MULTIPLE INPUT MULTIPLE OUTPUT ORTHOGONAL FREQUENCY DIVISION MULTIPLEXING WITH INDEX MODULATION, MIMO-OFDM-IM, COMMUNICATIONS SYSTEM

Abstract

A communication system for the next generation wireless communications technology standards. The communication system architecture is created by the combination of the index modulation technique and the multiple input, multiple output orthogonal frequency division multiplexing which eliminates the need to utilize complex equalizers by parsing high speed data strings and transmitting them over multiple orthogonal subcarriers, and allows the bits to be transmitted via active subcarrier indices. The OFDM-IM and multiple input multiple output communication techniques are used in tandem. The communication system can be used in future generation mobile communication systems and standards (5G and beyond), Local Area Network system and standards, terrestrial digital TV system and standards, multi-carrier communication systems and broadband digital communication systems.

Claims

1. A multiple input multiple output orthogonal frequency division multiplexing with index modulation (MIMO-OFDM-IM) communications system, comprising: at least two transmit antennas in a transmitter side for transmitting signals obtained after orthogonal frequency division multiplexing with index modulation, wherein each transmit antenna transmits OFDM-IM frames; at least two receive antennas in a receiver side for receiving the signals transmitted from the transmit antennas for detection and demodulation; a block interleaver in the transmitter side, wherein said block interleaver interleaves the OFDM-IM frames to minimize the correlation between channel coefficients corresponding to different OFDM-IM subcarriers; a minimum mean square error (MMSE) detection and log-likelihood ratio (LLR) calculation based detector in the receiver side to detect and demodulate the OFDM-IM frames received from said each transmit antenna, wherein the detector is enabled to separate the OFDM-IM frames received from said each transmit antenna; a successive minimum mean square error (MMSE) estimator to eliminate interference between different transmit antennas performing a filtering process, wherein the successive MMSE estimator is configured to have an enhanced operation mode based on ordered successive interference cancellation (OSIC) algorithm and calculates a quality measure for each transmit antenna and starts the detection from the one that has a maximum signal-to-noise ratio; a plurality of log-likelihood ratio (LLR) estimators to perform a search over possible realizations of data symbols for said each subcarriers to calculate a posterior probability ratio, wherein the posterior probability ratio is used as a benchmark for determining whether the corresponding subcarrier of a given transmit antenna is active or not; a block deinterleaver in the receiver side which performs an exact opposite function of the block interleaver; wherein each branch of the transmitter side uses reference look-up tables or combinatorial number theory for selection of indices of active subcarriers according to index selection data bits; wherein a cyclic prefix addition is performed after an inverse fast Fourier transform (IFFT) process on the signals in the transmitter section and a cyclic prefix subtraction is performed after the signals are received at the receive antennas followed by a fast Fourier transform process for the decoupling of the received signals without inter-carrier interference.

2. The multiple input multiple output orthogonal frequency division multiplexing with index modulation (MIMO-OFDM-IM) communications system according to claim 1, wherein the system is configured to adjust the number of active subcarriers

3. The multiple input multiple output orthogonal frequency division multiplexing with index modulation, MIMO-OFDM-LM communications system according to claim 1, wherein the system is configured to operate for different multiple input multiple output systems.

4. A 5G wireless network system comprising: at least two transmit antennas in a transmitter side for transmitting signals obtained after orthogonal frequency division multiplexing with index modulation, wherein each transmit antenna transmits OFDM-IM frames; at least two receive antennas in a receiver side for receiving the signals transmitted from the transmit antennas for detection and demodulation; a block interleaver in the transmitter side, wherein said block interleaver interleaves the OFDM-IM frames to minimize the correlation between channel coefficients corresponding to different OFDM-LM subcarriers; a minimum mean square error (MMSE) detection and log-likelihood ratio (LLR) calculation based detector in the receiving section to detect and demodulate the OFDM-IM frames received from said each transmit antenna, wherein the detector is enabled to separate the OFDM-LM frames received from said each transmit antenna; a successive minimum mean square error (MMSE) estimator to eliminate interference between different transmit antennas performing a filtering process, wherein the successive MMSE estimator is configured to have an enhanced operation mode based on ordered successive interference cancellation (OSIC) algorithm and calculates a quality measure for each transmit antenna and starts the detection from the one that has a maximum signal-to-noise ratio; a plurality of log-likelihood ratio (LLR) estimators to perform a search over possible realizations of data symbols for said each subcarriers to calculate a posterior probability ratio, wherein the posterior probability ratio is used as a benchmark for determining whether the corresponding subcarrier of a given transmit antenna is active or not; a block deinterleaver in the receiver side which performs an exact opposite function of the block interleaver; wherein each branch of the transmitting section uses reference look-up tables or combinatorial number theory for selection of indices of active subcarriers according to index selection data bits; wherein a cyclic prefix addition is performed after an inverse fast Fourier transform (IFFT) process on the signals in the transmitter section and a cyclic prefix subtraction is performed after the signals are received at the receive antennas followed by a fast Fourier transform process for the decoupling of the received signals without inter-carrier interference.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0007] The drawings and the related descriptions, which are used for the better explanation of the multiple input multiple output orthogonal frequency division multiplexing with index modulation, MIMO-OFDM-IM, communications system developed with the present invention, are provided below.

[0008] FIG. 1: Transmitter/Receiver Structure of the MIMO-OFDM-IM System

[0009] FIG. 2: Block Scheme of the OFDM Index Modulator

[0010] FIG. 3: BER performance of the MIMO-OFDM-IM system for BPSK modulation

[0011] FIG. 4: BER performance of the MIMO-OFDM-IM system for QPSK modulation

[0012] FIG. 5: BER performance of the MIMO-OFDM-IM system for 16-QAM modulation

[0013] FIG. 6: Reference Active Indices Selection Table 1

[0014] FIG. 7: Reference Active Indices Selection Table 2

DEFINITIONS OF THE ELEMENTS/SECTIONS/PARTS FORMING THE INVENTION

[0015] The parts and sections, which are presented in the drawings for a better explanation of the wireless communication system developed with the present invention, have been enumerated and the counterpart of each number is presented below. [0016] 1. Multiple input multiple output orthogonal frequency division multiplexing with index modulation, MIMO-OFDM-IM, communications system [0017] 2. Orthogonal frequency division multiplexing with index modulation (OFDM-IM) blocks [0018] 3. Block interleaver () [0019] 4. Inverse fast. Fourier transform (IFFT) [0020] 5. Cyclic prefix (CP) addition [0021] 6. Transmit antennas [0022] 7. Receive antennas [0023] 8. Cyclic prefix (CP) subtraction [0024] 9. Fast Fourier transform (FFT) [0025] 10. Block deinterleaver (.sup.1) [0026] 11. Successive minimum mean square error (MMSE) detector [0027] 12. Log-likelihood ratio (LLR) calculator [0028] 13. The Receiver [0029] 14. The Transmitter

DETAILED DESCRIPTION OF THE INVENTION

[0030] A MIMO system employing transmit and R receive antennas is considered. For the transmission of each frame, a total of riff bits enter the transmitter (14) and first split into groups (14) and the corresponding rte bits are processed in each branch of the transmitter by the OFDM index modulators (2). The incoming in information bits are used to form the N.sub.F1 OFDM-IM block x.sub.i=[x.sub.i(1) x.sub.i(2) . . . x.sub.i(N.sub.F)].sup.T, t=1,2, . . . ,T in each branch of the transmitter (2), where N.sub.F is the size of the fast Fourier transform (FFT) (4) and x.sub.i(n.sub.f){0,S},n.sub.f=1,2, . . . ,N.sub.F and S represents the signal constellation. According to the OFDM-IM principle (2), which is carried out simultaneously in each branch of the transmitter, these bits are split into G groups each containing p=p.sub.1+p.sub.2 bits, which are used to form OFDM-IM subblocks x.sub.i.sup.g=[x.sub.i.sup.g(1) x.sub.i.sup.g(2) . . . x.sub.i.sup.g(N)].sup.T, g=1,2, . . . ,G of length N=N.sub.F/G, where x.sub.i.sup.g(n){0,S}, n=1,2, . . . ,N. According to the corresponding p.sub.1=log.sub.2(C(N,K)) bits, only K out of N available subcarriers are selected as active by the index selector (2) at each subblock g, while the remaining NK subcarriers are inactive and set to zero. On the other hand, the remaining p.sub.2=K log.sub.2(M) bits are mapped onto the considered M-ary signal constellation (2). Active subcarrier index selection is performed by the reference look-up tables at OFDM index modulators (2) of the transmitter (14) for smaller N and K values. The considered reference look-up tables for N=4, K=2 and N=4, K=3 are given in FIGS. 6 and 7, respectively, where s.sub.kS for k=1,2, . . . K. For higher and K values, a combinatorial method based index selection algorithm is employed.

[0031] The OFDM index modulate (2) in each branch of the transmitter obtain the OFDM-IM subblocks first and then concatenate these G subblocks to form the main OFDM blocks x.sub.t, t=1,2, . . . ,T. In order to transmit the elements of the subblocks from uncorrelated channels, GN block interleavers () (3) are employed at the transmitter. The block interleaved OFDM-IM frames {tilde over (x)}.sub.t, t=1,2, . . . ,T are processed by the inverse FFT (IFFT) operators (4) to obtain {tilde over (q)}.sub.t, t=1,2, . . . ,T. After the addition of cyclic prefix of C.sub.p samples, parallel-to-serial and digital-to-analog conversions (5), the resulting signals sent simultaneously from T transmit antennas (6) over a frequency selective Rayleigh fading MIMO channel, where g.sub.r,t.sup.L1 represents the L-tap wireless channel between the transmit antenna t and the receive antenna r. Assuming the wireless channels remain constant during the transmission of a MIMO-OFDM-IM frame and C.sub.p>L, after removal of the cyclic prefix (8) and performing FFT operations in each branch of the receiver (9), the input-output relationship of the MIMO-OFDM-IM scheme in the frequency domain is obtained as (13)

{tilde over (y)}.sub.r=.sub.t=1.sup.Tdiag({tilde over (x)}.sub.t)h.sub.r,t+w.sub.r

[0032] for r=1,2, . . . ,R, where {tilde over (y)}.sub.r=[{tilde over (y)}.sub.r(1) {tilde over (y)}.sub.r(2) . . . {tilde over (y)}.sub.r(N.sub.F)].sup.T is the vector of the received signals for receive antenna r (13), h.sub.r,t.sup.N.sup.F.sup.1 represents the frequency response of the wireless channel between the transmit antenna t and receive antenna r, and w.sub.r.sup.N.sup.F.sup.1 is the vector of noise samples. The elements of h.sub.r,t and w.sub.r follow CN (0,1) and CN (0,N.sub.0,F) distributions, respectively, where N.sub.0,F denotes the variance of the noise samples in the frequency domain. We define the signal-to-noise ratio (SNR) as SNR=E.sub.b/N.sub.0,T where E.sub.b=(N.sub.F+C.sub.p)m [joules/bit] is the average transmitted energy per bit.

[0033] After block deinterlaving (10) in each branch of the receiver (13), the received signals are obtained for receive antenna as

y.sub.r=.sub.t=1.sup.Tdiag(x.sub.t){hacek over (h)}.sub.r,t+{hacek over (w)}.sub.r

[0034] Where {hacek over (h)}.sub.r,t and {hacek over (w)}.sub.r are deinterleaved versions of h.sub.r,t and w.sub.r t respectively. The detection of the MIMO-OFDM-IM scheme can be performed by the separation of the received signals for each subblock g=1, 2, . . . ,G as follows

y.sub.r.sup.g=.sub.t=1.sup.Tdiag(x.sub.t.sup.g){hacek over (h)}.sub.r,t.sup.g+{hacek over (w)}.sub.t.sup.g

[0035] for r=1,2, . . . ,R, where y.sub.r.sup.g=[y.sub.r.sup.g(1) y.sub.r.sup.g(2) . . . y.sub.r.sup.g(N)].sup.T is the vector of the received signals at receive antenna r (13) for subbblock g.sub.t x.sub.t.sup.g=[x.sub.t.sup.g(1) x.sub.t.sup.g(2) . . . x.sub.t.sup.g(N)].sup.T is the OFDM-IM subblock g for transmit antenna t (14), and {hacek over (h)}.sub.r,t=[{hacek over (h)}.sub.r,t.sup.g(1) {hacek over (h)}.sub.r,t.sup.g(2) . . . {hacek over (h)}.sub.r,t.sup.g(N)].sup.T and {hacek over (w)}.sub.r.sup.g=[{hacek over (w)}.sub.r.sup.g(1) {hacek over (w)}.sub.r.sup.g(2) . . . {tilde over (w)}.sub.r.sup.g(M)].sup.T. The use of the block interleaving (10) ensures the subcarriers in a subblock are affected from uncorrelated wireless fading channels for practical values of N.sub.F.

[0036] For the detection of the corresponding OFDM-IM subblocks of different transmit antennas (14), the following MIMO signal model is obtained for subcarrier n of subblock g:

y.sub.n.sup.g=H.sub.n.sup.gx.sub.n.sup.g+w.sub.n.sup.g

for n=1, 2, . . . ,N and g=1,2, . . .,G, where y.sub.n.sup.g is the received signal vector, H.sub.n.sup.g is the corresponding channel matrix which contains the channel coefficients between transmit (6) and receive antennas (7) and assumed to be perfectly known at the receiver, x.sub.n.sup.g is the data vector which contains the simultaneously transmitted symbols from all transmit antennas (6) and can have zero terms due to index selection in each branch of the transmitter and w.sub.n.sup.g is the noise vector. Due to the index information carried by the subblocks of the proposed scheme, it is not possible to detect the transmitted symbols by only processing y.sub.n.sup.k for a given subcarrier n in the MIMO-OFDM-IM scheme. Therefore, N successive MMSE detections (11) are performed for the proposed scheme using the MMSE filtering matrix

[00001]$W_n^g={\left({\left(H_n^g\right)}^H.Math.H_n^g+\frac{I_r}{}\right)}^{-1}.Math.{\left(H_n^g\right)}^H$

[0037] for n=1,2, . . . ,N, where =.sub.x.sup.2/N.sub.0,F, .sub.x.sup.2=K/N and E{x.sub.n.sup.g(x.sub.n.sup.g).sup.H}=.sub.x.sup.2I.sub.T. By the left multiplication of y.sub.n.sup.g with W.sub.n.sup.g, MMSE detection (11) is performed as

z.sub.n.sup.g=W.sub.n.sup.gy.sub.n.sup.g=W.sub.n.sup.gH.sub.n.sup.gx.sub.n.sup.g+W.sub.n.sup.gw.sub.n.sup.g

[0038] where z.sub.n.sup.g=[z.sub.n.sup.g(1) z.sub.n.sup.g(2) . . . z.sub.n.sup.g(T)].sup.T is the MMSE estimate of x.sub.n.sup.g. The MMSE estimate of MIMO-OFDM-IM subblocks {circumflex over (x)}.sub.t.sup.g=[{circumflex over (x)}.sub.t.sup.g(1) {circumflex over (x)}.sub.t.sup.g(2) . . . {circumflex over (x)}.sub.t.sup.g(N)].sup.T can be obtained by rearranging the elements of z.sub.n.sup.g, n=1,2, . . . ,N as {circumflex over (x)}.sub.t.sup.g=[z.sub.1.sup.g(t) z.sub.2.sup.g(t) . . . z.sub.N.sup.g(t)].sup.T for t=1,2,. . . ,T and g=1,2, . . . , G. As mentioned earlier, {circumflex over (x)}.sub.t.sup.g contains some zero terms, whose positions carry information; therefore, independent detection of the data symbols in {circumflex over (x)}.sub.t.sup.g (with linear decoding complexity) is not a straightforward problem for the proposed scheme. As an example, rounding off individually the elements of {circumflex over (x)}.sub.t.sup.g to the closest constellation points (the elements of {0,S} for the proposed scheme) as in classical MIMO-OFDM may result a catastrophic active index combination that is not included in the reference look-up table, which makes the recovery of index selecting n bits impossible.

[0039] In order to determine the active subcarriers in {circumflex over (x)}.sub.t.sup.g, the LLR detector (12) of the proposed scheme calculates the following ratio which provides information on the active status of the corresponding subcarrier index n of transmit antenna t:

[00002]${}_t^g\left(n\right)=\ln .Math.\frac{\underset{m=1}{\overset{M}{.Math.}}.Math.P\left({\hat{x}}_t^g\left(n\right)x_t^g\left(n\right)=s_m\right)}{P\left({\hat{x}}_t^g\left(n\right)x_t^g\left(n\right)=0\right)}$

[0040] for n=1,2, . . . ,N, where s.sub.mS. This calculation requires the conditional statistics of {circumflex over (x)}.sub.t.sup.g(n) (z.sub.n.sup.g(t)). However, due to successive MMSE detection (11), the elements of {circumflex over (x)}.sub.t.sup.g are still Gaussian distributed but have different mean and variance values. Let us consider the mean vector and covariance matrix of z.sub.n.sup.g conditioned on x.sub.t.sup.g(n){0,S}, which are given as

E{z.sub.n.sup.g}=W.sub.n.sup.gH.sub.n.sup.gE{x.sub.n.sup.g}=(W.sub.n.sup.gH.sub.n.sup.g).sub.tx.sub.t.sup.g(n)

cov(z.sub.n.sup.g)=W.sub.n.sup.gH.sub.n.sup.gcov(x.sub.n.sup.g)(H.sub.n.sup.g).sup.H(W.sub.n.sup.g).sup.H+N.sub.0,FW.sub.n.sup.g)(W.sub.n.sup.g).sup.H

where E{x.sub.n.sup.g} is an all-zero vector except its t th element is x.sub.t.sup.g(n), and cov(x.sub.n.sup.g)=diag([.sub.x.sup.2 . . . .sub.x.sup.2 0 .sub.x.sup.2 . . . .sub.x.sup.2]) is a diagonal matrix whose t th diagonal element is zero. Then, the conditional mean and variance of {circumflex over (x)}.sub.t.sup.g(n) are obtained as

E{{circumflex over (x)}.sub.t.sup.g(n)}=(W.sub.n.sup.gH.sub.n.sup.g).sub.t,tx.sub.t.sup.g(n), var({circumflex over (x)}.sub.t.sup.g(n))=(cov(z.sub.n.sup.g)).sub.t,t.

[0041] Using the above found statistics of the MMSE filtered signals, the LLR for the n th subcarrier of t th transmitter for subblock g can be calculated as (12)

[00003]${}_t^g\left(n\right)=\ln \left(\underset{m=1}{\overset{M}{.Math.}}.Math.-\frac{{.Math.{\hat{x}}_t^g\left(n\right)-{\left(W_n^g.Math.H_n^g\right)}_{i,t}.Math.s_m.Math.}^2}{{\left(\mathrm{cov}\left(z_n^g\right)\right)}_{i,t}}\right)+\frac{.Math.{\hat{x}}_t^g\left(n\right).Math.}{{\left(\mathrm{cov}\left(z_n^g\right)\right)}_{i,t}}$

for n=1,2, . . . ,N, t=1,2, . . . ,T and g=1,2, . . . ,G. After the calculation of N LLR values for a given subblock g and transmit antenna t, which results a linear decoding complexity of O(M) per subcarrier as in classical MIMO-OFDM, in order to determine the indices of the active subcarriers, the LLR detector (12) calculates the following LLR sums for c=1,2, . . . ,C according to the look-up table as d.sub.t.sup.g(c)=.sub.k=1.sup.K.sub.t.sup.g(i.sub.k.sup.c), where I.sup.c={i.sub.1.sup.c,i.sub.2.sup.c, . . . ,i.sub.K.sup.c} denotes the possible active subcarrier index combinations. The LLR detector determines the active subcarriers for a given subblock g and transmit antenna t as =arg max.sub.cd.sub.t.sup.g(c) and .sub.t.sup.g={i.sub.1.sup.,i.sub.2.sup., . . . ,i.sub.k.sup.}. The M-ary symbols transmitted by the active subcarriers are determined with ML detection a

.sub.t.sup.g(k)=argmin.sub.s.sub.m.sub.S|{circumflex over (x)}.sub.t.sup.g(i.sub.K.sup.)(W.sub.i.sub.K.sup.g.sup.H.sub.i.sub.K.sup.g.sup.).sub.t,ts.sub.m|.sup.2

for k=1,2, . . . ,K.sub.t, where these metrics were calculated for the LLR values calculated earlier and do not increase the decoding complexity. After this point, index selecting p.sub.1 bits are recovered from the look-up table and M-ary symbols are demodulated to obtain the corresponding p.sub.2 information bits.

[0042] In FIG. 3, we compare the bit error rate (BER) performance of the invention for N=4, K=2 with classical MIMO-OFDM for M=2 at same spectral efficiency values. As seen from FIG. 3, the proposed scheme provides significant BER performance improvement compared to classical MIMO-OFDM, which increases with higher order MIMO systems. As an example, the MIMO-OFDM-IM scheme achieves approximately 10.4 dB better BER performance than classical MIMO-OFDM at a BER value of 10.sup.5 for the 88 MIMO system.

[0043] In FIG. 4 and FIG. 5, we extend our simulations to higher spectral efficiency values and compare the BER performance of the proposed MIMO-OFDM-IM scheme (N=4, K=3) with classical MIMO-OFDM for M=4 and 16, respectively. As seen from FIG. 4 and FIG. 5, the proposed scheme still maintains its advantage over classical MIMO-OFDM in all considered configurations. It is interesting to note that the proposed scheme has the potential to achieve close or better BER performance than the reference scheme, even using a lower order MIMO system in most cases.