FRAME STRUCTURE FOR AN ADAPTIVE MODULATION WIRELESS COMMUNICATION SYSTEM

Abstract

A method of simplifying the encoding of a predetermined number of bits of data into frames including adding error coding bits so that a ratio of the frame length times the baud rate of the frame times he bit packing ratio of the data divided the total bits of data is always an integer. The method may also convolutionally encode the bits of data so that the same equation is also always an integer.

Claims

1. (canceled)

2. A method of operating a base station in an on-demand wireless cellular communication system for assigning portions of at least two downlink slots of a plurality of downlink slots of a frame to at least two mobile cellular telephones for transmitting modulated data according to a modulation scheme that may vary for each of the at least two mobile cellular telephones, the method comprising: (a) determining a characteristic of the modulation scheme employed in generating the modulated data for the at least two mobile cellular telephones; and (b) assigning the portions of the at least two slots to the at least two mobile cellular telephones based on the characteristic of the modulation scheme employed in generating the modulated data for the at least two mobile cellular telephones.

3. The method according to claim 2, wherein at (b) the portions of the at least two downlink slots are assigned to the at least two mobile cellular telephones from the least complex modulation scheme to the most complex modulation scheme.

4. The method according to claim 2 wherein (b) includes: (i) ordering the at least two mobile cellular telephones as a function of the complexity of the modulation scheme employed to generate the modulated data for each of the at least two mobile cellular telephones; and (ii) assigning the portions of the at least two slots to the at least two mobile cellular telephones based on the order of the at least two mobile cellular telephones.

5. The method according to claim 4, wherein (ii) of claim 3 includes assigning the portions of the at least two slots to the at least two mobile cellular telephones from the least complex modulation scheme to the most complex modulation scheme.

6. The method according to claim 5, wherein the frame is a time division duplex frame.

7. The method according to claim 2, wherein the bit per symbol rate of the modulation scheme employed to generate the modulated data may vary for each of the at least two mobile cellular telephones, wherein the determined characteristic of the modulation scheme employed in generating the modulated data for the at least two mobile cellular telephones determined in (a) is the bit per symbol rate.

8. The method according to claim 2, wherein (b) includes: (i) ordering the at least two mobile cellular telephones as a function of the bit per symbol rate of the modulation scheme employed to generate the modulated data for each of the at least two mobile cellular telephones; and (ii) assigning the portion of the at least two slots to the at least two mobile cellular telephones based on the order of the at least two mobile cellular telephones.

9. The method according to claim 8, wherein (ii) includes assigning the portion of the at least two slots to the at least two mobile cellular telephones from the lowest bit per symbol rate modulation scheme to the highest bit per symbol rate modulation scheme.

10. A base station for operation in an on-demand wireless cellular communication system, the base station comprising: a receiver operable to receive uplink transmissions from at least two mobile cellular telephones; a transmitter operable to transmit modulated data in at least two downlink slots of a plurality of downlink slots of a frame to the at least two mobile cellular telephones, the transmitter including a frame constructor operable to determine a characteristic of a modulation scheme employed in generating the modulated data for the at least two mobile cellular telephones, and assign the portions of the at least two slots to the at least two mobile cellular telephones based on the characteristic of the modulation scheme employed in generating the modulated data for the at least two mobile cellular telephones.

11. The base station according to claim 10 wherein the frame constructor assigns the portions of the at least two slots to the at least two mobile cellular telephones from the least complex modulation scheme to the most complex modulation scheme.

12. The base station according to claim 10, wherein the frame is a time division duplex frame.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0017] FIG. 1 is a diagram of an exemplary cell configuration with a base station and several CPEs associated with the cell.

[0018] FIG. 2 is a diagram an exemplary time division duplex (TDD) frame in accordance with the present invention.

[0019] FIG. 3 is a flowchart of an exemplary process of assigning time slots of a TDD frame in accordance with the present invention.

[0020] FIG. 4 is a flowchart of an exemplary process of simplifying the configuration of data to be inserted into a TDD frame in accordance with the present invention.

[0021] FIG. 5 is a block diagram of an exemplary transmitter for use with the present invention.

[0022] FIG. 6 is a block diagram of an exemplary receiver for use with the present invention.

[0023] FIG. 7 is a block diagram of a prior art Finite Impulse Response (FIR) filter suitable for use with the present invention.

[0024] FIGS. 8A to 8F are diagrams illustrating a method of the invention that changes weights of a FIR filter as new symbols propagate through the filter.

[0025] Like reference numbers and designations in the various drawings indicate like elements.

DETAILED DESCRIPTION OF THE INVENTION

[0026] Throughout this description, the preferred embodiment and examples shown should be considered as exemplars, rather than as limitations on the present invention.

[0027] The present invention includes an improved frame structure and a process of generating a frame structure for use in wireless communication systems employing adaptive modulation. Adaptive modulation includes varying the bit per symbol rate modulation scheme or modulation complexity of signals transmitted between a CPE and a base station as a function of channel interference of the signals or implementation or modem complexity of the CPE. FIG. 1 is a diagram of an exemplary cell 10 that includes a base station 20 located centrally in the cell 10 and a plurality of CPEs 30, 32, 34, 36, 38 associated with the base station. FIG. 1 does not shown buildings or other physical obstructions (such as trees or hills, for example), that may cause channel interference between signals of the CPEs.

[0028] As described above, the maximum bit per symbol rate modulation scheme or technique or most complex modulation scheme selected for use in the cell 10 is normally determined as a function of the channel interference between CPEs and the implementation or modem complexity of the CPEs. As also described above, the selection of a single maximum bit per symbol rate modulation technique based on the lowest bit per symbol rate modulation scheme supported by all CPEs may not optimize bandwidth utilization within the cell 10. In particular, lower channel interference between some CPEs (such as units 38, 30 for example), may permit the use of a higher bit modulation technique or more complex modulation scheme that has an error level below the maximum desirable error level. Adaptive bit-rate modulation or variable bit-rate modulation between different CPEs, however, usually requires complex transmitters and receivers in the CPEs where the CPEs may already have limited implementation or modem complexity.

[0029] As noted above, the frame structure is divided into a plurality of downlink and uplink slots. Each downlink time slots may be used to store data to be received by a number of users where a user identifies their data by an address or other label. Uplink time slots are commonly assigned to individual users for transmission of data from the user to another user or system via the base station. To maximize bandwidth utilization and minimize modulator complexity in the base station and associates CPEs, the present invention simplifies the configuration of data to inserted into the time slots. Briefly, data blocks are ideally parsed into an integer number of time slots. This process is described in detail below with reference to FIG. 4. Second, the present invention, orders or sorts the placement of data in the downlink and uplink time slots are a function of modulation complexity or bit per symbol rate modulation scheme employed to generate the data to be placed in the time slots. As described below with reference to FIG. 3, this technique reduces the complexity of CPE modulators and the number of modulation scheme transitions in a frame.

[0030] FIG. 2 is diagram of an exemplary frame structure to be employed in a cell that enables adaptive bit per symbol rate modulation schemes to be employed in a frame structure without increasing the complexity of receivers and transmitters of CPEs associated with the cell and reducing the number of modulation scheme transitions within each frame. As shown in FIG. 2, the frame 80 includes a plurality of time slots. In this example there are ten time slots where the first five time slots contain downlink data 82 (from the base station 10), and the remaining five time slots contain uplink data 84 (to the base station 10 from a CPE). In this example, the downlink slots have a modulation bit per symbol rate of DM.sub.1, DM.sub.2, DM.sub.3, and DM.sub.4 where the four downlink time slots are assigned to at least four CPEs where the CPEs will retrieve data located in these slots based on their respective assignment. It is noted that many CPEs may be assigned to any one downlink time slot where each CPE retrieves its data from such a slot based on an address or identifier. Consequently, a CPE may only retrieve data from only a portion of a downlink time slot.

[0031] In addition, the uplink slots have a modulation bit per symbol rate of UM.sub.1, UM.sub.2, UM.sub.3, and UM.sub.4 where the four uplink time slots are commonly assigned to four CPEs where the CPEs will insert data in these slots based on their respective assignment. It is noted that in some embodiments a CPE may be assigned more than one uplink slot. Further, downlink control information may be located at the start of the downlink time slots and an unreserved time slot may be located at the beginning of the uplink time slots. It is obviously desirable that any CPE associated a cell be able to retrieve data located in the downlink control information time slot regardless of the CPE's location within the cell. In addition, each CPE should be able to insert data into the unreserved uplink time slot.

[0032] As described above, in an adaptive bit per symbol rate modulation system the modulation scheme may vary for each CPE and thus for each downlink and uplink time slot. In order to minimize the complexity of CPEs and base stations employed in such a system and reduce the number modulation scheme transitions within a frame, the present invention requires that DM.sub.1DM.sub.2DM.sub.3DM.sub.4 and UM.sub.1UM.sub.2UM.sub.3UM.sub.4. Thus, ideally, the data in the time slots is arranged from the least complex modulation scheme to the most complex modulation scheme. As noted, this technique reduces the number of modulation transitions, which may simplify the implementation of a base station using this frame structure 80. Note this also enables the base station and CPEs to train on the least complex data, which may lower error rates.

[0033] Further, ideally the downlink control information is ideally encoded using the least complex modulation scheme of the system and the information placed in the unreserved uplink time slot is also encoded using the least complex modulation scheme of the system. This ensures that every CPE associated with the cell will be able to receive or encode information within desirable error levels. Ideally, the control information indicates where the modulation transitions occur within the frame. An exemplary process 90 of assigning time slots of frame 80 as shown in FIG. 2 is presented with reference to FIG. 3.

[0034] As shown in FIG. 3, the first step, 92 of the process 90 includes determining which CPEs will receive at least one time slot in the next frame. In duplex systems, as described above, a CPE receiving data in a downlink time slot may also transmit data in an uplink time slot. In other systems, such as point to multi-point or multi-cast systems, there may be more downlink time slots than uplink time slots. Then (in step 94) the most complex modulation scheme or maximum bit per symbol rate of the modulation scheme employed by the CPE is determined for each CPE. As stated above, the most complex modulation scheme or maximum bit per symbol rate modulation scheme may be determined as a function of channel interference of signals of a CPE and the maximum desirable error level and the implementation or modem complexity of the CPE.

[0035] In a preferred embodiment, Binary Phase Shift Keying (BPSK) modulation may be selected for the least complex modulation scheme. In BPSK, the bit per symbol rate, B.sub.I of the modulation scheme is one, i.e., each symbol represents one bit. B.sub.I could also be called the modulation scheme efficiency, i.e., how efficient the scheme encodes data. A Quadrature Amplitude Modulation (QAM) of four may be used for an intermediate modulation scheme. In QAM 4, the bit per symbol rate, B.sub.I of the modulation scheme is two, i.e., each symbol represents two bits. Higher quadrature amplitude modulations may be used for more complex modulation schemes, e.g., QAM 64 where the bit per symbol rate, B.sub.I of the modulation scheme is six, i.e., each symbol represents six bits. The modulation complexity or bit per symbol rate modulation scheme may be modified from frame to frame or remain constant for a plurality of frames for a particular CPE. Further, a CPE may select or indicate a desired modulation complexity or scheme.

[0036] Upon determination of the modulation complexity or bit per symbol rate modulation scheme to be used to encode data for each of the CPEs, in step 96 the CPEs are sorted in ascending order based on the selected modulation complexity or bit per symbol rate modulation scheme, i.e., from the lowest bit per symbol rate modulation scheme to the highest bit per symbol rate modulation scheme or least complex modulation scheme to the most complex modulation scheme. Finally, the time slots of a frame are allocated or assigned to the CPEs in their sorted order from the lowest bit per symbol rate modulation scheme to the highest bit per symbol rate modulation scheme or from the least complex modulation scheme to the most complex modulation scheme. As noted above, frames are constructed using this process to reduce the complexity of base stations and CPEs that insert or retrieve data therefrom. It is noted that even though modulation schemes may vary from CPE to CPE, the number of symbols to be transmitted in bursts is usually fixed to a predetermined number nS for all CPEs regardless of their modulation scheme.

[0037] It is desirable to simplify the configuration of time slots given fixed bursts of a group of symbols nS and variable modulation schemes. It is noted that the modulation of L bits generates a fixed number of symbols S where S=(L/B.sub.1) and B.sub.1 is the bits per symbol rate of the modulation scheme. To simplify time slot usage and bandwidth management, (L/B.sub.1) or S is ideally an integer multiple of length of the time slot T.sub.s times the baud rate R of the frame. Thus, ideally L bits fit into an integer number of time slots T.sub.s based on the modulation scheme. Note each frame has a fixed number of time slots where the length of the frame (and thus the number of time slots) is determined a function of a maximum desirable delay T.sub.D between signal transmissions and the baud rate R (symbols transmitted per second) of the system. Accordingly for each frame the number of symbols transmitted is equal to T.sub.D*R. It is desirable that the number of symbols nS or (L/B.sub.1) is an integer multiple of the number of symbols transmitted per frame. Thus, it is desirable that the ratio (T.sub.D*R)/(L/B.sub.1) is an integer. When the ratio (T.sub.D*R)/(L/B.sub.1) is an integer then a fixed number of bursts of nS symbols may be transmitted in each frame. This may simplify frame usage and bandwidth management.

[0038] In most systems, the L bits of data represent an encoded signal that includes overhead or Forward Error Correction (FEC) information where only L.sub.D of the L bits are pure data to be transmitted to a unit or base station. In these systems the number of data bits L.sub.D to be transmitted in a burst may be fixed, e.g., 256, 512, or 1024 bits. The FEC information commonly includes convolutional encoding bits and block codes including error correction encoding bits such as Reed-Solomon (RS(n,k)) data. In other embodiments, the convolutionally encoded data may also be interleaved prior to error encoding. Given that T.sub.D, R, and S are fixed due to system constraints and B.sub.1 is selected as a function of channel interference and modem or implementation complexity, L is ideally modified to simplify the time slot configuration or the bandwidth management of a frame. As noted, L.sub.D may also be fixed in a system. In such a system L would be determined for each possible modulation scheme of the system. FIG. 4 is a flowchart of a preferred process 60 of configuring or determining L based on T.sub.D, R, and B.sub.1 for the transmission of data by a unit or a base station so that frame usage is simplified.

[0039] As shown in FIG. 4, the first step, 62 of the process 60 determines the maximum allowable delay T.sub.D of the system. As noted above, the delay T.sub.D is set equal to the largest acceptable delay between transmissions of signals between CPEs or units and the base station. In step 64 the maximum bit per symbol rate modulation scheme or most complex modulation scheme that may be employed for the transmission of the L.sub.D bits is determined or selected (the process of which was described above.) Then in step 66 a convolution ratio (x/y) is selected for the L.sub.D data bits. In some embodiments no convolutional encoding is employed. In such embodiments, the ratio of (x/y) is set to 1. The convolutional ratio (x/y) is one of the parameters that may be modified to change the number of bits required to encode the L.sub.D bits of data. At step 68, the other variable parameter, the error encoding level is selected. A block code is used to reduce the Block Error Rate (BER) of the L.sub.D bits of data to a desirable level. In a preferred embodiment, a Reed-Solomon (RS) block code is used. The number of bits L required to encode the L.sub.D bits of data is thus set by the selection of the convolutional ratio (x/y) and the error code level.

[0040] At step 72, the value of the ratio Z of (T.sub.D*R)/(L/B.sub.I) is determined. The baud rate R is fixed, the delay T.sub.D was determined at step 62, B.sub.1 is determined at step 64, and L is determined as function of the parameters selected at steps 66 and 68. When it is determined at step 74 that the ratio Z is not integer, a different convolutional ratio (at step 66) or the error code level (at step 68) may be selected. In a preferred embodiment, the selection of the convolutional ratio and the error code level is varied as a function of the fractional remainder of the ratio Z, i.e., a convergence algorithm may be employed. As noted above, in some embodiments the convolution ratio is fixed to 1. In such embodiments, only the error code or block code level is modified. In order to ensure that the ratio Z is an integer, the number of bits used to generate the block code of data may be greater than necessary to meet the minimum BER. When at step 74, the ratio Z is determined to be an integer, then the process is complete and the block of L bits is optimized or simplified for the modulation scheme or bit per symbol rate B.sub.1.

[0041] A transmitter 40 and receiver 50 that may employed to transmit and receive frames of data in accordance with the present invention is presented with reference to FIGS. 5 and 6. FIG. 5 is a block diagram of an exemplary transmitter 40. As shown in this FIGURE, the transmitter 40 includes a convolutional encoder 42, block encoder 44, M-ary Modulator 46, frame constructor 48, and up-converter 49. The transmitter 40 receives the L.sub.D bits of data and encodes the data to generate L bits of data, packs the L bits of data into a frame and upconverts the frame of data to a transmission frequency. The convolutional encoder 42 and block coder 44 supply the FEC data that converts the L.sub.D bits of data into L bits of data. In particular, the convolutional encoder 42 uses the selected ratio (x/y) to encode the L.sub.D bits of data. The block coder uses the selected code level to encode the convoluted data to produce the encoded L bits of data to be transmitted to a base station or unit.

[0042] Then, the M-ary modulator converts the L bits of data into the nS symbols based on the selected bit per symbol rate B.sub.1. Due to the selection of the convolution ratio and error code level, the nS symbols can be inserted into an integer number of times slots of a frame. The frame constructor 48 ideally inserts the nS symbols into time slots of a frame based on the process presented with reference to FIG. 3 above, i.e., in order of the modulation scheme (from least complex to the most complex modulation scheme). Up-converter 49 frequency shifts the packed frame of data to a frequency suitable for transmission between a CPE or unit and base station based on techniques known to those of skill in the art.

[0043] The receiver 50 shown in FIG. 6 converts the frequency shifted frame of data back into groups of L.sub.D bits of data. As shown in FIG. 6, the receiver 50 includes a down-converter 59, frame deconstructor 58, M-ary demodulator 56, block decoder 54, and convolutional decoder 52. The down-converter 59 frequency shifts the received signal back to baseband using techniques known to those of skill in the art. The frame deconstructor separates the frame into groups of nS symbols for processing by the remaining components of the receiver 50. When the receiver 50 is part of a subscriber unit, the frame deconstructor selects one of the groups of nS symbols where the data is directed to the subscriber unit. Block decoder 54 decodes the nS symbols using techniques known to those of skill in the art. Then, the convolutional decoder decodes the data to produce L.sub.D bits of data.

[0044] The techniques and systems presented above may be modified while still falling within the scope of the appended claims. For example, symbol shaping may also be employed in a preferred embodiment to avoid spectrum spillage due to possible abrupt changes in modulation schemes in a frame as described above. Symbol shaping is commonly accomplished by filtering the nS symbols via a Finite Impulse Response (FIR) filter where an exemplary prior art FIR filter 60 is shown in FIG. 7. As shown in FIG. 7, the FIR filter 60 includes k multipliers and a summation node 66. Symbols S are received sequentially and stored in filter taps T0 to Tk 62. Each multiplier 64 has a tap weight W0 to Wk and tap T0 to Tk associated with a symbol stored in the taps 62. As can be seen from FIG. 7, the FIR filter 60 generates an output, y having the form

[00001]$y=\underset{i=0}{\overset{k}{.Math.}}.Math.\mathrm{Wi}*\mathrm{Ti}.$

[0045] It is noted that different modulation schemes, such as different QAM schemes (QAM-4, QAM-16, QAM-64) employ different alphabets to represent the x symbols of the scheme. For example, QAM-4 has four different symbols, QAM-16 has sixteen different symbols and QAM-64 has sixty-four different symbols. In addition, different modulation schemes may have different gains that are applied to the symbols for transmission due to varying back-off requirements. In prior art variable modulation systems when the modulation scheme changes, the memory of the FIR filter is not normally reset while the weights W0 to Wk are instantly changed to weights optimized for the modulation scheme or symbols of the scheme to prevent spectrum spillage.

[0046] This solution is not ideal, however, because the weights are then not optimized for the symbols in the memory (taps 62) of the filter that correspond to the previous modulation scheme or rate. One solution is to employ one set of weights for all modulation schemes. This solution is also not ideal, however, since the FIR filter is then not optimized for the alphabet of symbols for each modulation scheme. To prevent spectrum spillage and optimize the FIR filter 60, the present invention changes the filter taps sequentially with each new symbol from the new modulation scheme as shown in FIGS. 8A to 8F. In particular, the weight that corresponds to the first new symbol of a new modulation scheme is modified as the first new symbol propagates through the filter 60. In FIG. 8A, the filter weights W0 to Wk are optimized for the modulation scheme of the symbols currently being processed by the FIR filter 60. In FIG. 8B, the first symbol of a new modulation scheme is received in T0 of the FIR filter 60. At this point as shown in FIG. 8B, the present invention replaces the filter weight, W0 associated with T0 with a new filter weight W0 where W0 is optimized for the modulation scheme of the first new symbol stored in T0. Then, when the next symbol from the new modulation scheme is received and stored in T0 and the first new symbol is shifted to T1, the present invention replaces the filter weight, W1 associated with T1 with a new filter weight W1 where W1 is also optimized for the new modulation scheme as shown in FIG. 8C. This process is repeated as shown in FIGS. 8D to 8F until all the symbols stored in the taps 62 the FIR filter 60 belong to the new modulation scheme and all the filter weights W0 to Wk are associated with or optimized for the new modulation scheme. This technique reduces spectrum spillage while optimizing the weights employed in the FIR filter 60 to shape the symbols or varying modulation schemes.

[0047] Accordingly, it is to be understood that the invention is not to be limited by the specific illustrated embodiment, but only by the scope of the appended claims.