.3 is multiplied by (+1,1,+1) or vectors become u.sub.1,u.sub.2,u.sub.3 respectively from the binary encoding of the permutation group elements.

    (124) Let permutation table be defined for 3 elements

    (125) u.sub.1=[0 . . . u.sub.i,58 0 0].sup.T, u.sub.2=[0 . . . u.sub.i,23 0 0].sup.T, u.sub.3=[0 . . . u.sub.i,16 0 0].sup.T

    (126) then [u.sub.1+u.sub.2+u.sub.3]=[0 . . . u.sub.i,16 0 0 u.sub.i,23 0 . . . u.sub.1,58 . . . 0].sup.T.sub.MN1,

    (127) .sub.k is M1 atom of algebraically constructed dictionary

    (128) For 3-elements permutation group multiplication table using standard cycle notation can be defined as

    (129) .Math. ( 1 ) ( 12 ) ( 13 ) ( 23 ) ( 123 ) ( 132 ) ( 1 ) ( 1 ) ( 12 ) ( 13 ) ( 23 ) ( 123 ) ( 132 ) ( 12 ) ( 12 ) ( 1 ) ( 132 ) ( 123 ) ( 23 ) ( 13 ) ( 13 ) ( 13 ) ( 123 ) ( 1 ) ( 132 ) ( 12 ) ( 23 ) ( 23 ) ( 23 ) ( 132 ) ( 123 ) ( 1 ) ( 13 ) ( 12 ) ( 123 ) ( 123 ) ( 13 ) ( 23 ) ( 12 ) ( 132 ) ( 1 ) ( 132 ) ( 132 ) ( 23 ) ( 12 ) ( 13 ) ( 1 ) ( 123 ) binary encoding of sparse samples ( + 1 , + 1 , + 1 ) ( + 1 , + 1 , - 1 ) ( + 1 , - 1 , + 1 ) ( + 1 , - 1 , - 1 ) ( - 1 , + 1 , + 1 ) ( - 1 , + 1 , - 1 ) Equation 4

    (130) As shown in FIG. 22, a permutation group such as discussed above, may be mapped by using, for example, a permutation group mapper (Step 2208) onto the sparse PPM I and Q samples which are encoded at Step 2204. The jointly encoded (and permutation group mapped) PPM samples may them be modulated using, for example, spherical modulation (Step 2206) prior to being chaos waveform spreaded by, for example, chaos waveform spreader 1910.

    (131) FIG. 23 shows exemplary spherical demodulation and decoding, such as may be performed by spherical demodulator and decoder 1918 at receiver 1904. As shown, receiver 1904 receives the signal transmitted by transmitter 1902, and chaos waveform despreaded by chaos waveform despreader and rake receiver 1916. Chaos waveform despreader and rake receiver 1916 may provide the chaos waveform despreaded signal to spherical demodulator/decoder 1918 for demodulation and decoding. The output of the spherical demodulator/decoder 1918 is an estimate of the compressed sensing samples from first stage of compression of the clipping error sparse signal at compression sampling circuit 1906.

    (132) The jointly encoded (I,Q) samples from Step 2206 are recovered from the inputs from the chaos waveform despreader and RAKE receiver 1916 by using the same sparse recovery algorithms discussed with respect to FIG. 21 and compressive sampling unit 1906. Any number of sparse signal recovery algorithms can be used to estimate sparse signals as discussed above.

    (133) As discussed in the above example with respect to Equation 3, where (I,Q) sequence ordering is not preserved, new ordering may be an element of a permutation group, with 2-elements in the group. For this example, the 2 sparse I, Q vectors with 2-sparsity elements have been binary coded to an element of the permutation group (e.g., at Steps 2204 and 2208 above). One skilled in the art, will recognize this as a swap wherein the I and Q components of the vector are swapped in the vector sum, and wherein both sparse I and Q samples are multiplied by (1) in the sparse vectors u.sub.i,I,u.sub.i,Q. Likewise when the sparse I and Q samples of the vector are unswapped (I,Q) in the vector sum, then vector sum elements u.sub.i,I,u.sub.i,Q are multiplied by (+1). In some instances, the vector sum elements u.sub.i,I,u.sub.i,Q remain unchanged. For example,

    (134) Let permutation table be defined for 2 elements which has encoded 2-samples as below

    (135) u.sub.i,I=[0 . . . u.sub.i,58 0 0].sup.T, u.sub.i,Q=[0 . . . u.sub.i,16 0 0].sup.T

    (136) then v.sub.i=u.sub.i,I+u.sub.i,Q=[0 . . . u.sub.i,16 0 . . . 0 u.sub.i,58 0].sup.T

    (137) .sub.k is an atom of the constructed dictionary and now 2-vectors are selected from the dictionary. For 2-elements, a permuation group multiplication table using standard cycle construction can be defined as

    (138) ( 1 ) ( 12 ) binary encoding of sparse samples ( 1 ) ( 1 ) ( 12 ) ( + 1 , + 1 ) = > No swap ( 12 ) ( 12 ) ( 1 ) ( - 1 , - 1 ) = > Swap Equation 5

    (139) For decoding one must either swap or not swap the encoded samples to get the correct ordering. For example ab=1, for and encoded vector of (1,1) at the appropriate locations, the a=(12) from the permutation table, then the inverse element is b==(12).

    (140) Equation 5 shows the inverse operators (12) used to swap both sparse I and Q samples are multiplied by (1) in the sparse vector, such that I and Q elements are in correct sequence. Once the swapping is complete, the ordered sparse vectors u.sub.i,I,u.sub.i,Q are sent to a D/A converter (6-bits) for example to generate compressive sensing sample estimates for reconstruction of clipped error sparse signal estimates processed by compressive sampling unit 1906.

    (141) FIG. 24 shows exemplary processing of chaos waveform despreading and RAKE receiver 1916 output .sub.Wk(kT). The chaos waveform despreading and RAKE receiver 1916 processes high-rate samples received from the receiver channel symbols. In accordance with Step 2400, the high-rate samples are correlated over one spherical modulated symbol duration to determine direct path and multipath from the channel. Equation 6 shows the exemplary processing steps which may be carried out by chaos waveform despreading and RAKE receiver 1916. For example, the chaos waveform despreading and RAKE receiver 1916 cross-correlates input (y.sub.n) with a chaos sequence (C) generated according to various embodiments of the invention. The chaos waveform despreading and Rake receiver 1916 generates an output function that is a score function (Step 2400). The score function is searched for peaks that represent both direct and multipath returns to the receiver. Once peaks are found, the direct and multipath returns are ranked by amplitude strength and sent through a maximum ratio combining process. The maximum combining process according to this invention may be performed by a Maximum Ratio Combiner (MRC).

    (142) In a typical example of the invention, cross-correlation of input (y.sub.n) with a chaos sequence (C) may occur according to the following:

    (143) w n = .Math. p = 1 P y p + n C p , P
    is length of chaos waveform (C), w.sub.n is the resulting correlation surface There can be multiple peaks on the correlation due to multipath returns [w(n.sub.1T) w(n.sub.2T) w(n.sub.2T) . . . w(n.sub.m)]=FindPeaks (w.sub.n), determines all the potential peaks Create Sorted list of peaks from highest to lowest are
    [w(j.sub.iT)>w(j.sub.2T)>w(j.sub.3T) . . . >w(j.sub.mT)]=Sort[[w(n.sub.1T)w(n.sub.2T)w(n.sub.2T) . . . w(n.sub.m)]] Equation 6

    (144) As shown in FIG. 24, the maximum ratio combiner process (Step 2410) may coherently combine the first peak (i.e., maximum strength) from sorted list (Step 2400) and sends result for spherical demodulation and decoding, such as, by spherical demodulator/decoder 1918. Equation 7 shows the correlation surface peaks being summed by a MRC process (Step 2410) to form input samples to the spherical decoder 1918.

    (145) u ^ wK = .Math. i = 1 RMAX w ( j i T ) Equation 7

    (146) In some examples, it may be desired to provide an additional level of compression of the compressed samples received from compressive sampling unit 1906. By adding the additional level of compression, more bandwidth becomes available for chaos spreading. As shown by FIG. 25, the additional level of compression may be performed using Gabor Transform Multiple Symbol Encoding, performed by a Gabor Transform Multiple Symbol Encoder, as a block encoder (Step 2510). In one exemplary embodiment, the Gabor Transform Multiple Symbol Encoder can be used as a replacement for the spherical modulator 1908 to provide samples for chaos waveform spreading (Step 2520) by for example, chaos waveform spreader 1910. In yet another embodiment, block encoding is also preferred to symbol-by-symbol encoding for SNR enhancement at the receiver 1904.

    (147) It should be noted that, the compressed samples may undergo an A/D quantization process (Step 2500) similar to the A/D quantizer discussed with respect to Step 2202 of FIG. 22. For example, the A/D quantizer Step 2500 may include a 6-bit A/D quantizer for both (I, Q) complex samples, which generates sparse vectors of (I, Q) samples.

    (148) In contrast to symbol-by-symbol encoding/decoding for spherical modulation, the Gabor Transform Encoder described in Step 2510 also reduces time computational complexity at the receiver by exploiting sparse recovery algorithms. The sparse (I,Q) vectors are encoded to time/frequency elements by the Gabor transform. Each of these 2-D matrices are multiplied by a sensing vector (e.g. Alltop sequence) to map the matrix to a 1-D vector representing the time/frequency sparse element.

    (149) For example for (N=64)-(I,Q) samples of the sparse vector, the Gabor transform dictionary has at least (N.sup.2=4096)-elements. For reconstruction of a K-sparse signal, K<({square root over (N)}+1) or K<({square root over (64)}+1)=4.5 sparse quantized samples can be reconstructed with N=64 compressed samples. Therefore K=4 sparse quantized samples, can be jointly encoded using the Gabor Transform encoder. Joint encoding enhances noise immunity at the receiver. The overall sample rate of the incoming samples is decreased by a factor of 4, which means there is more chaos waveform spreading per symbol.

    (150) Gabor frames are used to model entries in range-Doppler maps (RDM) in radar or symbol-timing/frequency offset for wireless communications receivers. For example, let (H) be unknown matrix (e.g. RDM) of size (M, N) and assume that it can expanded in an orthonormal set of basis functions (H.sub.i), matrices. Then, when M=N, the Gabor frame for matrix (H) is shown in Equation 3 where each matrix is (NN) in size. In such case, the matrices are also orthonormal and given by:

    (151) H = .Math. i = 0 NM - 1 s i H i , H i * H j = i , j I Equation 8

    (152) Consequently, each matrix is formed as follows in Equation 8 where each matrix is a time/frequency matrix.

    (153) H i = P imodN T [ i / N ] T = [ 0 0 .Math. 1 1 0 .Math. 0 .Math. 0 0 .Math. 1 0 ] P = [ N 0 0 .Math. 0 0 N 1 .Math. 0 .Math. 0 0 0 .Math. 0 N N - 1 ] , N i = e j * 2 i / N Equation 9
    Compressed measurements are formed as (y) that is size N column vector, N.sup.2 is size of sparse vector (s)

    (154) y = Hf = .Math. i = 0 N 2 - 1 s i H i f = .Math. i = 0 N 2 - 1 s i i = s , s is sparse signal vector , ( e . g . an RDM cell ) i = H i f C N , = [ 0 1 N 2 - 1 ] C N N 2 Equation 10
    where (f) is a probe signal (e.g. Alltop) sequence and .sub.i are the atoms of a compressed sensing measurement dictionary. Usually the vector (f) is the randomization function which generates a dictionary with underlying coherency properties (e.g. Reciprocal Isometry Property (RIP)) but (f) can also inject knowledge about signal-of-interest (SOI) into the over-determined bases. Equation 11 shows the definition of the Alltop sequence according to this invention.

    (155) 0 f A = { f n } n = 0 N - 1 f n = 1 N e j * 2 n 3 / N , one ( i . e . cubic phase ) more general form .Math. ( f ( n ) = N - 1 / 2 N n 3 + n ) Equation 11

    (156) After A/D quantizing to unique position element in a sparse vector for both real and complex components (I,Q) at Step 2500, the Gabor Transform Encoder used at Step 2510 encodes each CS sample. Equation 1212 is an example encoding method that may be used for each sparse (I,Q).

    (157) As shown, Each (N1) time ordered sparse vector is stacked to form a Gabor domain (s) (N.sup.21). Let u.sub.i=[0 . . . u.sub.i,k.sub.i 0 0].sup.T, u.sub.i,k.sub.i=1, i=1, . . . , M, k.sub.iZ.sup.+, sparse address vector of A/D sample level

    (158) For K=4<<N.sup.2, CS samples of each sparse vector are stacked to form N.sup.21 vector. Note that since N must be a prime number for dictionary construction that is a property of the Alltop sequence, then N=67. Each N1 sample vector has an extra 0 at end. For this K, vectors are stacked in the Gabor Transform sparse vector.

    (159) [ b 1 b 2 .Math. b N - 1 b N ] N 1 = [ 1 2 3 .Math. N 2 - 2 N 2 - 1 N 2 ] [ u _ 1 u _ 2 .Math. u _ K 0 _ ] N 2 1 where k is an N 1 atom of the Gabor dictionary . Equation 12
    The time ordered sparse vectors u.sub.i results in time ordered selection of Gabor dictionary atoms. Since K<N, there is no modulo-N arithmetic to be concerned with. The compressed sample vector b from Gabor Transform encoder (Step 2510) is sent to chaos waveform spreading (Step 2520) performed for example by chaos waveform spreader 1910. After spreading, the samples are sent by the transmitter.

    (160) FIG. 26 shows the method for using the Gabor Transform Decoder with the present invention. As shown, incoming samples from the chaos waveform despreader and Rake receiver 1916 are processed using sparse recovery algorithms (Step 2600) to recover the sparse Gabor Transform vector. Next sparse joint decoding may be performed (Step 2610) by, for example, sparse joint decoder for extracting time ordered samples of u.sub.i, i=1, . . . , K (i.e., time ordered sparse vectors). AD/A converter may then take the time ordered sparse vectors and convert the vectors to estimates of compressed samples (Step 2620) for reconstruction to sparse clipped error signals by for example sparse reconstruction unit 1920.

    (161) FIG. 27 shows the method performed by a spherical modulation soft block symbol decoder that may be used with the present invention. The method may include receiving chaos waveform despreaded signals from, for example, chaos waveform despreader with a Rake receiver 1916. The despreaded chaos waveform signals may undergo a joint (I,Q) symbol-by-symbol recovery (Step 2700). Symbol-by-symbol recovery using sparse signal recovery requires thresholding and/or peak finding of the sparse vector to determine peaks. Block decoding of ALL symbols avoids thresholding/peak finding for each symbol by using the dynamic programming algorithm (DPA) (Step 2710) performed by for example a dynamic program algorithm block decoder. Furthermore, an SNR gain is achieved by coherently integrating over the symbol-by-symbol soft-decisions made by sparse recovery.

    (162) FIG. 27 shows a Spherical Modulation Soft Block Symbol. Joint sample recovery (Step 2700) receives samples from the chaos dispreading and RAKE receiver (Step 1940). The sparse vectors of samples are recovered using sparse recovery algorithms (Step 2700). However, these sparse vectors are not decoded but are maintained as soft measurements that are processed by the DPA (Step 2710) over a block of soft decisions for multiple symbols.

    (163) FIG. 28 is an example of the trellis used for DPA processing that may be used with Step 2710 and in accordance with the present invention. As shown, each complex sample of the sparse vector symbols, S(n) takes a position in the trellis at each time epoch. Each trellis node (i.e., S(1)-S(n)) represents a location in each sparse vector (u.sub.i,I+u.sub.i,Q).

    (164) In accordance with the DPA processing, each recovered sparse vector has 64 positions which could include noise positions along with the 2 clipped error samples of interest, where one sequence of sparse vectors through the trellis is designated as path-1 and other sequence designated as path-2. The sequence of symbols then form a path through each trellis node (trellis node 1-250) for each path. Each trellis node is fully connected to all other nodes in the trellis. The DPA uses forward and backward processing to determine optimal paths through the trellis for each of the 64 positions of each recovered sparse vector. A measurement for each set of trellis nodes at each time is the sparse vector reconstructed using sparse recovery, such as is performed by sparse reconstruction unit 1920, during the same time epoch. A cost metric is computed at each trellis node to find the best match of measurement to hypothesis (+1,1,+2) since those are 2 possibilities for a sparse symbol for 2 signals present. Otherwise each vector is 0 or 2 at locations when no signals are present or both signals are the same for the vector sum respectively.

    (165) Equation 13 shows exemplary processing that may be used for computing a cost metric for each trellis node in accordance with the present invention, wherein each node provides a sparse recovered vector sample at an identified point in time. The cost metric is approximated using the Turbo code approximation for reducing computational complexity. This metric identifies the most likely pairing of the measurement at a node with the most likely possible hypothesis. The term .sup.2 is the expected noise variance associated with each measurement. The cost metric is a standard normal probability density function to account whereby the signal power is the mean for the normal density function. Each trellis node is a tuple that has both metric score and hypothesis that resulted in that score.

    (166) For each trellis node (i) at time t.sub.k compute cost metric

    (167) d i ( t k ) = 1 2 max [ exp ( 1 2 2 ( .Math. z i ( t k ) - 1 .Math. 2 ) ) , exp ( 1 2 2 ( .Math. z i ( t k ) + 1 .Math. 2 ) ) exp ( 1 2 2 ( .Math. z i ( t k ) - 2 .Math. 2 ) ) ] , 1 2 exp [ 1 2 2 min { ( .Math. z i ( t k ) - 1 .Math. 2 ) , ( .Math. z i ( t k ) + 1 .Math. 2 ) , ( .Math. z i ( t k ) - 2 .Math. 2 ) } ] , Turbo - code approximation Let a node identified as a set with 2 parameters .Math. N i ( t k ) = { d i ( t k ) , M i } where M i is either ( + 1 , - 1 , + 2 ) depending which hypothesis gives min . The ( + 2 ) hypothesis occurs when both paths have the same trellis node in common . Equation 13
    Equation 14 shows epoch how a cumulative score is computed for each node at time based on all possible previous nodes for the next time. As shown, a cumulative path metric to each current node is stored to track of optimal path to the current node. Equation 14 also shows the computations involved in calculating cumulative score and cumulative path from the previous node that yields the top score. A directed edge is formed for each node at time (k) and the previous node that resulted in the maximum score with index (J).

    (168) C i ( t k ) = max j [ C j ( t k - 1 ) + d i ( t k ) ] , j = 1 , .Math. , , 64 , for each i = 1 , .Math. , 64 , C j ( t - 1 ) = 0 Define a cumulative path to current node as with N J ( t k - 1 ) is node that gives max score P i ( t k ) = EDGE ( N i ( t k ) , N J ( t k - 1 ) ) P J ( t k - 1 ) , with P i ( t 1 ) = , the null edge at the first time epoch Also the N J ( t k - 1 ) is expurgated from the trellis , so that the search for second path at ( t k ) does not include a node that is already part of another path . Equation 14
    At each time epoch, each SOFT measurement is used to update cumulative scores and the cumulative paths, for each path. At the last time epoch, there will be a cumulative score and a cumulative path defined for the last set of nodes in the forward path.

    (169) The backward path processing when the paths have no path nodes in common is as follows: Given all cumulative scores at the last set of nodes in the trellis, they are ranked in descending order. The maximum scores define the end nodes for all paths through the trellis, A backward trace is formed to define the optimal paths to the nodes with top scoreswhich has been defined during forward processingand the hypothesis associated with each path node to get to the final node. Equation 15 shows the explicit steps mathematically.
    C.sub.m(t.sub.N)=maximum(C.sub.i(t.sub.N)), for each i=1, . . . ,64 Equation 15 where (m) is the value of (i) where the cumulative score at the end node (N) is the largest. This node (m) is used to backtrack through the trellis to determine an optimal path. The sequence of (+1,1,+2) along each optimal backward path defines an optimal sequence of symbols through the trellis. To get the second optimal path through the trellis, one must first expurgate the trellis (i.e. delete the nodes of the optimal sequence of symbols from the previous backtracking step), so that these nodes are not duplicated when searching through the trellis in the forward direction to determine path-2 through the trellis. There is no need to re-compute the node metrics d.sub.i(t.sub.k) for the second forward pass through the trellis since the scores have already been computed. However new cumulative scores C.sub.i(t.sub.N) are computed based on this new trellis that doesn't include the nodes from the previous forward processing step. After forward processing is complete then Equation 5 is performed to find the largest cumulative score at the end node (N) from which the backtracking step is repeated with new set of optimal nodes. By using this 2-pass process, the 2 optimal nodes are determined for each time epoch (t.sub.i), since there are 2-nodes at each time epoch for each path. There is only NO thresholding step at the LAST set of nodes in the trellis to determine top 2 paths.

    (170) The backward path processing when the paths have trellis path nodes in common is essentially the same 2-pass processing as for the case when there are separate paths through the trellis, with the exception that any node with a hypothesis score (+2) is NOT expurgated from the trellis for the after the backtracking is complete from the first-path. Now the same steps are performed for the second forward pass and Equation 5 determines the end node for the second backtracking step.

    (171) After backward path processing is complete, the 2 paths provide hard decisions for the sequence of sparse symbols. The jointly decoded sparse signals may be re-ordered in time Step 2720, Permutation group mapped Step 2730, using for example, a permutation group element mapper, in similar manner as described with respect to Steps 2310 and 2320 respectively. These sparse signals are converted to compressive sensing samples with D/A converter as shown in (Step 2620) for further processing by Compressive sensing recovery algorithms, by for example, sparse reconstruction unit 1920.

    (172) It should be appreciated by one skilled in the art, that the foregoing description has been directed to specific and exemplary embodiments of this invention. It will be apparent; however, that other variations and modifications may be made to the described embodiments, with the attainment of some or all their advantages. Therefore, it is the object of the appended claims to cover all such variations and modifications as come within the true spirit and scope of the invention.