Method for decoding symbols transmitted via spatial modulation and device implementing same

Abstract

A method of decoding information encoded in respective activation modes of an activated subset of antennas of a MIMO transmitter transmitting orthogonally modulated signal components, the method comprising: decoupling a real part and an imaginary part of a signal component of a received signal vector, respective channel matrices of real and virtual channels are received, determined, or estimated, and real and virtual signal components are represented as respective products of transmitted symbol components and activation unit vectors. The respective activation unit vectors of the decoupled real and imaginary parts are then represented with corresponding random variable vectors that are dependent on a probability quality function of the activation pattern available at the transmitter. Then, a Gaussian belief propagation-based messaging process is applied to the random variables to jointly estimate activation unit vectors of the respective real and imaginary parts of the received signal vector, and finally the activation unit vectors are decoded to obtain the information.

Description

Technical Field

[0001] This invention relates to wireless communication via spatial modulation (SM), and more particularly to a method for decoding information transmitted via SM, and aims to realize or facilitate integrated sensing and communication (ISAC), also known as joint communication and sensing (JCAS) or joint radar and communication (JRC), in which wireless signals are used simultaneously for communication and environmental sensing, i.e., generating information about the environment and objects therein.

[0002] mark

[0003] In this specification, bold symbols represent vectors or matrices. Scalar values ​​are represented in italic lowercase letters, such as x. Superscript T and H These represent the transpose and complex conjugate transpose of a vector or matrix, respectively. Background Technology

[0004] Integrated Sensing and Communication (ISAC) has recently been recognized as a new key technology for surpassing fifth-generation (B5G) and sixth-generation (6G) wireless communication systems. While ISAC is expected to enable new applications, the fundamental trade-offs between sensing and communication functions will inevitably put pressure on the equally important goal of curbing power consumption in B5G and 6G systems. In fact, B5G and 6G systems not only need to achieve significant improvements over existing systems in terms of energy and spectral efficiency, but also in data throughput, reliability, latency, coverage, and user capacity. This will incentivize research into new technologies such as millimeter-wave (mmWave) / terahertz (THz) communication, massive MIMO and cell-free MIMO (CF-MIMO), reconfigurable smart surfaces (RIS), and indexed modulation (IM).

[0005] In IM (a special case of SM), a transmitter with multiple transmit antennas activates only a subset of available antennas for each transmission instance used for transmitting information symbols and beamforming. To compensate for information loss due to the limited number of transmitted symbols, the information is also encoded in a selection mode for the corresponding activated transmit antennas. If the decoder can determine which antennas are active, it can decode the corresponding information.

[0006] Among the various alternatives mentioned above, Intensive Modular Array (IM) is attractive for future wireless communication systems due to its economical use of transmitter resources and efficient operation through encoding information in a corresponding activated subset of fully available resources. It also leverages information encoded in a combination of "low-energy" or "no-energy" modes in resource activation patterns to achieve high spectral efficiency. The sparse utilization of the radio frequency (RF) chain achieved by IM is particularly attractive for millimeter-wave (mmWave) and terahertz (THz) systems, specifically for reducing the total transmit energy required. The characteristics of IM also add significant flexibility, as the actual transmitted symbols are not directly related to the antenna subset activation scheme itself. This makes it transparent to other promising technologies, as evidenced by the rapidly growing literature on integrating IM with mMIMO, CFMIMO, RIS, and ISAC.

[0007] Typically, an IM scheme that activates P selected modes out of a total of N resource allocation modes for transmission can transmit up to [number] bits of information beyond the information transmitted through the selected P resources themselves. An additional information bit. Therefore, even with a sufficiently large N and a suitable P, a large amount of information can be encoded in the pattern or index domain, making it possible for IM to outperform fully dense (i.e., P=N) systems in certain situations, while offering the additional benefits of improved energy and spectral efficiency and freeing up unused resources for other functions such as opportunistic communication and radar / sense.

[0008] Figure 1 A simplified schematic representation of an IM is shown. Among multiple antennas, at each transmission moment correspondingly defined by virtual vertical lines distributed across the time domain, only a subset of antennas is used for transmission; each subset is unique and therefore carries information itself. Within the IM, beamforming remains possible by appropriately controlling the phase and amplitude of the analog transmitted signal transmitted by each antenna and / or by appropriate precoding.

[0009] Inspired by the advantages of IM, rapid development of enhanced SM designs aimed at maximizing energy and spectral efficiency while minimizing error rate and decoding complexity can be found in recent literature. An excellent example is the Quadrature Spatial Modulation (QSM) scheme contained in “Quadrature Spatial Modulation,” IEEE Transactions on Vehicle Technology, Vol. 64, No. 6, pp. 2738–2742, June 2015, R. Mesleh, SS Ikki, and HM Aggoune, where SM techniques are applied independently to each in-phase and quadrature (IQ) component of the transmitted symbol, such that from a total of N… TP transmit antennas, selected from a total of P transmit antennas, independently transmit the P real parts and P imaginary parts of the symbol. The total number of antenna mode combinations in QSM can be represented by binomial coefficients in combinatorics. The QSM scheme, compared to the original SM scheme discussed below by R.R. Mesleh, H. Haas, S. Sinanovic, C.W. Ahn, and S. Yun, not only doubles the rate of spatially modulated information but also improves bit error rate (BER) performance: “Spatial modulation,” IEEE Transactions on Vehicle Technology, Vol. 57, No. 4, pp. 2228-2241, July 2008. Besides nearly doubling the codebook size and thus doubling the spectral efficiency in the spatial domain, QSM is also well-suited for ISAC applications because the actual transmitted symbols do not need to be used entirely to carry information; instead, some or even all of the symbols can be configured specifically for sensing purposes, for example, using them as pilot signals whose properties are known at the receiver.

[0010] Other examples of this type of design are improved generalized orthogonal spatial modulation (GQSM) designs that incorporate space-time code processing (STC) and other techniques to optimize antenna activation modes, such as those provided by... U. Aygölü, E. Panayirci, and HV Poor discuss “Space-time block coded spatial modulation” in IEEE Transactions on Communications, Vol. 59, No. 3, pp. 823-832, 2011; and HS Rou, GTF Abreu, H. Iimori, D. González G., and O. Gonsa discuss “Scalable quadrature spatial modulation” in IEEE Transactions on Wireless Communications, Vol. 21, No. 11, pp. 9293-9311, 2022.

[0011] In summary, the (G)QSM scheme achieves high bit error rate (BER) performance and high energy and spectral efficiency by activating the modulation information in both the transmitted symbols and the coded combination of subsets of multiple transmit antennas, thus outperforming conventional SM methods in all parameters.

[0012] However, considering the size is The codebook contains all possible resource activation modes. The aforementioned advantages of IM and SM are typically imposed on the receiver in terms of the computational complexity required to detect the activated resource at each given transmission instance.

[0013] Naive decoding of GQSM requires a size of The search is performed in the combined space of , where It refers to the number of transmitting antennas. It is the codebook amplification factor (for classic GQSM). ), It is the number of symbols transmitted, and It is the size of the complex sign constellation. Due to the nature of the binomial coefficients in factorial scaling, as N... T And P grows under the B5GmMIMO system, where N T >16, the total number of combinations in a QSM system grows exponentially. In other words, for a large number of N... T With the large transmitting antenna and the high data rate of large P in large MIMO scenarios, the codebook size becomes extremely large.

[0014] Therefore, this challenge has spurred work on low-complexity decoders for many variants of IM. On the other hand, as the number of transmit antennas increases, the amount of information that can be encoded in the "spatial domain" (i.e., in antenna-activated modes) increases significantly compared to the amount of information that can be encoded in the digital symbol constellation domain. This suggests that it may ultimately be possible to encode transmitted information solely in the antenna domain, where the transmitted symbols themselves can be fixed pilot symbols or specific waveforms used for other functions such as radar and sensing.

[0015] This means that, given the known transmitted symbols, the detector only needs to detect the antenna pattern. However, this still poses a challenge for the decoder because the knowledge of the symbols exists only in a relatively small digital domain, rather than in the real problem antenna combination domain, reducing the search space. For larger antenna array sizes, using conventional maximum likelihood (ML) decoding methods becomes infeasible.

[0016] For example, for search-reduction-based methods (such as lattice decoders and sphere decoders, as discussed by I. Al-Nahhal, E. Basar, OA Dobre, and S. Ikki in “Optimum Low-Complexity Decoder for Spatial Modulation”, IEEE Communications Selected Area Journal, Vol. 37, No. 9, pp. 2001–2013, 2019); by Z. Hu, F. Chen, Y. Liu, S. Liu, H. Yu, and F. Ji in “Low-Complexity Detection for Multiple-Mode OFDM with Index Modulation”, Physical Communications, Vol. 34, pp. 38–47, 2019; or by M. Simarro et al. in “Low Complexity Near-ML Sphere Decoding based on a MMSE ordering for Generalized Spatial Modulation”. (Low-complexity near-ML ball decoding based on MMSE ordering for generalized spatial modulation), IEEE 31st Annual International Workshop on Personal, Indoor and Mobile Radio Communications, 2020, pp. 1-6. Improvements to brute-force ML detection methods can reduce the total decoding complexity proportionally to the total codebook size, but their cost still depends on the combination factor.

[0017] Recently, methods have also been proposed aimed at directly reducing the order of binomial coefficients. For example, in the following article, "The achievable rate analysis of generalized quadrature spatial modulation and apair of low-complexity detectors", IEEE Transactions on Vehicle Technology, Vol. 71, No. 5, pp. 5203-5215, 2022, J. An, C. Xu, Y. Liu, L. Gan, and L. Hanzo, achieve a reduction in the order of complexity by utilizing the Orthogonal Matching Pursuit (OMP) algorithm to operate on the possible non-zero indices of the sparse signal of generalized orthogonal spatial modulation (GQSM). arrive The decrease, of which IM is a specific type of GQSM. This shows that a trade-off between computational complexity and optimal performance can be achieved by adjusting the value of α.

[0018] In the following article, “An efficient vector-valued belief propagation decoder for quadrature spatial modulation,” 56th Asilomar Conference on Signals, Systems and Computers, 2022, pp. 27–31, HS Rou, GTF de Abreu, and T. Takahashi propose a GQSM detector with vector-valued variables using a message-passing (MP) algorithm within the Gaussian belief propagation (GaBP) framework. This eliminates the quadratic exponent in the combination factor by employing a novel decomposition of the GQSM signal into two independent vectors. This reduces the decoding cost to the order of the square root of the binomial coefficients, i.e. It also achieves performance close to that of a fixed-complexity ML solution with a second-order gain.

[0019] Finally, there are non-classical approaches that are independent of the binomial coefficients, including techniques based on machine learning and quantum computing. However, such alternatives have their own limitations, such as requiring extensive offline training or expensive and not widely available quantum computers, respectively.

[0020] Furthermore, the complexity of all known detectors depends on the parameters of the combinatorial space, which is a factorial relationship between the number of P activatable modes from a total of N resources.

[0021] Therefore, it is desirable to provide an improved method for decoding symbols in spatially indexed modulated wireless transmissions that offers low complexity while maximizing spectral and energy efficiency, and allows for optimization of both communication and sensing performance in ISAC applications. Summary of the Invention

[0022] This objective is achieved by the method of claim 1 and the apparatus of claim 4. Claims 7 to 11 respectively provide a communication system, a vehicle equipped with the apparatus according to the invention, a computer program product, and a corresponding computer-readable medium. Advantageous embodiments and advancements are provided in the corresponding dependent claims.

[0023] The specific objective of this invention is to solve ISAC (Inter-Input Multiple Access Conversion) using SM (Signal Signal) in large-scale systems (e.g., B5G massive MIMO). Embodiments of this invention utilize pilot-transmitted SM signals, which, due to their enormous computational complexity, cannot be handled by known receivers in the target large-scale system. The novel decoding method proposed herein enjoys a complexity order independent of the combination factor.

[0024] As mentioned above, numerous IM technologies addressing various problems can be found in the literature, such as integration with other functions (e.g., radar and / or sensing), achievable data rates, decoding complexity, and, of course, the resource domain to be utilized and under what architecture. Without loss of generality, this specification exemplifies the GQSM approach and will utilize its system and signal models in the following derivations and analyses.

[0025] Before describing the invention in more detail, the underlying GQSM system model and transmitter structure will be discussed.

[0026] Consider a point-to-point (P2P) MIMO wireless communication system, where the transmitter and receiver are each equipped with One and One antenna, so that it can be used with the flat fading MIMO channel Transmission information vector The corresponding received signal vector Described as follows:

[0027]

[0028] in It is a complex-valued additive white Gaussian noise (AWGN) vector, which has element-wise variance This makes for To be honest .

[0029] In the GSQM transmitter, data is taken from a size of... Complex constellations of In-phase and quadrature (IQ) components of each transmitted symbol , (in The corresponding emission vector components are sparsely mapped to the following form. and

[0030]

[0031] Thus, the following transmitted signal vector is obtained.

[0032]

[0033] It should be noted that the superscripts R and I represent the real and imaginary counterparts of the corresponding parameters in the general spatial modulation (GSM) signal model, respectively, and the superscript c in the equation denotes either the real or imaginary part whenever the real and imaginary parts are treated the same way. It should be further noted that classical quadrature spatial modulation (QSM) is a specific case of GQSM, where... .

[0034] The real and imaginary parts of the transmitted symbol are independent, meaning they can be transmitted from different antennas in the (G)QSM. Therefore, the positions of the IQ vector components in equation (2) are determined by two independent index vectors. and Description: These two independent index vectors come from a size of The set of possible index vectors K. Specifically, This means only using A binary encodeable subset of possible activation pattern combinations, such that the size of the GQSM IQ activation vector codebook A is determined by... Given. It should also be noted that the position of the symbol component in the vector corresponds directly to the activated antenna element at the transmitter.

[0035] By utilizing the possible combinations of symbol component positions, the GQSM scheme not only transmits data from... China Information encoded by symbols, and also transmitted from... Choose from 10 possible locations Information for each location, respectively targeting and Both.

[0036] Based on the above, from GQSM signals The total information transmitted is given below.

[0037]

[0038] in It is due to the The antenna position of each symbol portion is selected to spatially encode the number of bits, and It is the number of bits digitally encoded by the transmitted symbol.

[0039] Below is a codebook example. For clarity, consider a small system, where... , making the total It is possible to have one activation vector, i.e.

[0040]

[0041] Non-zero elements express or This is because the signal codebook used for the two IQ parts is the same.

[0042] A valid codebook for such a system is a subset A of the codebook mentioned above, containing... An activation vector, for example

[0043]

[0044] It can also be represented by a set of corresponding index vectors, each index vector containing a corresponding activation vector. The index of non-zero elements, i.e.

[0045]

[0046] Understandable, from Obtaining a given codebook A from a larger set of possible activation vectors is an optimization topic, and has been solved under the channel diversity criterion, for example, in “Scalable quadrature spatial modulation” (ibid.), and by N. Ishikawa via QC under the decoding complexity criterion in “Quantum Speedup for Index Modulation”, IEEE Access, Vol. 9, pp. 111114–111124, 2021.

[0047] Given the description of the GQSM transmitted signal in equation (2), the received signal model in equation (1) can be rewritten in the following IQ decoupled form.

[0048]

[0049] in and They represent and The corresponding entity of IQ decoupling, and the channel matrix of IQ decoupling. It is further decomposed into the following two sub-matrices

[0050]

[0051] in and They represent and The effective channel components.

[0052] The real-domain IQ decoupling form shown in equation (5) It forms the basis of most known QSM / GQSM detection algorithms, which typically seek to... and To estimate the effective transmitted signal That is, to seek solutions to the recovery problem.

[0053]

[0054] in It is a valid signal The base is The discrete domain.

[0055] Although the linear recovery problem seems easy to solve, the challenge lies in... The base is discrete domain The unrealistic size, of which It is the floor operation up to the nearest power of 2.

[0056] As can be seen from the latter expression, the codebook size exist Scaling up at a geometric rate and in The scaling factorial rate makes the complexity prohibitively high, even for moderately large MIMO scenarios. This is why, even using the lowest complexity method, simulation results in the known literature only account for a maximum of [number missing]. , .

[0057] Due to the discrete solution domain The problem described by equation (7) is NP-hard, which hinders classical convex optimization approaches and necessitates combinatorial search in a high-dimensional space to find the optimal solution. To overcome this challenge, conventional low-complexity solutions utilizing various properties of GQSM signals have been proposed. For example, The inherent sparsity of the vector can be exploited by compressed sensing (CS)-based methods, and the discrete codebook structure can be used to design search-based methods, such as ball decoders and pruning search decoders. Furthermore, HS Rou et al. recently proposed the vector-valued MP method in "An efficient vector-valued belief propagation decoder for quadrature spatial modulation" (ibid.), where two variables are simultaneously estimated by utilizing the independent and identically distributed (iid) bivariate property of the decoupled vectors in equation (6). and The bivariate vector-valued MP method in “An efficient vector-valued belief propagation decoder for quadrature spatial modulation” (ibid.) is shown to expand the search space from... Reduce to .

[0058] Building upon the above, this invention proposes a novel low-complexity decoder that utilizes a new GQSM system model based on Unit Vector Decomposition (UVD) integrated into the GaBP framework. This achieves [the desired result] with [the desired combination factor]. Extremely low decoding complexity order that is completely unrelated.

[0059] The following description considers a pilot-based GQSM scenario where some or all symbol component values ​​are known at the receiver. Pilot symbols are considered arbitrary and can be used for other functions such as authentication, radar, channel estimation, etc. However, even with known pilot symbols, estimating the unknown indices of symbols in the combination space remains challenging, as seen in equation (2).

[0060] Utilizing the following fact: Each IQ symbolic component and The GQSM transmit vector described in equation (2) In this case, the signal occupies only a single position, meaning it is transmitted from only a single antenna element. The transmitted signal can be rewritten as multiplied by the basic activation vector. The superposition of the symbolic components yields...

[0061]

[0062] Where the activation vector (in )yes identity matrix The Column, its definition The following set of orthogonal unit vectors on

[0063]

[0064] Given the restatement in equation (8), the received signal vector of IQ decoupling in equation (5) becomes

[0065]

[0066] in The linear recovery has been transformed into The problem of joint estimation of activation (unit) vectors. The restatement in equation (9) better highlights the construction of GQSM signals with inherently independent spatial and digital codes.

[0067] Figure 2 and Figure 3 The factor graphs of the prior art GQSM system based on the undecoupled univariate model, as shown in equation (5), and the fully decoupled GQM system, as shown in equation (9), are presented respectively. The factor plot of the multivariate UVD model illustrates the differences between the conventional solution and the solution proposed in this paper.

[0068] Based on the above, a random vector variable 'a' is introduced to model the unit vector, where the discrete uniform prior probability mass function (PMF) is given below.

[0069]

[0070] Where a represents an instance of a, It is the set of events of a, and Denotes the unit impulse function, where if ,but ,otherwise, .

[0071] because The vector variables are instances of variable a, therefore the estimation problem is rewritten in the following UVD form.

[0072]

[0073] Among them random variables For unit vectors respectively Modeling is performed, and for Similarly, this is shown as Figure 3 Factor plot in the diagram.

[0074] Using the above, we can derive the following based on the well-known Gaussian Belief Propagation (GaBP) framework: Figure 3 The objective of operations on the factor graph fits the vector value MP rule. This makes the operation on the factor graph... The joint estimate of the activation vector variables (unit vectors) is perfectly within their respective signal domains, each of which is of size . .

[0075] First, regarding the first Factor nodes (of which ), activate vector variables and (in The soft copy vectors of ) are defined as follows: and Soft copy The corresponding expected error covariance matrix is ​​defined as

[0076]

[0077] The following applies only to real components, that is, to... The derivation is provided because the expressions for the corresponding imaginary components are the same; the difference lies in the superscript. Change to And vice versa.

[0078] When the soft copy is available, the factor node receives the signal. Perform the following soft interference cancellation (IC) operation.

[0079]

[0080] in and They represent from Defined channel components and The OK.

[0081] The latter error term and AWGN term The sum (excluding the true symbolic part) is approximated as a Gaussian scalar via the Central Limit Theorem (CLT), which yields the soft IC symbol with respect to a given activation vector. The following is the conditional probability density function (PDF).

[0082]

[0083] The conditional variance is obtained through the following method.

[0084]

[0085] in .

[0086] Then, for each variable node, the conditional PDF from the connected factor nodes is aggregated to compute extrinsic beliefs with self-interference cancellation. ,follow

[0087]

[0088] in and External beliefs The information vector and precision matrix are given below.

[0089]

[0090] Subsequently, the posterior Bayesian optimal soft copy is calculated from external beliefs using the following method.

[0091]

[0092] The corresponding error covariance matrix is ​​given below.

[0093]

[0094] Equations (13) through (19) describe the estimation of the GQSM, which is restated as equation (11). The first MP iteration of the activation vector yields a refined posterior soft copy vector and the corresponding error covariance matrix. Furthermore, in such a... At the end of the MP iteration, the soft replica vector and error covariance matrix are updated with damping to prevent premature convergence to a local optimum, following the... ,in It is the damping factor, and It represents the number of iterations.

[0095] Next, in order to obtain the activation vectors A hard verdict, in all The information between the factor nodes is aggregated to calculate the consensus belief without self-interference cancellation, i.e., equation (14), which yields the belief. The external consensus PDF.

[0096] Finally, through evaluation of The optimal activation vector estimate is selected from the PDFs of the effective states, i.e.

[0097]

[0098] The following summarizes the proposed decoding method via a novel UVD-GaBP decoder for pilot-based GQSM as described in equations (12) to (20). Reference numbers refer to... Figure 4 The flowchart shown.

[0099] The input to method 100 is the received signal. Effective Channel and Pilot symbols for all p and and noise variance The corresponding output is the estimated activation vector for all p. and .

[0100] After receiving the above input in step 110, in step 120, activation vectors are initialized for all n and p. and A soft copy. In step 130, the corresponding error covariance matrix is ​​calculated via equation (12). and In subsequent iterations, MP iterations are performed on both the real and imaginary components for all n and p until a termination criterion is met, for example, until a predetermined number of iterations have been performed. The iteration continues until the expected convergence of the soft replica is reached—based on the first criterion satisfied. In step 140 of the iterative loop, the received signal is subjected to soft IC via equation (13). In step 150, the conditional variance is calculated via equation (15). and And in step 160, external beliefs are calculated via equation (17). Information vector and and precision matrix and Next, in step 170, the posterior Bayesian optimal soft copy is computed from the external beliefs via equation (18). and In step 180, the corresponding error covariance matrix is ​​calculated via equation (19). and Then, in step 190, the Bayesian optimal soft copy is updated via damping. and Step 200 checks if the termination criterion is met. If not (No branch of step 200), the result obtained in step 190 is used as input for the next iteration in step 140. If met (Yes branch of step 200), the iteration loop terminates, and in step 210, it is evaluated according to equation (20). of The optimal activation vector estimate is selected from the PDFs of the effective states. and Finally, the optimal activation vector is estimated. and It can be used for output in step 220, for example, output to the overall method 10 that calls method 100.

[0101] Figure 5 The method is further illustrated using a self-explanatory frameflow type diagram.

[0102] Simulations demonstrate the performance of the proposed method. Figure 6 It shows in and as well as GQSM simulation results of mMIMO systems under different values, where the BER ratio is based on the SNR per bit. To evaluate performance. Fixed MP parameterization has been applied to all scenarios, where the damping factor is... And the maximum number of MP iterations is The proposed method is referred to as "Prop." in the figure.

[0103] Since the computational complexity of brute-force ML and other conventional decoders is prohibitively high at the system scales under consideration, a Genie-assisted matched filter bound (MFB) is instead introduced as an absolute performance bound (e.g., as discussed by T. Takahashi, S. Ibi, and S. Sampei in “Design of adaptively scaled belief in multi-dimensional signal detection for higher-order modulation”, IEEE Communications Bulletin, Vol. 67, No. 3, pp. 1986–2001, 2019). This absolute performance bound is obtained by providing perfect prior knowledge of the activation vectors and pilot symbols for the UVD-GaBP method.

[0104] Simulations show that, even with the cost-effectiveness of a typical personal computer, the proposed UVD-GaBP decoder exhibits unprecedented performance. and The ability to efficiently demodulate high-speed GQSM signals in an mMIMO setting. It can be observed that in both cases, this is aimed at... Optimal performance is achieved with UVD-GaBP, and is designed for increased performance. In high The area showed approximately Slight performance loss and incorrect planarization. Note the extremely low performance of GQSM. The range, thanks to the fact that no transmit power is needed to encode most of the information, confirms the initial motivation for achieving energy- and spectrally efficient mMIMO. However, it should be noted that in larger systems (where... In this model, the negative impacts on both BER performance and error flattening are reduced, thanks to the increased sparsity in the system and therefore the increased orthogonality in the unit vector random variables.

[0105] The proposed method exhibits low complexity, as will be discussed below. Table I compares the decoding complexity of the proposed UVD-GaBP method with exemplary conventional methods. For fairness, the complexity of symbol-level detection is ignored for conventional methods, since the scenario is considered entirely via pilots.

[0106]

[0107] Table I: Complexity order of various GQSM decoders.

[0108] It can be seen that conventional methods reduce the quadratic combination terms that occur in brute-force ML search. That is, the ordered continuous IC (ROMP-OSIC) decoder based on relaxed orthogonal matched pursuit proposed in "The achievable rate analysis of generalized quadrature spatial modulation and apair of low-complexity detectors" achieves a factorial reduction. (in The upper index of the binomial coefficients is relaxed, and the IQ-decoupled GaBP decoder proposed in "An efficient vector-valued belief propagation decoder for quadrature spatial modulation" (ibid.) eliminates the quadratic factor on the binomial coefficients, and This represents the number of MP iterations. However, conventional methods still retain binomial coefficients, which are practically unscalable for the mMIMO system under consideration.

[0109] On the other hand, the proposed UVD-GaBP decoder enjoys a significantly reduced complexity that is completely independent of the binomial coefficients. This is due to the variable... The fact that the vectors are unit vectors results in highly sparse soft copies and covariance matrices at convergence, further reducing the computational complexity with each MP iteration. This low complexity enables the decoding capability of the GQSM scheme in mMIMO systems significantly larger than conventional methods, as verified in the performance evaluation results.

[0110] In view of the above description, according to a first aspect of the present invention, a method for decoding information that is encoded in the corresponding activation modes of an activated subset of activated subsets of a first plurality of activatable transmit antennas of a multiple-input multiple-output (MIMO) transmitter. Each of the activated antennas transmits a quadrature-modulated signal component of a transmitted symbol. The method includes receiving the quadrature-modulated signal at a second plurality of receive antennas, and then using the received signal vector... This indicates that, further, a copy of at least a subset of the transmitted quadrature-modulated signal components is received, derived from the transmitted signal vector. The method indicates that receiving at least a subset may include retrieving that subset from memory. The subset of transmitted quadrature-modulated signals may include pilot signals and / or signals optimized for non-communication purposes, including signals optimized for purposes such as radar. The method further includes: adjusting the received signal vector... The real and imaginary parts of the signal components are decoupled, and the received signal vector is received, determined, or estimated. The corresponding channel matrix of the channel for the real and imaginary parts of the signal components. and In a further step, the method includes: receiving a signal vector The corresponding real and imaginary signal components of the decoupled real and imaginary parts are represented as the transmitted symbol components. and activation unit vector The corresponding product, and through the corresponding random variable vector For the received signal vector The corresponding activation unit vectors of the decoupled real and imaginary parts A model is constructed whereby the vector of random variables depends on a probability mass function (PMF) of the activation modes available at the transmitter. Then, a message-passing (MP) process based on Gaussian belief propagation (GaBP) is applied to the random variables. To jointly estimate the received signal vector The activation unit vectors of the real and imaginary parts are obtained, and the activation vector estimate output by the GaBP process is decoded to obtain the emitted information.

[0111] In one or more embodiments of the method, the GaBP-based MP process includes: receiving the received signal vector for all p active subsets of transmit antennas. The corresponding decoupled real and imaginary parts of the signal components, and the received signal vector. The channel matrix of the corresponding channel for the real and imaginary parts of the signal components. and Received signal vector The decoupled real and imaginary parts of the signal components are a vector of random variables. and the transmitted signal vector The representation includes at least a subset of the transmitted quadrature-modulated signal components, and the activation vectors of all factor nodes and all variable nodes in the factor graph used to model the system. and A soft copy. The corresponding error covariance matrix is ​​then calculated. and Then, iterative loops are performed on the real and imaginary parts of the signal respectively until the termination criterion is met. This iterative loop includes: calculating the conditional variance (150) after performing soft interference cancellation (IC) on the received signal. and Calculate external beliefs Information vector and and precision matrix and From external beliefs Calculate the posterior Bayesian optimal soft copy and And calculate the corresponding error covariance matrix. and Then, the computed variables are used to update the Bayesian optimal soft copy via damping. and Perform checks to verify if the termination criterion is met; if not, repeat the iterative loop using the previously calculated value as input. If the criterion is met, select and output the optimal activation vector estimate. and .

[0112] Termination criteria may include, for example, a predetermined number of iterations or the Bayesian optimal soft copy. and It converges to a predetermined range or below a predetermined value. For example, when the continuous subsequent iterations of the Bayesian optimal soft copy converge to this value. and When the average change between them is lower than the predetermined value, such a convergence criterion can be satisfied.

[0113] According to a second aspect of the invention, a wireless device is provided, configured to receive orthogonally modulated signal components of transmitted symbols. The wireless device includes a plurality of antennas, associated circuitry for processing radio frequency signals, one or more microprocessors, and associated volatile and non-volatile memories. The elements or components are connected via one or more data lines and / or signal lines or buses. The non-volatile memory stores computer program instructions that, when executed by the one or more microprocessors, configure the elements or components of the receiver to implement or perform one or more embodiments of the method according to the first aspect of the invention.

[0114] In one or more embodiments, the wireless device is further configured to transmit quadrature-modulated signals.

[0115] In one or more embodiments, the circuitry for processing radio frequency signals includes a low-noise amplifier and / or mixer configured to provide a representation of the received signal at an intermediate frequency. The mixer preferably uses the same oscillator signal as the transmitter located in the same location as the receiver in the wireless device. The latter can support environmental awareness using signals reflected from objects and transmitted by an entity including the receiver.

[0116] Two or more wireless devices according to the second aspect of the invention can form a communication system when configured to transmit quadrature-modulated signals.

[0117] The receiver according to the third aspect of the invention can be arranged in a vehicle.

[0118] As those skilled in the art will understand, aspects of the embodiments may be embodied as systems, devices, methods, or program products. Therefore, embodiments may take the form of entirely hardware embodiments, entirely software-implemented embodiments (including firmware, resident software, microcode, etc.), or embodiments combining software and hardware aspects.

[0119] For example, the disclosed embodiments can be implemented as hardware circuitry, including custom-designed very large-scale integration (VLSI) circuitry or gate arrays, off-the-shelf semiconductors (such as logic chips, transistors, or other discrete components). The disclosed embodiments can also be implemented in programmable hardware devices such as field-programmable gate arrays, programmable array logic, programmable logic devices, etc. As another example, the disclosed embodiments may include one or more physical or logical blocks of executable code, which may, for example, be organized as objects, procedures, or functions.

[0120] The method described above can be represented by computer program instructions. Therefore, according to a fourth aspect of the invention, the computer program product includes computer program instructions that, when executed by a microprocessor of a wireless device according to a second aspect of the invention, cause the microprocessor to perform the method according to a first aspect of the invention and accordingly control the hardware and / or software blocks or modules of the wireless device.

[0121] The computer program instructions or code used to perform the operations of the embodiments can be any number of lines and can be written in any combination of one or more programming languages, including object-oriented programming languages ​​(such as Python, Ruby, Java, Smalltalk, C++, etc.), as well as conventional procedural programming languages ​​(such as the "C" programming language, etc.) and / or machine languages ​​(such as assembly language). The code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the latter scenario, the remote computer can be connected to the user's computer via any type of network (including a local area network (LAN), wireless LAN (WLAN), or wide area network (WAN)), or can be connected to the external computer, for example, via the Internet through an Internet service provider (ISP).

[0122] Computer program instructions may be retrievably stored on or transmitted on a computer-readable medium or data carrier. This medium or data carrier may be tangibly or physically embodied, for example, in the form of a hard disk, solid-state drive, flash memory device, etc. However, the medium or data carrier may also include modulated electromagnetic, electrical, or optical signals that are received by the computer through a corresponding receiver and transmitted to and stored in the computer's memory.

[0123] The features, structures, or characteristics described in the embodiments can be combined in any suitable manner. Numerous specific details, such as examples of programming, software modules, user selection, network transactions, database queries, database structures, hardware modules, hardware circuits, hardware chips, etc., are provided in this description to provide a thorough understanding of the embodiments.1 However, those skilled in the art will recognize that the embodiments can be practiced without one or more of the specific details described herein or using other methods, components, materials, etc. In other instances, well-known structures, materials, or operations have not been shown or described in detail to avoid obscuring aspects of the embodiments. References to “an embodiment,” “embodiment,” or similar language throughout the specification mean that a particular feature, structure, or characteristic described in connection with that embodiment is included in at least one embodiment. Therefore, unless expressly specified otherwise, the phrases “in one embodiment,” “in an embodiment,” and similar language appearing throughout the specification may, but not necessarily all, refer to the same embodiment, but rather mean “one or more, but not all, embodiments.” Unless expressly specified otherwise, the terms “comprising,” “including,” “having,” and variations thereof mean “including, but not limited to,” “including.” Unless expressly specified otherwise, the enumeration of items does not imply that any or all items in the item are mutually exclusive. Unless otherwise expressly specified, the terms “a,” “an,” and “the” also mean “one or more.”

[0124] In this specification, where aspects of embodiments are described with reference to schematic flowcharts and / or block diagrams of methods, apparatus, systems, and program products according to embodiments, it should be understood that each block of the schematic flowcharts and / or block diagrams, and combinations of blocks of the schematic flowcharts and / or block diagrams, can be implemented by code. This code can be provided to a processor of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus to produce a machine such that instructions executable via the processor of the computer or other programmable data processing apparatus establish components for implementing the functions / actions specified in the flowcharts and / or block diagrams.

[0125] It should be noted that in some implementations or embodiments, the functions presented in the exemplary embodiments shown in the figures may not be executed in the order shown in the figures. For example, depending on the functionality involved, two consecutively shown blocks may actually be executed substantially concurrently, or the blocks may sometimes be executed in reverse order. Other steps and methods that are functionally, logically, or effectively equivalent to one or more blocks or portions thereof shown in the figures are conceivable.

[0126] The method described above is particularly advantageous in large antenna arrays, where the binomial coefficients increase rapidly with the number of antennas. Other advantageous applications of this method include ISAC / JCAS systems, which benefit from an even larger number of antennas. Here, a large number of antennas whose activation modes can be efficiently detected by the method according to the invention can be used to transmit radar or sense optimized symbols. Thus, the method according to the invention combines the advantages of large-scale IM and ISAC techniques.

[0127] The methods and devices presented in this paper can be used with great advantages in mobile communications, such as vehicle-to-the-world (V2X), high-speed rail communication systems, low Earth orbit (LEO) satellite communications, communications with unmanned aerial vehicles (UAVs) (especially in swarm environments), massive machine-type Internet of Things (IoT) communications (e.g., in wireless factories and other industrial settings), ultra-large-scale MIMO and OFDM systems, and even underwater acoustic communications.

[0128] Although the present invention has been described herein using an exemplary mMIMO IM QSM system, it will be apparent to those skilled in the art who have read and understood this specification that the principles and methods developed and presented above are also applicable to general SM-based systems, including QSM, GSM, GQSM, STC-QSM, and OS-QSM.

[0129] The specific embodiments described below with reference to the accompanying drawings are intended as descriptions of various configurations and are not intended to represent only configurations in which the concepts described herein can be practiced. The detailed description includes specific details and is intended to provide a thorough understanding of the various concepts. However, it will be apparent to those skilled in the art that these concepts can be practiced without these specific details. Specifically, although terms from 3GPP 5G NR may be used in this disclosure to exemplify embodiments herein, this should not be construed as limiting the scope of the invention. Attached Figure Description

[0130] The figures in the accompanying drawings are used to illustrate various aspects of the invention.

[0131] Figure 1 A simplified schematic representation of IM combination is shown.

[0132] Figure 2 The factor plot of a conventional GSQM system model is shown.

[0133] Figure 3 A factor plot of the decoupled UVD GSQM system model according to the present invention is shown.

[0134] Figure 4 A flowchart of the GaBP-based MP process according to the present invention is shown.

[0135] Figure 5 The method is illustrated using a frame flow type diagram.

[0136] Figure 6 It shows in and as well as Simulation results of GQSM performance of mMIMO system under different values, according to BER ratio SNR per bit. To evaluate performance.

[0137] Figure 7 A flowchart illustrating an exemplary method for decoding according to the present invention is shown, the method including or invoking Figure 4 The method, and

[0138] Figure 8 An exemplary block diagram is shown of a device configured to implement an embodiment of the method according to the present invention.

[0139] In the figures, the same or similar elements may be referenced using the same reference numerals. Detailed Implementation

[0140] Figures 1 to 6 This has already been discussed above and will not be elaborated upon again.

[0141] Figure 7 A flowchart of an exemplary method 10 for decoding according to the present invention is shown, the method including or invoking reference Figure 4 Method 100 is discussed. This method decodes information encoded in the corresponding activation modes of an activated subset of activated antennas in a first plurality of activatable transmit antennas of a multiple-input multiple-output (MIMO) transmitter. Each of the activated antennas transmits a quadrature-modulated signal component of the transmitted symbol. In step 12, the quadrature-modulated signal is received at a second plurality of receive antennas, by the received signal vector. In step 14, a copy of at least one subset of the transmitted quadrature-modulated signal components is received, as indicated by the transmitted signal vector. This indicates that, and in step 16, the received signal vector... The real and imaginary parts of the signal components are decoupled. In step 18, the received signal vector is received, determined, or estimated. The corresponding channel matrix of the channel for the real and imaginary parts of the signal components. and Next, in step 20, the received signal vector will be... The corresponding real and imaginary signal components of the decoupled real and imaginary parts are represented as the transmitted symbol components. and activation unit vector The product of. Then, in step 22, through the corresponding random variable vector. For the received signal vector The corresponding activation unit vectors of the decoupled real and imaginary parts Modeling is performed on the random variable vector, which depends on the probability mass function (PMF) of the activation modes available at the transmitter. In step 24, a Gaussian belief propagation (GaBP)-based message passing (MP) process 100 is applied to the random variable. To jointly estimate the received signal vector The corresponding real and imaginary activation unit vectors are obtained, and in step 26, the activation vector estimate output by GaBP process 100 is decoded to obtain the transmitted information.

[0142] Figure 8 A schematic block diagram of an exemplary wireless device 300 according to the present invention is shown. The wireless device 300 is configured to receive orthogonally modulated signal components of transmitted symbols and includes two or more antennas 302, a circuit system 304 for processing radio frequency signals, one or more microprocessors 306, and associated volatile memory 308 and non-volatile memory 310. Various components and elements of the device 300 are communicatively connected via one or more data and / or signal lines or buses 312. The non-volatile memory 310 stores computer program instructions that, when executed by one or more microprocessors 306, configure the components of the wireless device 300 to implement or perform embodiments of the method according to the first aspect of the present invention, as described herein.

[0143] List of reference numerals (part of the instruction manual)

[0144]

Claims

1. A method (10) for decoding information, said information being encoded in corresponding activation modes of an activated subset of activated subsets of a first plurality of activatable transmit antennas of a multiple-input multiple-output (MIMO) transmitter, each of the activated antennas transmitting orthogonally modulated signal components of a transmitted symbol, said method comprising: - The orthogonally modulated signal described in (12) is received at the second plurality of receiving antennas, by the received signal vector ( )express, - Receive a copy of at least a subset of the quadrature-modulated signal components transmitted by (14), generated by the transmitted signal vector ( )express, - For the received signal vector ( The real and imaginary parts of the signal components are decoupled (16). - Receive, determine, or estimate (18) the received signal vector ( The corresponding channel matrix of the channel for the real and imaginary parts of the signal components () , ), - The received signal vector ( The corresponding real and imaginary signal components of the decoupled real and imaginary parts of (20) are represented as the transmitted symbol components. ) and activation unit vector ( The corresponding product of ) - Through the corresponding random variable vector ( ) for the received signal vector ( The corresponding activation unit vectors of the decoupled real and imaginary parts of ) Modeling is performed (22), where the vector of random variables depends on the probability mass function (PMF) of the activation modes available at the transmitter. - Applying the message passing (MP) process (100) based on Gaussian belief propagation (GaBP) to random variables ( ), to jointly estimate the received signal vector ( The corresponding activation unit vectors of the real and imaginary parts of ), and - Decode (26) the activation vector estimate output (220) of the GaBP process to obtain the information.

2. The method (10) of claim 1, wherein the GaBP-based MP process (100) comprises: - For all p active subsets of transmit antennas, receive the received signal vector (110) The corresponding decoupled real and imaginary parts of the signal components, the received signal vector ( The channel matrix of the corresponding channel for the real and imaginary parts of the signal components () , The received signal vector ( The random variable vector (the decoupled real and imaginary parts of the signal components) ), and the transmitted signal vector ( The subset of the transmitted quadrature-modulated signal components represented by ) - Initialize (120) the activation vectors of all factor nodes and all variable nodes in the factor graph used to model the system. , A soft copy of ) as well as - Calculate the error covariance matrix corresponding to (130) ( , ), The method further includes iterative loops performed respectively for the real part and the imaginary part of the signal, wherein - Perform (140) soft interference cancellation (IC) on the received signal. - Calculate the conditional variance (150) , ), - Calculate (160) extrinsic beliefs ( ) information vector ( , ) and precision matrix ( , ), - From the aforementioned external beliefs ( )Calculate (170) posterior Bayesian optimal soft copy ( , ), - Calculate the error covariance matrix corresponding to (180) ( , ), - Bayesian optimal soft copy via damped update (190) , ), - Check if (200) meets the termination criteria, and If the conditions are not met, - Repeat the iteration loop using the previously calculated value as input, or Under the condition that, - Select (210) optimal activation vector estimation ( , ),as well as - Output (220) is the selected activation vector estimate ( , ).

3. In one or more embodiments of the method (10), the subset of the transmitted quadrature-modulated signals includes pilot signals and / or signals optimized for non-communication purposes.

4. A wireless device (300) configured to receive orthogonally modulated signal components of transmitted symbols, the wireless device comprising two or more antennas (302) connected via one or more data and / or signal lines or buses (312), a circuit system (304) for processing radio frequency signals, one or more microprocessors (306), volatile memory (308) and non-volatile memory (310), wherein the non-volatile memory (310) stores computer program instructions that, when executed by the one or more microprocessors (306), configure the components of the wireless device (300) to implement or perform the method as described in any one of claims 1 to 3.

5. The wireless device (300) of claim 4, wherein the device (300) is further configured to transmit a quadrature-modulated signal.

6. The wireless device (300) of claim 4 or 5, wherein the circuit system (304) for processing radio frequency signals includes a low-noise amplifier and / or mixer configured to provide a representation of the received signal at an intermediate frequency.

7. A communication system (400) comprising two or more wireless devices (300) according to any one of claims 5 to 6.

8. A vehicle comprising a wireless device (300) according to any one of claims 4 to 6.

9. Use of the wireless device (300) according to any one of claims 4 to 6 and / or the communication system according to claim 7 and / or the method (10, 100) according to any one of claims 1 to 3.

10. A computer program product comprising computer program instructions that, when executed by a microprocessor of a wireless device (300) according to one or more of claims 4 to 6, cause the wireless device (300) and / or control hardware blocks, modules or components that respectively control the wireless device (200) to perform the method (10, 100) according to one or more of claims 1 to 3.

11. A computer-readable medium or data carrier that can retrievably transmit or store the computer program product as claimed in claim 10.

Images