Intelligent MIMO signal transmission method and system based on distributed cooperation

Abstract

The invention discloses an intelligent MIMO signal transmission method and system based on distributed cooperation, and the detection method comprises the steps: continuously employing different distributed units in an MIMO system to execute preset detection steps until the number of iterations reaches a preset number-of-iterations threshold L; sending the estimation vector subjected to the L times of iteration to a central processing unit, and quantizing the estimation vector by the central processing unit to obtain a final detection signal; the preset detection step comprises the following steps: processing a received radio frequency signal into a baseband signal; performing channel estimation by using the pilot signal to obtain a local channel matrix; based on a multi-step conditional random block Kaczmarz algorithm, updating an estimation vector by using a baseband signal and a local channel matrix, and selecting a next distributed unit based on a multi-step conditional sampling probability; and sending the updated estimation vector to the selected next distributed unit based on the topological structure of the MIMO system. According to the invention, not only can higher performance and lower complexity be realized, but also the required data transmission bandwidth is smaller.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of signal transmission in large-scale MIMO systems, and specifically relates to an intelligent MIMO signal transmission method and system based on distributed cooperation. BACKGROUND

[0002] With the rapid growth of the number of users and data traffic in wireless communication systems, large-scale multiple-input multiple-output (MIMO) technology has become one of the key technologies of the sixth generation (6G) wireless communication system due to its high spectral efficiency, high energy efficiency and high reliability. However, with the increasing number of antennas at the base station side, centralized baseband processing faces problems such as high computational complexity and large data transmission bandwidth requirements. To alleviate the above problems, a variety of distributed signal detection algorithms have been proposed, such as the decentralized baseband processing (DBP) architecture, the partially decentralized (PD) architecture, and the fully decentralized (FD) architecture. However, the existing schemes have the following shortcomings: the ADMM (Alternating Direction Method of Multipliers) and CG (Conjugate Gradient) methods under the DBP architecture have high communication complexity, and the detection schemes under the PD and FD architectures have performance loss and insufficient scalability. The existing recursive algorithms based on daisy chain or other topologies, such as recursive least squares, stochastic gradient descent, and alternating stochastic gradient descent, have certain limitations in practicality and performance, such as restrictions on the number of antennas within the distributed unit or poor convergence. In addition, although the classic Kaczmarz method has low complexity, the traditional method has an error bound that cannot be eliminated for inconsistent linear systems, resulting in performance degradation, and its random variants such as the Randomized Kaczmarz method and the Randomized Block Kaczmarz method have the risk of repeated sampling in the distributed detection environment, which cannot effectively exploit the potential. SUMMARY

[0003] To solve the above problems, the present application proposes an intelligent MIMO signal transmission method and system based on distributed cooperation, which not only achieves high performance and low complexity, but also requires less data transmission bandwidth, and has a globally convergent property that is theoretically proven, and can adapt to different types of distributed MIMO system application scenarios.

[0004] In order to achieve the above technical purposes and achieve the above technical effects, the present application is implemented by the following technical solutions:

[0005] In a first aspect, the present application provides an intelligent MIMO signal transmission method based on distributed cooperation, comprising:

[0006] continuously performing a preset detection step by different distributed units in the MIMO system until the number of iterations reaches a preset iteration threshold ;

[0007] the estimated vector after iterations is sent to a central processor, and the central processor quantizes the estimated vector to obtain a final detection signal ; ;

[0008] wherein the preset detection step comprises:

[0009] processing the received radio frequency signal to obtain a baseband signal;

[0010] performing channel estimation using a pilot signal to obtain a local channel matrix;

[0011] updating the estimated vector using the baseband signal and the local channel matrix based on a multi-step conditional random block Kaczmarz algorithm, and selecting the next distributed unit based on a multi-step conditional sampling probability in the multi-step conditional random block Kaczmarz algorithm;

[0012] based on the topology of the MIMO system, the updated estimated vector is sent to the selected next distributed unit.

[0013] In combination with the first aspect, optionally, the formula of the multi-step conditional sampling probability is:

[0014] ;

[0015] wherein denotes the multi-step conditional sampling probability, denotes the sampling probability of the random block Kaczmarz algorithm, is the memory length of the multi-step conditional sampling, used to control the non-repetition of the distributed unit in continuous iterations.

[0016] In combination with the first aspect, optionally, the updating of the estimated vector using the baseband signal and the local channel matrix comprises:

[0017] calculating the dynamic step size at the th iteration;

[0018] updating the dynamic step size based on the baseband signal, the local channel matrix, and the updated dynamic step size and the second aspect the estimation vector obtained after the first iteration , the updated estimation vector is calculated.

[0019] In combination with the first aspect, optionally, when the dynamic step update is performed, the formula used is:

[0020] ;

[0021] In the formula, is the total number of transmitting antennas in the MINO system, is the total number of receiving antennas in the MINO system, is the number of iterations, is the total number of receiving antennas in a single distributed unit.

[0022] In combination with the first aspect, optionally, the formula for calculating the updated estimation vector is:

[0023] ;

[0024] In the formula, is the updated estimation vector, is the local channel matrix, denotes the pseudo-inverse of the local channel matrix , is the baseband signal, is the index of the distributed unit, , . .

[0025] In combination with the first aspect, optionally, the formula for calculating the final detection signal is:

[0026] ;

[0027] In the formula, is the constellation point rounded to the nearest , is the set of constellation points corresponding to the transmitting signals.

[0028] The second aspect, the present application provides a kind of MIMO signal detection system based on distributed cooperation, including several distributed units and central processing unit;Between each distributed unit and the central processing unit, there is bidirectional interaction link;Each distributed unit includes receiving antenna array, radio frequency processing unit, channel estimation unit and signal detection unit;

[0029] Each distributed unit performs preset detection step, until the number of iterations reaches the preset iteration number threshold ​;

[0030] The central processor receives the The estimated vector of the iteration , and the estimated vector Perform quantification processing to obtain the final detection signal ;

[0031] Wherein, the preset detection steps include:

[0032] The receiving antenna array is used to receive radio frequency signals;

[0033] The radio frequency processing unit processes the received radio frequency signal to obtain a baseband signal;

[0034] The channel estimation unit performs channel estimation using the pilot signal to obtain a local channel matrix;

[0035] The signal detection unit updates the estimation vector based on the multi-step conditional random block Kaczmarz algorithm using the baseband signal and the local channel matrix, selects the next distributed unit based on the multi-step conditional sampling probability in the multi-step conditional random block Kaczmarz algorithm; and sends the updated estimation vector to the selected next distributed unit based on the topology structure of the MIMO system.

[0036] In combination with the first aspect, optionally, the formula for the multi-step conditional sampling probability is:

[0037] ;

[0038] Where, represents the multi-step conditional sampling probability, represents the sampling probability of the random block Kaczmarz algorithm, The memory length of multi-step conditional sampling is used to control the non-repeatability of distributed units in consecutive iterations.

[0039] In combination with the first aspect, optionally, the updating of the estimation vector using the baseband signal and the local channel matrix includes:

[0040] Calculate the Dynamic step size at iteration ;

[0041] Based on baseband signal, local channel matrix, updated dynamic step size , and The estimated vector obtained after iterations , calculate the updated estimate vector.

[0042] In combination with the first aspect, optionally, when performing dynamic step length The formula used in the updating is:

[0043] ;

[0044] In the formula, is the total number of transmitting antennas in the MINO system, is the total number of receiving antennas in the MINO system, is the iteration number, is the total number of receiving antennas in a single distributed unit;

[0045] The calculation formula of the updated estimation vector is:

[0046] ;

[0047] In the formula, is the updated estimation vector, is the local channel matrix, represents the pseudo-inverse of the local channel matrix , is the baseband signal, is the index of the distributed unit, , ;

[0048] The calculation formula of the final detection signal is:

[0049] ;

[0050] In the formula, is the constellation point rounded to the nearest , is the set composed of the constellation points corresponding to the transmitting signals.

[0051] Compared with the prior art, the present application has the beneficial effects that:

[0052] In the present application, the multi-step conditional random block Kaczmarz algorithm introduces a multi-step conditional sampling probability parameter, which can avoid the problem of repeated sampling of the same distributed unit in continuous iterations, and considers the estimation vectors (i.e. iteration results) obtained in the previous multiple iterations, thereby significantly improving the convergence speed of the multi-step conditional random block Kaczmarz algorithm and reducing the error bound.

[0053] Further, the present application designs a dynamic step mechanism, which eliminates the inherent error bound in the multi-step conditional random block Kaczmarz algorithm in the iteration updating process, so that it can accurately converge to the solution of linear detection, and ensures the global convergence of the algorithm.

[0054] Further, the MIMO signal transmission method and system in the application is suitable for various decentralized baseband architectures, such as ring or star topologies, and the number of receiving antennas in each distributed unit can be flexibly selected, thus breaking the limitation of the number of receiving antennas in the distributed unit in the traditional Kaczmarz algorithm.

[0055] In summary, the application achieves excellent performance and balance among performance, complexity and data transmission bandwidth, is widely applicable to distributed large-scale MIMO practical application scenarios, and realizes distributed cooperative intelligent signal transmission. BRIEF DESCRIPTION OF DRAWINGS

[0056] In order to more clearly illustrate the technical solutions in the embodiments of the application or the prior art, the following will briefly introduce the drawings needed in the embodiments. Obviously, the drawings in the following description are only some embodiments of the application, and for those skilled in the art, other drawings can also be obtained from these drawings without creative labor, wherein:

[0057] Figure 1 is a scene schematic diagram of a MIMO system in an embodiment of the application;

[0058] Figure 2 is a structure diagram of a MIMO signal detection system in which the distributed units are in ring distribution;

[0059] Figure 3 is a structure diagram of a MIMO signal detection system in which the distributed units are in star distribution;

[0060] Figure 4 is a convergence diagram of a MIMO signal transmission method of the application;

[0061] Figure 5 is a comparison diagram of the bit error rate change trend of a MIMO signal transmission method of the application and various detection schemes in a 16-QAM 128x16 MIMO system;

[0062] Figure 6 is a comparison diagram of the bit error rate change trend of a MIMO signal transmission method of the application and various detection schemes in a 16-QAM 128x32 MIMO system;

[0063] Figure 7 is a comparison diagram of the complexity change trend of a MIMO signal transmission method of the application and various detection schemes in a 16-QAM 256xK MIMO system;

[0064] ​​​Figure 8 is one of the application scenarios of the MIMO signal transmission method.

[0065] Figure 9 is another application scenario of the MIMO signal transmission method. DETAILED DESCRIPTION

[0066] The technical solutions in the embodiments of the present application will be clearly and completely described with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0067] In addition, if the present application embodiments involve descriptions such as "first", "second", etc., the descriptions of "first", "second", etc. are only for the purpose of description, and cannot be understood as indicating or implying the relative importance of the indicated technical features or implicitly indicating the number of the indicated technical features. Therefore, the features limited by "first", "second" can explicitly or implicitly include at least one of the features. In addition, the technical solutions of each embodiment can be combined with each other, but it must be based on the fact that a person of ordinary skill in the art can realize it. When the combination of technical solutions contradicts each other or cannot be realized, it should be considered that the combination of technical solutions does not exist and is not within the protection scope of the present application.

[0068] Embodiment 1

[0069] The present application embodiment provides an intelligent MIMO signal transmission method based on distributed cooperation, comprising the following steps:

[0070] Continuously performing a preset detection step by using different distributed units in the MIMO system until the number of iterations reaches a preset iteration threshold

[0071] The estimated vector after times of iterations is sent to a central processor, and the central processor quantizes the estimated vector to obtain a final detection signal

[0072] The preset detection step comprises:

[0073] processing the received radio frequency signal to obtain a baseband signal;

[0074] performing channel estimation by using a pilot signal to obtain a local channel matrix;

[0075] ​​​Based on a multi-step conditional random block Kaczmarz algorithm, the baseband signal and the local channel matrix are used to update the estimation vector, and the next distributed unit is selected based on the multi-step conditional sampling probability in the multi-step conditional random block Kaczmarz algorithm;

[0076] Based on the topology of the MIMO system, the updated estimation vector is sent to the next distributed unit selected.

[0077] In a specific implementation of the embodiment of the present invention, the formula for the multi-step conditional sampling probability is:

[0078] ;

[0079] Where, represents the multi-step conditional sampling probability, represents the sampling probability of the random block Kaczmarz algorithm, is the memory length of multi-step conditional sampling, which is used to control the non-repeatability of distributed units in continuous iterations. The convergence of the MIMO signal transmission method in the embodiment of the present invention is the best when .

[0080] In a specific implementation of the embodiment of the present invention, the updating of the estimation vector using the baseband signal and the local channel matrix includes:

[0081] Calculate the first Dynamic step size at iteration ;

[0082] Based on baseband signal, local channel matrix, updated dynamic step size , and The estimated vector obtained after iterations , calculate the updated estimate vector.

[0083] In a specific implementation of the embodiment of the present invention, when performing dynamic step length When updating, the formula used is:

[0084] ;

[0085] Where, is the total number of transmitting antennas in the MINO system, is the total number of receiving antennas in the MINO system, is the number of iterations, is the total number of receiving antennas in a single distribution unit.

[0086] In a specific implementation of the embodiment of the present invention, the calculation formula of the updated estimated vector is:

[0087] ;

[0088] Where, is the updated estimate vector, is the local channel matrix, represents the local channel matrix The pseudo-rebellion, is the baseband signal, is the index of the distributed unit, , .

[0089] In a specific implementation of the embodiment of the present invention, the final detection signal The calculation formula is:

[0090] ;

[0091] Where, Round to the nearest The constellation point, for The set of constellation points corresponding to the transmitted signals.

[0092] The MIMO signal transmission method in the embodiment of the present invention is mainly used in the uplink signal detection scenario of a large-scale MIMO system. The transmitting end is composed of multiple terminal user devices, including Antenna, the receiving end is equipped with The base station with root antennas, K≤N, such as Figure 1 The MIMO signal transmission method in the embodiment of the present invention can be applied to cooperative intelligent distributed communication networks that meet this basic structure, such as non-cellular MIMO, Internet of Vehicles, and Industrial Internet of Things.

[0093] In a specific implementation process, the MIMO signal transmission method in the embodiment of the present invention includes the following steps:

[0094] Step 1: For a Root transmitting antenna and MIMO system with 1 receiving antenna, the base station side The antennas are divided into distributed units, and record these distributed units as index sets , each distributed unit contains Antenna, satisfy The channel estimation unit in each distributed unit uses the pilot to perform channel estimation, and the dimension is The local channel matrix . In each distributed unit The radio frequency signal received by the root antenna is processed by the radio frequency processing unit to obtain a baseband signal .

[0095] Step 2: Parameter initialization is performed on the multi-step conditional stochastic block Kaczmarz algorithm. In each distributed unit, the estimated vector and the dynamic step size are initialized, and the number of iterations is set. The specific operation steps are as follows:

[0096] Step 2.1: Set the initial value of the estimated vector to , and the dimension is .

[0097] Step 2.2: Set the initial value of the dynamic step size to .

[0098] Step 2.3: Set the iteration threshold value .

[0099] Step 3: Based on the multi-step conditional stochastic block Kaczmarz algorithm, the estimated vector is continuously updated using the local channel matrix and the baseband signal. The specific operation steps are as follows:

[0100] Step 3.1: Calculate the multi-step conditional sampling probability, and select the next distributed unit based on the multi-step conditional sampling probability. The calculation formula of the multi-step conditional sampling probability is:

[0101]

[0102] In the formula, denotes the multi-step conditional sampling probability, denotes the sampling probability of the stochastic block Kaczmarz algorithm, is the iteration step size.

[0103] Step 3.2: In each distributed unit, the multi-step conditional stochastic block Kaczmarz algorithm is used to update the estimated vector and the dynamic step size . The specific calculation steps are as follows:

[0104] Step 3.2.1: At the th iteration, the dynamic step size is updated, and the update formula is:

[0105]

[0106] Step 3.2.2: Based on the last iteration result and the updated dynamic step size The estimated vector is iteratively updated as follows:

[0107]

[0108] wherein, is the updated estimated vector, is the local channel matrix, represents the pseudo-inverse of the local channel matrix , is the baseband signal, is the index of the distributed unit, , .

[0109] Step 3.3 compares the size of and , if , jump to step 3.1. If , jump to step 3.4.

[0110] Step 3.4, the estimated vector after iterations is obtained, which is passed to the central processor and step 4 is performed.

[0111] Step 4: the central processor quantizes the vector , rounds by rounding based on the constellation , obtains as the final output:

[0112]

[0113] wherein, is rounded to the nearest constellation point .

[0114] The MIMO signal transmission method in the embodiment of the application is further described below through specific embodiments.

[0115] As shown in FIG. 1, the MIMO system includes Figure 1 transmit antennas and receive antennas, the relationship between the signals received by the receive antennas and the signals sent by the transmit antennas is shown in the following formula:

[0116]

[0117] wherein, represents the channel matrix, the dimension is ; ​denotes the Gaussian white noise interference with dimension , which obeys the complex Gaussian distribution with mean and variance . denotes the sending signal sent by the user side with dimension , wherein each element belongs to the constellation point set modulated by M-QAM. denotes the receiving signal on the base station side with dimension . In order to recover the sending signal in such a MIMO system , the optimal maximum likelihood (ML) detection aims to solve the following integer least square (ILS) problem:

[0118]

[0119] Such a problem is NP-hard and is difficult to solve in theory. The existing linear detection scheme transforms the problem into solving the following least square (LS) problem under the condition that

[0120]

[0121] The traditional linear detection scheme can well approximate the solution of the ML detection, and the estimated signal is:

[0122] and

[0123] Finally, according to the modulation constellation quantize and to obtain the final decision and :

[0124] and

[0125] However, the calculation complexity of the matrix inversion operation involved in such a linear scheme is , which cannot be used in large-dimensional systems.

[0126] As shown in Figure 2 and Figure 3 , in the embodiment of the present application, the base station side is connected using a distributed framework such as a ring or a star. The root receiving antennas on the base station side are evenly divided into distributed units, and the distributed units are recorded as an index set . Each distributed unit contains root antennas, satisfying ​In each distributed unit, The RF signal received by the antenna is processed by the RF processing unit through amplification, filtering, down-conversion and analog-to-digital conversion to obtain the baseband signal (i.e. digital baseband received signal). The channel estimation unit performs local channel estimation through the pilot signal and obtains the dimension The local channel matrix , and sent to the signal detection unit. The signal detection unit uses the local channel estimation result and baseband signals , combined with the estimated vector from the previous distributed unit , based on the multi-step conditional random block Kaczmarz algorithm proposed in this invention, signal detection is performed. The final estimated vector in the distributed unit at iteration The signal is transmitted to the central processing unit, which completes the channel decoding task through the corresponding quantization operation to obtain the final detection signal. . Figure 2 and Figure 3 In the figure, RF represents the radio frequency unit, Detection represents the signal detection unit, and CHEST represents the channel estimation unit.

[0127] Below is a transmitting antenna , receiving antenna , the number of distributed units Taking the distributed massive MIMO system as an example, the specific implementation of the present invention is described.

[0128] Step 1: For a Root transmitting antenna and MIMO system with 10 receiving antennas, the base station side The antennas are divided into distributed units, and record these distributed units as index sets , each distributed unit contains Antenna, satisfy Each distributed unit uses the pilot to perform channel estimation, and the dimension is The local channel matrix . Base station The RF signal received by the root antenna is processed by the RF processing unit to obtain the received baseband signal For the transmitting antenna , receiving antenna Distributed massive MIMO systems, The distributed units are each assigned to Root receiving antenna, Local channel matrix of distributed units and local received signals , 1, 2, 3, …, 16, can be expressed as:

[0129]

[0130] wherein, has a dimension of , has a dimension of . denotes the channel response between the th transmit antenna and the th receive antenna, denotes the received data signal on the th receive antenna.

[0131] Step 2: Initialization based on the multi-step conditional random block Kaczmarz algorithm. In each distributed unit, initialize the estimated vector and the dynamic step size , and set the number of iterations . The specific operation steps are as follows:

[0132] Step 2.1 Set the initial value of the estimated vector to a zero vector with a dimension of .

[0133]

[0134] Step 2.2 Set the initial value of the dynamic step size to

[0135]

[0136] Step 2.3 Set the number of iterations , which can be set to .

[0137] Step 3: Based on the multi-step conditional random block Kaczmarz algorithm, the estimated vector is constantly updated using the local channel matrix and the baseband signal. The specific operation is as follows:

[0138] Step 3.1 Select the next distributed unit for detection to obtain the corresponding estimated vector. Here, the selection of the next distributed unit is based on the multi-step conditional sampling probability:

[0139]

[0140] Here, denotes the sampling probability of the general random block sampling method.

[0141] Step 3.2: In each distributed unit, use the multi-step conditional random block Kaczmarz algorithm to estimate the vector and dynamic step size To update, is the current number of iterations. The specific calculation steps are as follows:

[0142] Step 3.2.1: After each iteration, At the iteration, the dynamic step size To perform such an update, you can use the following methods:

[0143]

[0144] Step 3.2.2: Based on the results of the previous iteration And the updated dynamic step size , iteratively update the estimated vector:

[0145]

[0146] in, represents the channel matrix Pseudo-reversal.

[0147] Step 3.3 Comparison and The size of t , skip to step 3.1. , then skip to step 3.4.

[0148] Step 3.4 now gets the result The estimated vector of the iteration , pass it to the CPU and execute step 4.

[0149] Step 4: CPU quantizes the vector , by constellation-based Rounding to the nearest integer ,get As the final output:

[0150]

[0151] in, Round to the nearest constellation points.

[0152] Since the MIMO signal transmission method of the embodiment of the present invention is an iterative algorithm, its complexity mainly depends on the complexity of the iterative formula. In the iterative formula of the embodiment of the present invention, the main computational complexity lies in the matrix pseudo-inverse. Calculation and operation , and the total computational complexity of each iteration is . Therefore, in the case of , the overall computational complexity does not exceed , which is superior to most existing distributed detection algorithms, and has certain advantages in terms of bandwidth cost compared with existing distributed detection algorithms.

[0153] Since the MIMO signal transmission method of the embodiment of the application belongs to a distributed algorithm, the bandwidth of the interconnection link needs to be considered. The data transmission bandwidth is determined by the average complex value transmitted on each link, which can reflect the actual overhead of the hardware interface. In the application, since only the vector needs to be transmitted by the distributed unit in one direction, the data bandwidth requirement is only .

[0154] Theoretical derivation shows that for an inconsistent linear system , given the multi-step conditional sampling probability on the fixed partition , the convergence of the multi-step conditional random block Kaczmarz algorithm proposed in the embodiment of the application is described as:

[0155] ;

[0156] wherein, E ( ) represents the mean square error of the algorithm using the multi-step conditional sampling probability . And:

[0157] ;

[0158] wherein, and respectively represent the square of the minimum and maximum singular values of the corresponding matrix. And:

[0159] ;

[0160] Based on such a conclusion, in order to achieve the best convergence performance and error bound, the optimal selection of in the multi-step conditional random block Kaczmarz algorithm proposed in the application is the maximum , which can be specifically written as:

[0161] ;

[0162] wherein, represents the convergence coefficient of each algorithm, and the larger it is, the faster the convergence speed is. And:

[0163] ;

[0164] in, is the lower limit of the minimum singular value of the row block matrix, which reflects the error limit of each algorithm. The larger it is, the smaller the error limit is. For example, in a In an uncoded massive MIMO system with Rayleigh fading channels, The intuitive results are as follows. Figure 4 As shown, the convergence coefficient along with Increased with the increase of .

[0165] When using multiple algorithms such as ADMM (Alternating Direction Method of Multipliers), CG (Conjugate Gradient), MMSE (Minimum Mean Square Error), CD (coordinate descent), DN (decentralized Newton), RLS (Recursive Least Square), SGD (Stochastic Gradient Descent), ASGD (Averaged Stochastic Gradient Descent), SDK (standard distributed Kaczmarz), etc. to perform the multi-step conditional random block Kaczmarz algorithm (MRCBK- ) as an alternative, as these algorithms either have very high computational complexity or data bandwidth costs, or limit the number of distributed unit antennas, making them inflexible. Taking into account the comparison of detection performance, computational complexity, and transmission bandwidth, the Multi-Step Conditional Random Block Kaczmarz Algorithm (MCRBK) proposed in this paper achieves the best compromise between multiple constraints and performance. Figures 5-7 This can be proved by Figure 6 The MRCBK in this example refers to the absence of dynamic step size updates. To demonstrate the superiority of the MIMO signal transmission method in this embodiment of the present invention, the algorithm of the present invention is compared with other existing solutions in terms of complexity, bandwidth, number of antennas, global convergence, and distributed architecture, as shown in the following table: .

[0166] In the table: Computational Complexity represents complexity; Data Bandwidth represents bandwidth; Flexible number of antennas in each DU represents whether the number of receiving antennas in each distributed unit can be flexibly set; Global Convergence represents global convergence; Decentralized Architecture represents distributed architecture. Ring represents ring architecture; Star represents star architecture, and Daisy-chain represents daisy-chain architecture. represents the number of iterations.

[0167] The specific application of the MIMO signal transmission method in the embodiments of the application is described in detail below.

[0168] Compared with the traditional massive MIMO system, the extremely large MIMO (XL-MIMO) can provide higher spectral efficiency. However, with the extremely large increase in the antenna dimension, the users in the XL-MIMO system are mainly located in the near-field area of the base station, and such a near-field propagation scenario makes the modeling of the system change compared with the traditional massive MIMO. Moreover, due to the non-stationarity of electromagnetic wave propagation, the signal propagation mode and detection requirement in the XL-MIMO system become more complex. Specifically, due to the near-field electromagnetic propagation, the signal transmitted by the user can only reach a part of the antenna array of the base station, resulting in that each UE (user equipment) has a corresponding visible region (VR). Due to these reasons, in the case of XL-MIMO, the antenna array is divided into small, separable and flexible sub-arrays, and then distributed computing is performed in these processing units, which can further improve the system performance and flexibility.

[0169] The MIMO signal transmission method in the embodiments of the application can be easily applied to XL-MIMO in a daisy-chain manner. As shown in FIG. 1, if the processing of the sub-array outside the visible region is ignored, only the active sub-array (i.e., the distributed unit) in the visible region needs to be sequentially processed. Moreover, the more active sub-arrays in the visible region, the more distributed units can be considered, and the better the detection performance. Specifically, each distributed unit is only responsible for processing the signal data in the sub-array where it is located, and then uses the multi-step conditional stochastic block Kaczmarz algorithm to quickly and effectively recover the signal. Through the adjustment of the dynamic step size, the best detection result can be achieved in fewer iterations, and the computational burden is reduced, thereby realizing efficient signal processing and flexible architecture adaptability. Figure 8

[0170] ​Such a distributed cooperation framework solves the problem of changing propagation scenarios in XL-MIMO systems to some extent. However, in XL-MIMO systems, the visible area of users changes dynamically, resulting in significant differences in channel quality between different distributed units, further increasing the complexity of signal detection and the resource requirements of the system. To optimize detection efficiency and improve the adaptability and resource utilization of XL-MIMO systems, the present invention proposes an intelligent dynamic scheduling mechanism that dynamically selects distributed units involved in calculation based on real-time channel state information (CSI) and signal quality feedback, thereby achieving optimal cooperation.

[0171] Specifically, the system periodically collects channel state information in the coverage area of each distributed unit, including signal quality indicators such as signal-to-noise ratio, path loss, and multipath delay, and constructs a channel quality prediction model based on historical data. These data will be transmitted to the central controller, which uses a pre-set priority algorithm (such as a weighted comprehensive evaluation method) to calculate the priority score of each distributed unit in real time, to determine the participation priority of each distributed unit. Intelligent scheduling decisions driven by machine learning are used to deploy lightweight neural network models to learn the correlation between channel quality and detection performance using historical data. Through this process, the system can predict the channel attenuation trend of each distributed unit in the future several iteration periods, thereby dynamically generating the optimal distributed unit scheduling sequence. For example, the system will preferentially schedule distributed units with stable channels and less interference, while avoiding subarrays that are severely affected by instantaneous interference.

[0172] To further improve scheduling efficiency, the system will introduce an adaptive exclusion mechanism to ensure that the system remains efficient in unstable environments. When the detection result error of a distributed unit exceeds the set threshold for consecutive multiple iterations, the system will automatically trigger the exclusion process, suspend the participation of the distributed unit, and fill the task vacancy through data sharing and collaborative calculation between adjacent distributed units to ensure the continuity and reliability of the entire detection process. Through this intelligent dynamic scheduling mechanism, the present invention can achieve efficient signal detection in XL-MIMO systems, ensuring that each distributed unit intelligently adjusts its task allocation based on the current channel quality and system load, thereby improving the overall performance and resource utilization efficiency of the system.

[0173] Cell-Free large-scale MIMO networks remove the physical boundaries of traditional base stations and adopt user-centric data transmission, enabling multiple access points (APs) to work cooperatively and provide services to the same user. For example, Figure 9As shown, the same user in the cell-free massive MIMO network can be served by multiple access points within a geographical area, which are connected to a central processor through dedicated fronthaul links. Therefore, the cell-free massive MIMO network can be seen as the intersection of massive MIMO, distributed MIMO and borderless cell. In this way, the overall performance of the system can be improved, especially in an environment where the user location is not fixed and the channel changes greatly. Since cooperative computation and information transmission are needed between multiple access points, how to efficiently allocate computing tasks and ensure the real-time performance of signal detection becomes a key problem. Moreover, since the distribution density and load of access points will change with the dynamic changes of the network environment, the traditional static cooperation strategy is difficult to cope with this complex and real-time changing environment. Based on the above problems, the present application proposes an intelligent cooperation framework based on distributed cooperative detection to optimize the resource allocation and signal detection efficiency in the cell-free MIMO network.

[0174] The MIMO signal transmission method in the embodiment of the present application can be applied to the cell-free massive MIMO network in two ways. The first way is that each access point can be used as a distributed unit and send its estimation of the transmission signal to the central processor. Then, the central processor transmits information to another access point for further processing, and so on. The other way is that the access point can also send its local information, i.e. channel estimation and baseband signal, directly to the central processor for detection. The computational complexity of this way is lower. No matter which way is adopted, since the distributed processing reduces the computing pressure of each access point, the throughput of the overall network is improved. Moreover, in the signal detection of each access point, the dynamic step mechanism enables each iteration to converge more quickly, ensuring a good balance between the accuracy and real-time performance of the detection. Accordingly, the borderless nature of the cell-free network also enables the detection algorithm of the present application to adapt flexibly to different network topologies and load changes, further improving the flexibility and scalability of the network.

[0175] Based on such cooperative detection, an intelligent cooperation framework is proposed for the signal detection problem of the cell-free MIMO network. The core of the framework is a distributed intelligent decision layer, each access point is equipped with a local decision module, which monitors the computing load, channel capacity and user demand (such as quality of service QoS level) of itself in real time. Based on the reinforcement learning algorithm, the access point can autonomously generate a local task allocation strategy, so as to dynamically adjust the allocation of computing resources when the network conditions change. For example, when high-priority users access, the access point will preferentially allocate the signals of these users to idle computing units for processing, to ensure that high-priority tasks are processed in time. In addition, in the global cooperative optimization layer, the central processor collects the state data of each access point through the front-end link, and models the topological relationship and interference mode between the access points by using a graph neural network (GNN). Based on the output of the GNN, the system can dynamically adjust the signal transmission path between the access points and the weight of the computing task. For example, in areas with high user density, the system intelligently increases the number of cooperative access points to reduce the computing pressure of a single access point; and when the channel conditions change suddenly, the system can quickly switch to a backup access point cluster to ensure the stability and efficiency of signal detection. In order to realize flexible resource allocation, the present application introduces a digital twin simulation module, which uses simulation technology to predict the impact of different cooperation strategies. The module constructs a virtual twin based on real-time network state, simulates the impact of different cooperation strategies on throughput, delay and bit error rate.

[0176] By introducing the hierarchical intelligent cooperation framework and the digital twin simulation, the cell-free MIMO network can realize efficient task cooperation and resource flexible allocation across access points, greatly improving the cooperation efficiency and resource utilization between access points. The present application ensures that the system can adaptively optimize the signal detection process under different network topologies and channel conditions, improves the overall network performance, and provides an extensible technical foundation for future ultra-dense network deployment.

[0177] Embodiment 2

[0178] The embodiment of the present application provides a MIMO signal detection system based on distributed cooperation, which comprises a plurality of distributed units and a central processing unit; there are bidirectional interaction links between each distributed unit and the central processing unit; each distributed unit comprises a receiving antenna array, a radio frequency processing unit, a channel estimation unit and a signal detection unit;

[0179] Each distributed unit executes a preset detection step until the number of iterations reaches a preset iteration threshold , the preset detection step comprises:

[0180] The receiving antenna array is used for receiving radio frequency signals;

[0181] The radio frequency processing unit processes the received radio frequency signal to obtain a baseband signal;

[0182] The channel estimation unit estimates a channel by using a pilot signal to obtain a local channel matrix;

[0183] The signal detection unit updates an estimation vector by using the baseband signal and the local channel matrix based on a multi-step conditional random block Kaczmarz algorithm, selects a next distributed unit based on a multi-step conditional sampling probability in the multi-step conditional random block Kaczmarz algorithm, and sends the updated estimation vector to the selected next distributed unit based on a topology of the MIMO system;

[0184] The central processor receives the updated estimation vector after the (n-1)th iteration , quantizes the estimation vector , and obtains a final detection signal .

[0185] In one specific embodiment of the present application, the formula of the multi-step conditional sampling probability is as follows:

[0186] ;

[0187] In the formula, Pn represents the multi-step conditional sampling probability, P represents a sampling probability of the random block Kaczmarz algorithm, and n represents a memory length of the multi-step conditional sampling, which is used to control the non-repetition of the distributed unit in continuous iterations. It is verified that the convergence of the MIMO signal transmission method in the embodiment of the present application is the best when n is equal to 2. In one specific embodiment of the present application, the updating of the estimation vector by using the baseband signal and the local channel matrix comprises the following steps.

[0188] In the formula, n represents the iteration number, and n represents the iteration number after the updating of the dynamic step size.

[0189] The dynamic step size after the nth iteration is calculated. ;

[0190] The updated estimation vector is calculated based on the baseband signal, the local channel matrix, the updated dynamic step size , and the estimation vector obtained after the (n-1)th iteration.

[0191] In one specific embodiment of the present application, the formula used when the dynamic step size is updated is as follows:

[0192] ​​​ ;

[0193] wherein, is the total number of transmitting antennas in the MINO system, is the total number of receiving antennas in the MINO system, is the iteration number, is the total number of receiving antennas in a single distributed unit;

[0194] The calculation formula of the updated estimation vector is:

[0195] ;

[0196] wherein, is the updated estimation vector, is the local channel matrix, denotes the pseudo-inverse of the local channel matrix , is the baseband signal, is the index of the distributed unit, , ;

[0197] The calculation formula of the final detection signal is:

[0198] ;

[0199] wherein, is the constellation point rounded to the nearest , is the set of constellation points corresponding to the transmit signals.

[0200] The remaining parts are the same as those of Embodiment 1.

[0201] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage, etc.) containing computer-usable program code.

[0202] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 one or more flow or blocks. Figure 1 one or more flow or blocks.

[0203] These computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture including instructions which implement the function specified in the flowchart block or blocks. Figure 1 one or more flow or blocks. Figure 1 one or more flow or blocks.

[0204] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 one or more flow or blocks. Figure 1 one or more flow or blocks.

[0205] The embodiments of the present application described above are merely intended to illustrate the principles and main features of the present application. It should be understood by those skilled in the art that the present application is not limited to the above-described embodiments and that various modifications and improvements can be made without departing from the spirit and scope of the present application. Such modifications and improvements are also intended to fall within the scope of the present application. The scope of the present application is defined by the appended claims and the equivalents thereof.

[0206] The basic principles and main features of the present application and the advantages of the present application have been shown and described above. It should be understood by those skilled in the art that the present application is not limited to the above-described embodiments and that the above-described embodiments and descriptions in the specification are merely illustrative of the principles of the present application. Various changes and modifications can be made without departing from the spirit and scope of the present application, and such changes and modifications are intended to fall within the scope of the present application. The scope of the present application is defined by the appended claims and the equivalents thereof.

Claims

1. A distributed collaborative intelligent MIMO signal transmission method, characterized in that: include: Continuously use different distributed units in the MIMO system to perform the preset detection steps until the number of iterations reaches the preset iteration threshold ; will pass The estimated vector of the iteration Sent to the CPU, the CPU quantizes the estimated vector , and obtain the final detection signal ; Wherein, the preset detection steps include: Process the received radio frequency signal to obtain a baseband signal; Use pilot signals to perform channel estimation and obtain the local channel matrix; Based on a multi-step conditional random block Kaczmarz algorithm, the baseband signal and the local channel matrix are used to update the estimation vector, and the next distributed unit is selected based on the multi-step conditional sampling probability in the multi-step conditional random block Kaczmarz algorithm; Based on the topology of the MIMO system, the updated estimation vector is sent to the next distributed unit selected.

2. The distributed collaborative intelligent MIMO signal transmission method according to claim 1, characterized in that: The formula for the multi-step conditional sampling probability is: ； Where, represents the multi-step conditional sampling probability, represents the sampling probability of the random block Kaczmarz algorithm, is the memory length of multi-step conditional sampling, which is used to control the non-repeatability of distributed units in consecutive iterations; For the Iterations.

3. The distributed collaborative intelligent MIMO signal transmission method according to claim 1, wherein: The updating of the estimation vector using the baseband signal and the local channel matrix includes: Calculate the first Dynamic step size at iteration ; Based on baseband signal, local channel matrix, updated dynamic step size , and The estimated vector obtained after iterations , calculate the updated estimate vector.

4. The distributed collaborative intelligent MIMO signal transmission method according to claim 3, characterized in that: In dynamic step When updating, the formula used is: ； Where, is the total number of transmitting antennas in the MINO system, is the total number of receiving antennas in the MINO system, is the number of iterations, is the total number of receiving antennas in a single distribution unit.

5. The distributed collaborative intelligent MIMO signal transmission method according to claim 3, wherein: The calculation formula of the updated estimated vector is: ； Where, is the updated estimate vector, is the local channel matrix, represents the local channel matrix The pseudo-rebellion, is the baseband signal, is the index of the distributed unit, , .

6. The distributed collaborative intelligent MIMO signal transmission method according to claim 1, characterized in that: The final detection signal The calculation formula is: ； Where, Round to the nearest The constellation point, for The set of constellation points corresponding to the transmitted signals.

7. A MIMO signal detection system based on distributed collaboration, characterized in that: It includes several distributed units and a central processing unit; there is a two-way interactive link between each distributed unit and the central processing unit; each distributed unit includes a receiving antenna array, a radio frequency processing unit, a channel estimation unit and a signal detection unit; Each distributed unit executes the preset detection steps until the number of iterations reaches the preset iteration threshold ; The central processor receives the The estimated vector of the iteration , and the estimated vector Perform quantification processing to obtain the final detection signal ; Wherein, the preset detection steps include: The receiving antenna array is used to receive radio frequency signals; The radio frequency processing unit processes the received radio frequency signal to obtain a baseband signal; The channel estimation unit performs channel estimation using the pilot signal to obtain a local channel matrix; The signal detection unit updates the estimation vector based on the multi-step conditional random block Kaczmarz algorithm using the baseband signal and the local channel matrix, selects the next distributed unit based on the multi-step conditional sampling probability in the multi-step conditional random block Kaczmarz algorithm; and sends the updated estimation vector to the selected next distributed unit based on the topology structure of the MIMO system.

8. The distributed collaborative intelligent MIMO signal transmission system according to claim 7, characterized in that: The formula for the multi-step conditional sampling probability is: ； Where, represents the multi-step conditional sampling probability, represents the sampling probability of the random block Kaczmarz algorithm, is the memory length of multi-step conditional sampling, which is used to control the non-repeatability of distributed units in consecutive iterations; For the Iterations.

9. The distributed collaborative intelligent MIMO signal transmission system according to claim 7, characterized in that: The updating of the estimation vector using the baseband signal and the local channel matrix includes: Calculate the first Dynamic step size at iteration ; Based on baseband signal, local channel matrix, updated dynamic step size , and The estimated vector obtained after iterations , calculate the updated estimate vector.

10. The distributed collaborative intelligent MIMO signal transmission system according to claim 9, characterized in that: In dynamic step When updating, the formula used is: ； Where, is the total number of transmitting antennas in the MINO system, is the total number of receiving antennas in the MINO system, is the number of iterations, is the total number of receiving antennas in a single distribution unit; The calculation formula of the updated estimated vector is: ； Where, is the updated estimate vector, is the local channel matrix, represents the local channel matrix The pseudo-rebellion, is the baseband signal, is the index of the distributed unit, , ; The final detection signal The calculation formula is: ； Where, Round to the nearest The constellation point, for The set of constellation points corresponding to the transmitted signals.

Images