Methods of beam-indexed spatial modulation

Abstract

This invention provides methods of beam-indexed spatial modulation (BISM) for multiple-input and multiple-output (MIMO) technology. It does not only enhance the efficiency of MIMO using, but also address the compatibility problems in Spatial Modulation (SM-MIMO) and Orthogonal Frequency Division Multiplexing Index Modulation (OFDM-IM). Furthermore, the BISM improves the speed limitation problem and spectral efficiency issue in the current spatial modulation architectures.

Claims

1. A first method of beam-indexed spatial modulation (BISM) for a multiple-input and multiple output (MIMO) module having a transmitter end and a receiver end, comprising: step a) confirming that the MIMO module has a plurality of beams, the number of the plurality of beams is N.sub.B, each of the plurality of beams is an individual channel constituted by one of antennas at the transmitter end and one of antennas at the receiver end; step b) providing an index to the N.sub.B beams through a N.sub.TN.sub.R channel matrix H, wherein N.sub.T are an integer number of a plurality of transmitting antennas and N.sub.R are an integer number of a plurality of receiving antennas, which H is decomposed by the singular value decomposition and represented by $H=U{V}^H=\underset{i=1}{\overset{N_B}{.Math.}}{}_i{u}_i{v}_i^H$ wherein U is a unitary matrix, V is a unitary matrix, u.sub.i=1, . . . , N.sub.B, are orthonormal column vectors of U, and v.sub.i, i=1, . . . , N.sub.B, are orthonormal column vectors of V, is an N.sub.BN.sub.B diagonal matrix with singular values .sub.i=.sub.1, .sub.2, . . . , .sub.N.sub.B of the MIMO module, and N.sub.B is an integer; step c) evaluating a communication condition of the plurality of beams and selecting a plurality of sub-channels, which number is n.sub.B, out of the individual channels of the plurality of beams as transmit beams, and transmitting a signal and the index via the antennas of the transmit beams at the transmitter end; and step d) receiving and recognizing the index at the receiver end, by the antennas, and further detecting the symbols of the transmit beams received at the receiver end.

2. The method of claim 1, wherein the MIMO module is compatible with orthogonal frequency-division multiplexing (OFDM).

3. The method of claim 1, wherein the index is represented by either a plurality of M-ary quadrature amplitude modulation symbols or a plurality of spatial bits.

4. The method of claim 3, wherein the spatial bits are binary.

5. The method of claim 1, wherein the number of the transmit beams equals to the number of the sub-channel, n.sub.B, plus one.

6. The method of claim 5, wherein the index comprises a zero bit to represent those unselected sub-channels.

7. A second method of beam-indexed spatial modulation (BISM) for a multiple-input and multiple output (MIMO) module having a transmitter end and a receiver end, comprising: step a) confirming that the MIMO module has a plurality of beams, the number of the plurality of beams is N.sub.B, each of the plurality of beams is an individual channel constituted by one of antennas at the transmitter end and one of antennas at the receiver end; step b) providing an index to the N.sub.B beams through a channel matrix H which is decomposed by the angular steering vectors and represented by: $H=\underset{i=1}{\overset{N_B}{.Math.}}{a}_i{e}_r\left({}_{\mathrm{ri}}\right)e_t^{*}\left({}_{\mathrm{ti}}\right)$ wherein a.sub.i is an attenuation of the i-th beam, e.sub.t (.sub.ti) is a vector of angle of departure representing by $e_t\left({}_{\mathrm{ti}}\right)=\frac{1}{\sqrt{N_T}}[1,\exp \left(-j2{}_t{}_{\mathrm{ti}}\right),\exp \left(-j22{}_t{}_{\mathrm{ti}}\right),\mathrm{.Math.},{\exp \left(-j2\left(N_T-1\right){}_t{}_{\mathrm{ti}}\right)]}^T$ e.sub.r (.sub.ri) is a vector of angle of arrival representing by $e_r\left({}_{\mathrm{ri}}\right)=\frac{1}{\sqrt{N_T}}[1,\exp \left(-j2{}_r{}_{\mathrm{ri}}\right),\exp \left(-j22{}_r{}_{\mathrm{ri}}\right),\mathrm{.Math.},{\exp \left(-j2\left(N_R-1\right){}_r{}_{\mathrm{ri}}\right)]}^T$ where .sub.t is the distance between the antennas at the transmitter, .sub.r is the distance between the antennas at the receiver, .sub.ti=cos(.sub.ti) is cosine of the angle of departure at the transmitter, wherein .sub.ti is an angle of departure at the transmitter, and .sub.ri=cos(.sub.ri) is cosine of the angle of arrival at the receiver end, wherein .sub.r, is an angle of arrival at the receiver, N.sub.T are the number of a plurality of transmitting antennas and N.sub.R are the number of a plurality of receiving antennas; and step c) evaluating a communication condition of the plurality of beams and selecting a plurality of sub-channels, which number is n.sub.B, out of the individual channels of the plurality of beams as transmit beams, and transmitting a signal and the index via the antennas of the transmit beams at the transmitter end; and step d) receiving and recognizing the index at the receiver end, by the antennas, and further detecting the symbols of the transmit beams received at the receiver end.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIG. 1 shows a BER comparison of SM-MIMO with VBLAST MIMO detection scheme and maximum ratio combining (MRC) MIMO detection scheme.

(2) FIG. 2 shows an GSM-MIMO block diagram under which more than one antennas are activated (i.e. n.sub.T1) and M-QAM symbols are transmitted on the activated antennas.

(3) FIG. 3 illustrates a block diagram of OFDM-IM where a subset of available subcarriers is selected to be active based on the information bits.

(4) FIG. 4 illustrates an uncoded BER performance of a MIMO-OFDM-IM and a typical MIMO-OFDM schemes for three MIMO configurations n.sub.Tn.sub.R=22; 44, and 88.

(5) FIG. 5 shows a flow chart of the present invention.

(6) FIG. 6 illustrates a scheme of performing antenna-index modulation on SM-MIMO.

(7) FIG. 7 shows a scheme of performing subcarrier-index modulation on OFDM-IM.

(8) FIG. 8 shows a BISM scheme of the present invention.

(9) FIG. 9 shows an example of performing BISM on MIMO-OFDM.

(10) FIG. 10 exemplarily demonstrates performing BISM on MIMO.

(11) FIG. 11 illustrates the degrees of BER performance in both conventional MIMO and BISM scenarios.

(12) FIG. 12(a) shows trade-off between SE and EE.

(13) FIG. 12(b) illustrates trade-off between SE and EE in the condition where there are eight data stream (N.sub.S=8) and the active data streams n.sub.B are 8, 7, . . . and 1.

(14) FIG. 13 shows ergodic capacities of a conventional MIMO and MIMO-BISM systems in view of various SNR.

(15) FIG. 14 shows an exemplary system of applying BISM to a conventional MIMO

(16) FIG. 15 is a block diagram showing how to apply BISM into a MIMO-OFDM having N subcarriers.

(17) FIG. 16 is an example of applying BISM to a MIMO-OFDM system.

(18) FIG. 17 is an exemplary joint design of BISM and OFDM-IM system.

DETAILED DESCRIPTION OF THE INVENTION

(19) FIG. 6 illustrates a scheme of performing antenna-index modulation on SM-MIMO. As mentioned, SM requires hardware components to switch among the selected antennas. The speed of spatial bits (i.e. index modulation bit) is therefore suffered. Further, the SM-MIMO is incompatible with OFDM-based communication systems because the subcarriers are affected by the antenna switching simultaneously. The incompatibility makes SM-MIMO impractical for real-world telecommunications.

(20) FIG. 7 shows a scheme of performing subcarrier-index modulation on OFDM-IM. Although OFDM-IM is inherently compatible with OFDM-based systems, it does not exploit the degree-of-freedom that MIMO can provide wisely. Those unselected and inactive subcarriers will not be able to transmit at all.

(21) This invention proposes methods of beam-indexed spatial modulation (BISM) to resolve the above issues. This invention can adopt the combinations of MIMO channel eigen-modes to carry out index modulation. Thus, the low processing speed and incompatibility issues occurred in SM-MIMO can be solved. Additionally, the present invention provides a higher degree of efficiency with respect to spectrum and energy comparing to OFDM-IM.

(22) In the context of MIMO, the eigen-modes are independent spatial channels in the MIMO channel decomposition systems. Each of the eigen-modes carries independent data stream in the MIMO communication system. In a MIMO-OFDM system, the maximum number of data streams in each subcarrier represents the degree-of-freedom of the system. The combinations of the selected data streams are encoded for index modulation in BISM.

(23) Additionally, unlike OFDM-IM, the proposed BISM achieves a higher degree of efficiency with respect to energy (bps/Joule) and/or spectrum (bps/Hz) in view of different channels and system configurations. With comparison to conventional MIMO-OFDM, BISM improves the efficiencies of spectrum and energy and provides a higher degree of flexibility.

(24) The following paragraphs will discuss the features and advantages that BISM can achieve as well as analysis results of performance simulations and capacity.

(25) Assuming a MIMO channel has N.sub.T transmitting antennas and N.sub.R receiving antenna. Further, defining N.sub.B as the number of channel ranks of the MIMO channel and n.sub.B as the number of transmit beam in every transmission. It is noted that N.sub.T>N.sub.B>n.sub.B>1. It should also be noted that there are many alternatives to generate the transmit beams; they will be discussed below.

(26) In one embodiment, the transmit beams can be generated by adopting singular value decomposition (SVD) to decompose the channel matrix HC.sup.N.sup.T.sup.N.sup.R. The channel matrix H is:

(27) $H=U{V}^H=\underset{i=1}{\overset{N_B}{.Math.}}{}_i{u}_i{v}_i^H$
U and V are unitary matrices and the column vectors, u.sub.i,v.sub.i of U and V, respectively, are orthonormal. The is an N.sub.BN.sub.B diagonal matrix with singular values .sub.i=.sub.1, .sub.2, . . . , .sub.N.sub.B of the channel. Each sub-channel H.sub.i is obtained by H.sub.i=.sub.iu.sub.iv.sub.i.sup.H, wherein i=1, 2, . . . , N.sub.B representing the i-th eigen-mode as well as the i-th eigen-beam.

(28) In one embodiment, the transmit beam can be obtained by using .sub.ti (angle of departure) and .sub.ri, (angle of arrival) to decompose the channel matrix H. The channel matrix H can be decomposed by the angular steering vectors and is represented by

(29) $H=\underset{i=1}{\overset{N_B}{.Math.}}{a}_i{e}_r\left({}_{\mathrm{ri}}\right)e_t^{*}\left({}_{\mathrm{ti}}\right)$
a.sub.i is an attenuation of the ith beam, e.sub.t is a vector of angle of departure which equation is:

(30) 0$e_t\left({}_{\mathrm{ti}}\right)=\frac{1}{\sqrt{N_T}}[1,\exp \left(-j2{}_t{}_{\mathrm{ti}}\right),\exp \left(-j22{}_t{}_{\mathrm{ti}}\right),\mathrm{.Math.},{\exp \left(-j2\left(N_T-1\right){}_t{}_{\mathrm{ti}}\right)]}^T$
Additionally, e.sub.r is a vector of angle of arrival which equation is:

(31) $e_r\left({}_{\mathrm{ri}}\right)=\frac{1}{\sqrt{N_T}}[1,\exp \left(-j2{}_r{}_{\mathrm{ri}}\right),\exp \left(-j22{}_r{}_{\mathrm{ri}}\right),\mathrm{.Math.},{\exp \left(-j2\left(N_R-1\right){}_r{}_{\mathrm{ri}}\right)]}^T$
where .sub.t is the distance between the antennas at the transmitter, .sub.r is the distance between the antennas the receiver, .sub.ti=cos(.sub.ti) is cosine of the angel of departure at the transmitter, and .sub.ri=cos(.sub.ri) is cosine of the angle of arrival at the receiver. Each sub-channel H.sub.i is obtained by H.sub.i=a.sub.ie.sub.r(.sub.ri)e.sub.t*(.sub.ti), i=1, . . . , N.sub.B representing the i-th angular beam.

(32) The transmitter beamformer can allocate a power on the n.sub.B transmit beams (out of N.sub.B beams) at every transmission. The receiver can detect the reliability in the spatial domain by using the independent spatial signatures H.sub.l. Therefore, the achievable rate R.sub.BISM can be represented by the spatial bits and the n.sub.B streams of signal bits:

(33) $R_{\mathrm{BISM}}=.Math.{\log }_2\left(\begin{array}{c}{N}_B\\ n_B\end{array}\right).Math.+n_B{\log }_{2}M$

(34) The first term on the right-hand side of the above R.sub.BISM equation is the spatial bits through generalized spatial modulation in the beam space. The second term is M-ary amplitude-phase modulation through n.sub.13 spatial multiplexing.

(35) The BISM codebook can be designed in accordance with the transmission sub-channel. The number of data streams equals to n.sub.B plus one stream provided for the index modulated spatial bits. The spatial bits determine the active eigen-beams for transmitting M-QAM symbols, and the inactive eigen-beams send over null signal space.

(36) FIG. 8 shows a BISM scheme of the present invention. Assuming the number of channel ranks N.sub.B equals to four (N.sub.B=4), and the number of transmit beams n.sub.B equals to 3 (n.sub.B=3). Thus, a number of stream (i.e. N.sub.S) for BISM transmission can be designed accordingly: N.sub.S=n.sub.B+1 (i.e. in this case, N.sub.S=4).

(37) Table 2 demonstrates a BISM spatial constellation based on eigen decomposition with the previous assumptions: N.sub.B=4, n.sub.B=3. It should be noted that x.sub.i is the QAM modulated symbol and i=1, 2, . . . , n.sub.B

(38) TABLE-US-00002 TABLE 2 demonstrates a BISM spatial constellation based on the previous assumptions Selected eigen- Spatial BISM beam index bits Sub-channel codeword 1, 2, 3 00 $\underset{i\left\{1,2,3\right\}}{.Math.}{H}_i=\underset{i\left\{1,2,3\right\}}{.Math.}{}_i{u}_i{v}_i^H$ [x.sub.1, x.sub.2, x.sub.3, 0].sup.T 1, 2, 4 01 $\underset{i\left\{1,2,4\right\}}{.Math.}{H}_i=\underset{i\left\{1,2,4\right\}}{.Math.}{}_i{u}_i{v}_i^H$ [x.sub.1, x.sub.2, 0, x.sub.3].sup.T 1, 3, 4 10 $\underset{i\left\{1,3,4\right\}}{.Math.}{H}_i=\underset{i\left\{1,3,4\right\}}{.Math.}{}_i{u}_i{v}_i^H$ [x.sub.1, 0, x.sub.2, x.sub.3].sup.T 2, 3, 4 11 $\underset{i\left\{2,3,4\right\}}{.Math.}{H}_i=\underset{i\left\{2,3,4\right\}}{.Math.}{}_i{u}_i{v}_i^H$ [0, x.sub.1, x.sub.2, x.sub.3].sup.T

(39) Assuming the spatial bits in transmitted data is binary constituted by 10. The sub-channels H.sub.1, H.sub.3, H.sub.4 containing the 1.sup.st, 3.sup.rd, and 4.sup.th eigen-beams (or eigen-modes) are selected to transmit signals. The BISM codeword uses the corresponding stream index to carry different QAM modulated symbols, such as [x.sub.1, 0, x.sub.3, x.sub.4].sup.T. As discussed, SVD is adopted to decompose the channel matrix, the effective channel, composed of the precoder, channel and combiner, is equivalent to parallel channel (e.g. diagonal matrix). Each symbol in the codeword maps to the corresponding eigen-beam; thus, the eigen-beam can be selected by using difference codeword.

(40) BISM originated from spatial modulation (SM) and generalized spatial modulation (GSM). BISM is advantageous over SM-MIMO and OFDM-IM because it uses eigen-beam selection rather than antenna selection to represent the spatial bits. The advantages of BISM include but no limit to the following:

(41) i) it provides an enhanced transmission data rate without hardware constraint caused by antenna switch;

(42) ii) it is compatible with MIMO-OFDM system;

(43) iii) it is energy efficient;

(44) iv) it provides a better performance of bit error rate (BER);

(45) v) the trade-off between spectral efficiency (SE) and energy efficiency (EE) is balanced; and

(46) vi) it is compatible with multiuser MIMO-OFDM systems.

(47) Enhanced Transmission Data Rate without Hardware Constraint Caused by Antenna Switch

(48) As discussed, the processing speed of index modulation bit is restrained by antenna switching in SM-MIMO. However, BISM uses eigen-beam combinations to represent spatial bits. Therefore, instead of switching antennas on and off, the eigen-beam combinations can be configured in the baseband.

(49) Further, the transmission data rate in BISM is as fast as the system's symbol rate determined by channel bandwidth. For example in IEEE 802.11a, the bandwidth for each sub-channel and the symbol rate can be as high as 20 MHz. On the other hand, in SM-MIMO systems, the index bit transmission rate is dominated by the antenna switching rate which is as slow as 2000 times of switching per second. Hence, in the absence of antenna switching restriction, BISM helps to enhance the index bit transmission rate than the SM-MIMO system by 10.sup.4 times faster.

(50) MIMO-OFDM System Compatibility

(51) The issue of SM-MIMO's incompatibility with MIMO-OFDM has been iterated. The reason is the subcarriers in OFDM system are vulnerably affected by the antenna index selection. The BISM of the present invention exploits the combinations of eigen-beams for index modulation. The eigen-beams are equivalent to the MIMO channel eigen-modes in each of the subcarriers in MIMO-OFDM system.

(52) FIG. 9 shows an example of performing BISM on MIMO-OFDM system (MIMO-OFDM-BISM). Assuming N.sub.B=4 and n.sub.B=3. Additionally, each eigen-mode carries one independent data stream; the combinations of data stream index carry the index modulation in baseband.

(53) Energy Efficiency

(54) Unlike conventional MIMO system having N.sub.B ranks, MIMO-OFDM-BISM of the present invention selects a subset of n.sub.B ranks for the transmission of M-ary QAM symbols. Although less ranks (or data streams) are used to transmit data, BISM nevertheless is capable of carrying extra spatial bits by index modulation. Assuming a fixed power P is given to each data stream of conventional MIMO and MIMO-OFDM-BISM systems. Although in both scenarios the BER of M-ary QAM symbols in each data stream of MIMO-OFDM-BISM system remains the same, the overall power consumption after performing BISM is down to n.sub.B/N.sub.B.

(55) FIG. 10 exemplarily demonstrates performing BISM on MIMO system (MIMO-BISM). Assuming N.sub.B=4 and n.sub.B=3. Further assuming that a fixed power is given to each data stream, the overall power consumption is then down to after performing BISM.

(56) Bit Error Rate (BER) Performance

(57) FIG. 11 illustrates the comparison of BER performance in both conventional MIMO and MIMO-BISM scenarios. It appears that the MIMO-BISM system can achieve the same BER as the conventional MIMO system (without BISM) but with lower E.sub.b/N.sub.0. For instance, assuming the transmission rate for transmitting a subcarrier is 8-bit and three transmitted data streams (N.sub.B=3) are chosen, the MIMO-BISM system shows the performance of 1.21 dB E.sub.b/N.sub.0 at 10.sup.5 BER. The lower E.sub.b/N.sub.0 value shows BISM can effectively increase energy efficiency.

(58) Spectral Efficiency (SE) and Energy Efficiency (EE) Trade-Off

(59) The question of how to balance between spectral efficiency (SE) and energy efficiency (EE) is a common question concerning in wireless communication systems. In BISM, assuming n.sub.B=N.sub.B1, the achievable rate R.sub.BISM is:

(60) $R_{\mathrm{BISM}}=.Math.{\log }_2\left(\begin{array}{c}{N}_B\\ N_B-1\end{array}\right).Math.+\left(N_B-1\right){\log }_{2}M=.Math.{\log }_2\left(\begin{array}{c}{N}_B\\ 1\end{array}\right).Math.+\left(N_B-1\right){\log }_{2}M{N}_B{\log }_{2}M,\mathrm{if}{N}_{B}M$

(61) It is noted that M is the order of M-QAM modulation symbols. As far as spectral efficiency is concerned, if N.sub.BM, then a more spatial bits can be obtained by performing BISM than using a mere spatial multiplexing in conventional MIMO. In such instance, only N.sub.T1 RF chains are used so as to reduce the complexity of fully turn-on spatial multiplexing.

(62) In another extreme case where only one beam is active. The achievable rate R.sub.BISM is:

(63) $R_{\mathrm{BISM}}=.Math.{\log }_2\left(\begin{array}{c}{N}_B\\ 1\end{array}\right).Math.+{\log }_{2}M$

(64) In the extreme case, BISM is equivalent to spatial modulation, featuring with low complexity and high energy efficiency (bps/Joule).

(65) FIG. 12(a) shows the trade-off between SE and EE. Assuming there are five data stream (N.sub.S=5) and the active data streams n.sub.B are 4, 3, 2 and 1. It should be noted that if n.sub.B=N.sub.S=4, that means a comparison is made between a conventional MIMO and a MIMO-BISM systems.

(66) The active data streams are loaded with M-ary QAM modulation symbols. Assuming various modulations including BPSK, QPSK, 16-QAM and 32-QAM are all considered. It appears that a higher degree of spectral efficiency can be achieved in both conventional MIMO and MIMO-BISM systems when a higher order M-ary QAM presents. As shown in FIG. 12(a), it appears that the MIMO-BISM system always has a higher EE than the conventional MIMO system if the same order of M-ary QAM symbols are loaded. With comparison to MIMO-BISM, the conventional MIMO system generally has a higher spectral efficiency. Also illustrated in FIG. 12(a), in some cases the MIMO-BISM system can still have a higher spectral efficiency even at a low order QAM.

(67) Similarly, FIG. 12(b) illustrates trade-off between SE and EE in the condition where there are eight data stream (N.sub.S=8) and the active data streams n.sub.B are 8, 7, . . . and 1.

(68) The capacity of BISM (C.sub.BISM) can be expressed by the below equation contributed by the spatial domain (i.e. the first term at the right-hand side) and signal domain (i.e. the second term also at the right-hand side).

(69) $C_{\mathrm{BISM}}=\underset{i=1}{\overset{n_B}{.Math.}}{\log }_2\left(.Math.I_{N_B}+\frac{E_s}{N_0}{H}_i{H}_i^H.Math.\right)+.Math.{\log }_2\left(\begin{array}{c}{N}_B\\ n_B\end{array}\right).Math.\left(1-P_{e,\mathrm{bit}}\right)$
The capacities from the two domains are traded off with each other, or reconfigurable with different setup of n.sub.B. In the absence of loss of generality, BISM sum rate can be expressed as the summation of n.sub.B MIMO sub-channels and the index modulation bits throughput.

(70) 0$P_{e,\mathrm{bit}}\underset{j}{.Math.}\underset{k}{.Math.}\frac{d\left(j,k\right)}{M}{2}^{2N_R+1}{\left(\frac{N\left(j,k\right)}{n_T}\right)}^{-N_T}{\left(\frac{E_s}{N_0}\right)}^{-N_R}$

(71) It is noted that E.sub.S is the M-ary QAM symbol energy, N.sub.0 is the variance of AWGN (additive white Gaussian) noise, d(j, k) is the number of bits in difference (or Hamming distance) between two index modulation codeword, and N(j, k) is the number of distinct columns between two sub-MIMO channels.

(72) FIG. 13 shows ergodic capacities of a conventional MIMO and MIMO-BISM systems in view of various SNR. The grey part illustrates the distribution of MIMO-BISM, while the black part is that of the conventional MIMO.

(73) As can be seen, when at the low SNR region, it is more spectral efficient to turn on less beams. On the other hand, as the increase of SNR, in order to achieve a higher data rate than that in MIMO, more beams need to be turned on. As mentioned early, in a high SNR scenario and when a sufficient number of data streams (n.sub.B) are utilized, BISM's capacity increases proportionally and surpass the capacity of the conventional MIMO as more beam patterns being used. As more beam patterns are turned on, more diversity gains and multiplexing gains can be achieved.

(74) Compatible with Multiuser MIMO-OFDM Systems

(75) In a multi-user MIMO-OFDM communication system, the subcarriers are divided into sub-blocks as radio resources. The blocks of the radio resources are then allocated to multiple users for multiple accesses.

(76) The BISM of the present invention exploits index modulation in the eigen-mode dimension which can consequently be compatible with the way in which resources are allocated in a multiuser MIMO-OFDM system. This makes BISM advantageous over SM-MIMO and OFDM-IM systems. The former cannot adopt the same resource allocation method, while the resource allocation of the latter is less flexible because OFDM-IM exploits index modulation in its subcarriers.

Exemplary Embodiments

(77) General Diagram of Switchable System Between BISM and MIMO

(78) FIG. 14 shows an exemplary system of performing BISM to a conventional MIMO (i.e. MIMO-BISM). The procoders F.sub.BB and F.sub.RF, and combiners W.sub.BB and W.sub.RF in the system are decided by the channel information and hardware constraint. The reconfiguration mechanism determines to adopt either the MIMO-BISM or the conventional MIMO based on the channel information and user requirements. The settings of precoders and combiners are irrelevant to the switching between MIMO-BISM and the conventional MIMO. Therefore, it is feasible to apply BISM and reconfiguration mechanism to the current communication system.

(79) BISM Realization Example for MIMO OFDM System

(80) FIG. 15 is a block diagram showing how to performing BISM on a MIMO-OFDM having N subcarriers. In one embodiment, the number of data stream N.sub.S equals to MIMO channel rank in each subcarrier. As for as the k-th (k=1, 2, . . . , N) subcarrier is concerned, the data vector d[k] of length N.sub.S contains N.sub.Sn.sub.B digital modulation symbols (e.g. M-QAM symbol) and n.sub.B null elements. By the beam index on the data vector, the

(81) $\left(\begin{array}{c}{N}_s\\ n_B\end{array}\right)$
data vector patterns represent the spatial symbols. The transmit precoding symbol vector is:
x[k]=F.sub.BB[k]d[k]

(82) It is noted that F.sub.BB [k] is a baseband precoder matrix. The post-IFFT symbol vector is s[n], n=1, 2, . . . , N where each element within s[n] is:

(83) $s_l\left[n\right]=\frac{1}{\sqrt{N}}\underset{k=0}{\overset{N-1}{.Math.}}{x}_l\left[k\right]e^{j\frac{2}{N}\mathrm{kn}},l=0,1,\mathrm{.Math.},N_s-1$

(84) Assuming the system is operated with coherence time and coherence bandwidth, analog beamforming matrix (F.sub.RF) of a hybrid beaming MIMO can be applied during one OFDM symbol.

(85) The transmit symbol vector at time n is b[n] representing as b[n]=F.sub.RFs[n], n=0, 1, . . . , N1. It is noted that F.sub.RF is the analog beamforming matrix. The received symbol vector on k-th subcarrier of an OFDM symbol at the receiver in a baseband is:
y[k]=W.sub.RFH[k]F.sub.RFF.sub.BB[k]d[k]+W.sub.RFn[k]

(86) It is noted that H[k] is the frequency domain full channel matrix for the k-th subcarrier, W.sub.RF is the analogy post-coder matrix and n[k] is the complex AWGN vector.

(87) The MIMO channel in the k-th subcarrier, H[k], can be decomposed by using SVD, so that
H[k]=U[k][k]V.sup.H[k].
Additionally, the eigen-modes in each of the subcarriers are exploited in BISM for index modulation.

(88) For example, as shown in FIG. 16, assuming BIMS is performing on a MIMO-OFDM system where there are four eigen-modes in each subcarrier, N.sub.S=4, n.sub.B=3 and x.sub.1, x.sub.2, x.sub.3 are M-ary QAM. The location of zero is selected by the input bits with random distribution. As shown, each location of zero represents

(89) $.Math.{\log }_2\left(\begin{array}{c}{N}_s=4\\ n_B=3\end{array}\right).Math.=2\mathrm{bits}.$
Joint Design of BISM in OFDM-IM

(90) The concept of BISM in the eigen-mode dimension can be implemented jointly with OFDM-IM as two-dimensional index modulation scheme. Assuming in a MIMO system there are N.sub.S eigen-modes and N subcarriers dividing into G smaller and manageable OFDM-IM sub-block each containing N.sub.c subcarriers to perform IM, wherein N=G*N.sub.c.

(91) For each sub-block, n.sub.0 out of N.sub.c available subcarriers can be selected as the active subcarriers based on the index selection bits. Each of the active subcarriers sends over M-QAM modulation symbols based on the p.sub.2 modulation bits.

(92) For each OFDM-IM frame, the spectral efficiency R is

(93) $R=\left({\log }_2\left(\begin{array}{c}{N}_s{N}_c\\ n_0\end{array}\right)+n_0{\log }_2\left(M\right)\right)G$

(94) FIG. 17 is an exemplary joint design of BISM and OFDM-IM where N.sub.s=4, N.sub.c=4, and n.sub.0. Each resource block of N.sub.sN.sub.c degree-of-freedoms can be exploited for index modulation. Thus, as shown in FIG. 17, there are N.sub.sN.sub.c=44 degree-of-freedoms for index modulation.

(95) The above-described embodiments of this invention are presented for purposes of illustration and not of limitation. Of course, those skilled in the art will recognize many modifications may be made to this configuration without departing from the scope of the disclosed aspects.