Transmitting and receiving device based on negentropy statistics and compressive sensing

Abstract

The invention discloses a receiving and transmitting device based on negentropy statistics and compressive sensing, and relates to the field of high-speed wireless communication. The device comprises a channel encoder, a modulator, a sparse encoder, a multi-channel transmitter, a multi-channel receiver, a digital preprocessor, an interference eliminator based on negentropy statistics, a signal reconstruction device based on compressive sensing, a symbol synchronizer, a multi-channel demodulator, a channel decoder and the like. According to an interference elimination technology based on negentropy statistics and a nonlinear signal reconstruction technology based on compressive sensing, under the condition that the complexity of the system is not increased, the multi-path high-speed parallel transmission capacity of the system is improved, high-speed reliable transmission is guaranteed, the anti-interference capacity of the system is improved, and the reliability of the system is improved. The antenna is especially suitable for high-speed wireless communication scenes such as microwave communication, mobile communication and scatter communication.

Description

Technical Field

[0001] The invention discloses a transceiver based on negative entropy statistics and compressed sensing, and relates to the field of high-speed wireless communications. Background Art

[0002] In modern high-speed wireless communications, in order to meet the requirements of high-throughput wireless transmission, the use of multi-path parallel transmission is an effective means. However, the parallel transmission communication system is highly complex, and there is interference between multiple paths. As the throughput increases, the interference between multiple paths leads to a decrease in system performance. In addition, how to reduce the impact of external interference during high-speed wireless transmission is also a problem that needs to be solved. Summary of the invention

[0003] The purpose of the present invention is to avoid the shortcomings of the above-mentioned background technology and provide a transceiver based on negative entropy statistics and compressed sensing. The present invention uses interference elimination technology based on negative entropy statistics to extract multi-dimensional independent signals and eliminate interference from received signals, and purifies and reconstructs multi-dimensional signals through nonlinear signal reconstruction technology based on compressed sensing, thereby improving the system's multi-channel high-speed transmission capability, ensuring high-speed and reliable transmission, and improving the system's anti-interference ability.

[0004] The object of the present invention is achieved in that:

[0005] A transceiver based on negative entropy statistics and compressed sensing, comprising a channel encoder 1, a modulator 2, a sparse encoder 3, a multi-channel transmitter 4, a multi-channel receiver 5, a digital preprocessor 6, a symbol synchronizer 9, a multi-channel demodulator 10, a channel decoder 11 and a power supply 12, and further comprising an interference eliminator 7 based on negative entropy statistics and a signal reconstructor 8 based on compressed sensing;

[0006] Among them, the channel encoder 1 performs channel encoding on the received service information, and then sends the encoded signal to the modulator 2;

[0007] The modulator 2 digitally modulates the encoded signal and sends the modulated signal to the sparse encoder 3;

[0008] The sparse encoder 3 performs sparse encoding on the modulated signal, and then sends the digital sparse modulated signal to the multi-channel transmitter 4;

[0009] The multi-channel transmitter 4 performs digital / analog conversion, up-conversion, power amplification and filtering on the digital sparse modulation signal, and then outputs the radio frequency transmission signal to the multi-channel receiver 5;

[0010] The multi-channel receiver 5 performs low-noise amplification, filtering, down-conversion and analog / digital conversion on the received RF signal, and outputs the digital sampling signal to the digital preprocessor 6;

[0011] The digital preprocessor 6 performs digital down-conversion and filtering on the digital sampling signal, and sends the processed digital baseband signal to the interference canceller 7 based on negative entropy statistics;

[0012] The interference canceller 7 based on negative entropy statistics performs de-averaging, whitening, multi-dimensional negative entropy statistics, parallel iterative extraction and interference elimination on the received digital baseband signal to obtain a de-interference signal, and sends it to the signal reconstructor 8 based on compressed sensing;

[0013] The signal reconstructor 8 based on compressed sensing performs sparse orthogonal optimization, compressed sampling, tensor analysis, nanofiltering and nonlinear reconstruction processing on the interference removal signal, and outputs the processed purified reconstructed signal to the symbol synchronizer 9 and the multi-channel demodulator 10;

[0014] The symbol synchronizer 9 extracts the synchronization information and sends the synchronization information to the multi-channel demodulator 10;

[0015] The multi-channel demodulator 10 synchronously demodulates the purified reconstructed signal according to the synchronization information, and sends the processed demodulated signal to the channel decoder 11;

[0016] The channel decoder 11 performs efficient channel decoding according to the demodulated signal to obtain decoded service information.

[0017] Further, the interference eliminator 7 based on negative entropy statistics includes a de-averager 13, a signal whitener 14, a multi-dimensional negative entropy statistician 15, a multi-dimensional threshold generator 16, a multi-dimensional signal parallel iterative extractor 17, an interference signal eliminator 18 and a signal anti-whitener 19;

[0018] The digital baseband signal output by the digital preprocessor 6 is preprocessed by the de-averaging device 13, and the de-averaged signal is sent to the signal whitening device 14;

[0019] The signal whitener 14 performs whitening processing on the signal after the mean value is removed, and sends the obtained whitened signal to the multidimensional negative entropy statistician 15, the multidimensional threshold generator 16 and the multidimensional signal parallel iterative extractor 17;

[0020] The multi-dimensional negative entropy statistics device 15 is used to perform multi-dimensional negative entropy statistics on the whitening signal, and send the statistical results to the multi-dimensional signal parallel iterative extractor 17;

[0021] The multidimensional threshold generator 16 obtains the characteristic value of the whitening signal, adaptively generates a multidimensional threshold according to the characteristic value, and sends the multidimensional threshold to the multidimensional signal parallel iterative extractor 17;

[0022] The multidimensional signal parallel iterative extractor 17 performs parallel iterative extraction processing on the whitened signal input by the signal whitener 14 according to the negative entropy statistical result input by the multidimensional negative entropy statistician 15 and the multidimensional threshold value input by the multidimensional threshold value generator 16, and sends the processed independent multidimensional signal to the interference signal eliminator 18;

[0023] The interference signal eliminator 18 separates the interference signal from the independent multi-dimensional signal, and sends the purified signal after the interference is eliminated to the signal de-whitening device 19;

[0024] The signal de-whitening device 19 performs de-whitening processing on the purified signal, and then sends the processed interference-free signal to the signal reconstructor 8 based on compressed sensing.

[0025] Further, the signal reconstructor 8 based on compressed sensing includes a low-pass filter bank 20, a sparse orthogonal optimizer 21, a compressed sampler 22, an orthogonal matching tracker 23, a signal tensor analyzer 24, a Wiener filter 25 and a nonlinear reconstructor 26;

[0026] The low-pass filter group 20 is used to perform multiple groups of filtering on the interference removal signal output by the interference eliminator 7 based on the negative entropy statistics, and output the filtered signal to the sparse orthogonal optimizer 21;

[0027] The sparse orthogonal optimizer 21 performs sparse orthogonal processing on the filtered signal sent by the low-pass filter bank 20, and sends the sparse signal to the compression sampler 22 and the signal tensor analyzer 24;

[0028] The compression sampler 22 is used to compress and sample the sparse signal from the sparse orthogonal optimizer 21, highlight the dominant information in the sparse signal, and send the compressed sampled signal to the orthogonal matching tracker 23;

[0029] The orthogonal matching tracker 23 performs orthogonal matching and dominant information tracking on the compressed sampling signal from the compressed sampler 22, and sends the purified dominant information to the nonlinear reconstructor 26;

[0030] The signal tensor analyzer 24 performs tensor analysis on the sparse signal from the sparse orthogonal optimizer 21 to obtain a signal feature vector and sends it to the Wiener filter 25;

[0031] The Wiener filter 25 performs optimal Wiener filtering on the signal feature vector from the signal tensor analyzer 24 and sends the optimized feature vector to the nonlinear reconstructor 26;

[0032] The nonlinear reconstructor 26 is used to perform nonlinear reconstruction processing on the purified advantage information from the orthogonal matching tracker 23 and the optimized feature vector from the Wiener filter 25 to obtain a purified reconstructed signal, and send it to the symbol synchronizer 9 and the multi-channel demodulator 10.

[0033] Compared with the background technology, the present invention has the following advantages:

[0034] 1. The present invention adopts an interference eliminator 7 based on negative entropy statistics, and adopts interference elimination technology based on negative entropy statistics. Through negative entropy statistics and parallel iterative signal extraction, multi-dimensional independent signal extraction and interference elimination are performed on the received signal, interference between transmission channels is eliminated, the reliability of the system is improved, and the anti-interference ability of the system is enhanced.

[0035] 2. The present invention adopts a signal reconstructor 8 based on compressed sensing, and adopts nonlinear signal reconstruction technology based on compressed sensing to purify and reconstruct multidimensional signals, thereby enhancing the high-speed multi-channel transmission capability and greatly improving the throughput of the system, which can be improved by an order of magnitude.

[0036] 3. The circuit components of the present invention are made of large-scale field programmable devices, so the working parameters can be flexibly modified by configuring different programs, which greatly simplifies the structure of the equipment and significantly reduces the cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 It is an electrical principle block diagram of a transceiver based on negative entropy statistics and compressed sensing in an embodiment of the present invention.

[0038] Figure 2 yes Figure 1 The electrical schematic diagram of the interference eliminator 7 based on negative entropy statistics.

[0039] Figure 3 yes Figure 1 The electrical schematic diagram of the signal reconstructor 8 based on compressed sensing. DETAILED DESCRIPTION

[0040] The present invention is further described in detail below with reference to the accompanying drawings.

[0041] A transceiver based on negative entropy statistics and compressed sensing, referring to Figure 1 , including a channel encoder 1, a modulator 2, a sparse encoder 3, a multi-channel transmitter 4, a multi-channel receiver 5, a digital preprocessor 6, an interference eliminator based on negative entropy statistics 7, a signal reconstructor based on compressed sensing 8, a symbol synchronizer 9, a multi-channel demodulator 10, a channel decoder 11 and a power supply 12. Figure 1 This is a block diagram of the electrical principle of a transceiver device based on negative entropy statistics and compressed sensing of the present invention. Figure 1 Connect the lines.

[0042] The service information inputted from the service input port A is channel encoded by the channel encoder 1, and then the encoded signal is sent to the modulator 2; the modulator 2 digitally modulates the encoded signal, and sends the modulated signal to the sparse encoder 3; the sparse encoder 3 sparsely encodes the modulated signal, and then sends the digital sparse modulated signal to the multi-channel transmitter 4; the multi-channel transmitter 4 performs digital / analog conversion, up-conversion, power amplification and filtering on the digital sparse modulated signal, and then outputs the RF transmission signal to the RF output port B; the RF signal received by the RF input port C is low-noise amplified, filtered, down-converted and analog / digital converted by the multi-channel receiver 5, and then outputs the digital sampling signal to the digital preprocessor 6; the digital preprocessor 6 performs digital down-conversion and filtering on the digital sampling signal, and sends the processed digital baseband signal to the interference eliminator 7 based on negative entropy statistics ; The interference eliminator 7 based on negative entropy statistics performs de-averaging, whitening, multi-dimensional negative entropy statistics, parallel iterative extraction and interference elimination on the received digital baseband signal to obtain a de-interference signal, and sends it to the signal reconstructor 8 based on compressed sensing; the signal reconstructor 8 based on compressed sensing performs sparse orthogonal optimization, compressed sampling, tensor analysis, nano-filtering and nonlinear reconstruction on the de-interference signal, and outputs the processed purified reconstructed signal to the symbol synchronizer 9 and the multi-channel demodulator 10; the symbol synchronizer 9 extracts the synchronization information and sends the synchronization information to the multi-channel demodulator 10; the multi-channel demodulator 10 performs synchronous demodulation on the purified reconstructed signal according to the synchronization information, and sends the processed demodulated signal to the channel decoder 11; the channel decoder 11 performs efficient channel decoding according to the demodulated signal, obtains the decoded service information and outputs it to the service output port D.

[0043] In the embodiment, the channel encoder 1, modulator 2, sparse encoder 3, digital preprocessor 6, interference eliminator based on negative entropy statistics 7, signal reconstructor based on compressed sensing 8, symbol synchronizer 9, multi-channel demodulator 10 and channel decoder 11 are all made of V7 series FPGA chips produced by Fudan Micro. The multi-channel transmitter 4 is made of multi-channel transmission component TS-FHTRC of Shijiazhuang Dongtail Company. The multi-channel receiver 5 is made of high-performance receiving component ZKJ-TRS08103 of Zhongkexin Company.

[0044] The function of the interference eliminator 7 based on negative entropy statistics is to perform de-averaging, whitening, multi-dimensional negative entropy statistics, parallel iterative extraction and interference elimination on the received digital baseband signal to obtain the interference-free signal and send it to the signal reconstructor 8 based on compressed sensing. It consists of a de-averaging device 13, a signal whitening device 14, a multi-dimensional negative entropy statistician 15, a multi-dimensional threshold generator 16, a multi-dimensional signal parallel iterative extractor 17, an interference signal eliminator 18 and a signal anti-whitening device 19. Figure 2 The electrical schematic diagram of the interference canceller 7 based on negative entropy statistics is shown in the embodiment. Figure 2 Connect the lines.

[0045] Among them, the digital baseband signal output by the digital preprocessor 6 is preprocessed by de-averaging device 13, and the de-averaged signal is sent to the signal whitening device 14; the signal whitening device 14 whitens the de-averaged signal, and sends the obtained whitened signal to the multidimensional negative entropy statistics device 15, the multidimensional threshold generator 16 and the multidimensional signal parallel iterative extractor 17; the multidimensional negative entropy statistics device 15 is used to perform multidimensional negative entropy statistics on the whitened signal, and send the statistical results to the multidimensional signal parallel iterative extractor 17; the multidimensional threshold generator 16 obtains the characteristic value of the whitened signal, adaptively generates a multidimensional threshold according to the characteristic value, and sends the multidimensional threshold to the multidimensional signal parallel iterative extractor 17. The multidimensional signal is input into the multidimensional signal parallel iterative extractor 17; the multidimensional signal parallel iterative extractor 17 performs parallel iterative extraction processing on the whitened signal input by the signal whitener 14 according to the negative entropy statistical result input by the multidimensional negative entropy statistician 15 and the multidimensional threshold value input by the multidimensional threshold generator 16, and sends the processed independent multidimensional signal to the interference signal eliminator 18; the interference signal eliminator 18 separates the interference signal in the independent multidimensional signal, and sends the purified signal after the interference is eliminated to the signal de-whitening device 19; the signal de-whitening device 19 performs de-whitening processing on the purified signal, and then sends the processed interference-free signal to the signal reconstructor 8 based on compressed sensing.

[0046] In the embodiment, the de-averaging device 13, the signal whitening device 14, the multi-dimensional negative entropy statistic device 15, the multi-dimensional threshold generator 16, the multi-dimensional signal parallel iterative extractor 17, the interference signal eliminator 18 and the signal de-whitening device 19 are all made of the V7 series FPGA chip produced by Fudan Microelectronics Corporation.

[0047] The function of the signal reconstructor 8 based on compressed sensing is to perform sparse orthogonal optimization, compressed sampling, tensor analysis, nanofiltering and nonlinear reconstruction on the interference-removing signal, and output the processed purified reconstructed signal to the symbol synchronizer 9 and the multi-channel demodulator 10. It consists of a low-pass filter group 20, a sparse orthogonal optimizer 21, a compressed sampler 22, an orthogonal matching tracker 23, a signal tensor analyzer 24, a Wiener filter 25 and a nonlinear reconstructor 26. Figure 3 The schematic diagram of the signal reconstructor 8 based on compression sensing is shown in the embodiment. Figure 3 Connect the lines.

[0048] Among them, the low-pass filter group 20 is used to perform multiple groups of filtering processing on the interference removal signal output by the interference eliminator 7 based on negative entropy statistics, and output the filtered signal to the sparse orthogonal optimizer 21; the sparse orthogonal optimizer 21 performs sparse orthogonal processing on the filtered signal sent by the low-pass filter group 20, and sends the sparse signal to the compression sampler 22 and the signal tensor analyzer 24; the compression sampler 22 is used to perform compression sampling on the sparse signal from the sparse orthogonal optimizer 21, highlight the dominant information in the sparse signal, and send the compressed sampled signal to the orthogonal matching tracker 23; the orthogonal matching tracker 23 performs orthogonal processing on the compressed sampled signal from the compression sampler 22. Cross-matching and advantage information tracking, and sending the purified advantage information to the nonlinear reconstructor 26; the signal tensor analyzer 24 performs tensor analysis on the sparse signal from the sparse orthogonal optimizer 21 to obtain the signal feature vector, and sends it to the Wiener filter 25 for filtering; the Wiener filter 25 performs optimal Wiener filtering on the signal feature vector from the signal tensor analyzer 24, and sends the optimized feature vector to the nonlinear reconstructor 26; the nonlinear reconstructor 26 is used to perform nonlinear reconstruction processing on the purified advantage information from the orthogonal matching tracker 23 and the optimized feature vector from the Wiener filter 25 to obtain the purified reconstructed signal, and send it to the symbol synchronizer 9 and the multi-channel demodulator 10.

[0049] In the embodiment, the low-pass filter group 20, the sparse orthogonal optimizer 21, the compressed sampler 22, the orthogonal matching tracker 23, the signal tensor analyzer 24, the Wiener filter 25 and the nonlinear reconstructor 26 are all made of V7 series FPGA chips produced by Fudan Microelectronics.

[0050] The power supply 12 provides a DC working voltage for the entire receiving device. The embodiment is made of a commercially available general-purpose integrated voltage-stabilized DC power supply module, and its output +V voltage is +5V.

[0051] The brief working principle of the present invention is as follows:

[0052] The service information inputted from the service input port A is channel encoded by the channel encoder 1, and then the encoded signal is sent to the modulator 2; the modulator 2 digitally modulates the encoded signal, and sends the modulated signal to the sparse encoder 3; the sparse encoder 3 sparsely encodes the modulated signal, and then sends the digital sparse modulated signal to the multi-channel transmitter 4; the multi-channel transmitter 4 performs digital / analog conversion, up-conversion, power amplification and filtering on the digital sparse modulated signal, and then outputs the RF transmission signal to the RF output port B; the RF signal received by the RF input port C is low-noise amplified, filtered, down-converted and analog / digital converted by the multi-channel receiver 5, and then outputs the digital sampling signal to the digital preprocessor 6; the digital preprocessor 6 performs digital down-conversion and filtering on the digital sampling signal, and sends the processed digital baseband signal to the interference eliminator 7 based on negative entropy statistics ; The interference eliminator 7 based on negative entropy statistics performs de-averaging, whitening, multi-dimensional negative entropy statistics, parallel iterative extraction and interference elimination on the received digital baseband signal to obtain a de-interference signal, and sends it to the signal reconstructor 8 based on compressed sensing; the signal reconstructor 8 based on compressed sensing performs sparse orthogonal optimization, compressed sampling, tensor analysis, nano-filtering and nonlinear reconstruction on the de-interference signal, and outputs the processed purified reconstructed signal to the symbol synchronizer 9 and the multi-channel demodulator 10; the symbol synchronizer 9 extracts the synchronization information and sends the synchronization information to the multi-channel demodulator 10; the multi-channel demodulator 10 performs synchronous demodulation on the purified reconstructed signal according to the synchronization information, and sends the processed demodulated signal to the channel decoder 11; the channel decoder 11 performs efficient channel decoding according to the demodulated signal, obtains the decoded service information and outputs it to the service output port D.

[0053] The installation structure of the present invention is as follows: Figure 1 The multi-channel transmitter 4 and the multi-channel receiver 5 are mounted on a printed circuit board with a size of 100×80 mm in length×width; Figure 1 The channel encoder 1, modulator 2, sparse encoder 3, digital preprocessor 6, symbol synchronizer 9, multi-channel demodulator 10, and channel decoder 11 are mounted on a printed circuit board with a size of 100×80 mm in length and width. Figure 2 , Figure 3 All circuit components are installed on a printed circuit board with a size of 80×80 mm in length×width, and then the three printed circuit boards are installed in a shielding box plug-in with a size of 200×100×60 mm in length×width×height. The front panel of the shielding box plug-in is installed with a cable socket connected to external ports A, B, C, and D, and the rear panel is installed with a power input socket to assemble the present invention.

[0054] The present invention improves the system's multi-channel high-speed transmission capability without increasing the system complexity based on the interference elimination technology based on negative entropy statistics and the nonlinear signal reconstruction technology based on compressed sensing, ensures high-speed and reliable transmission, and enhances the system's anti-interference capability. The system is particularly suitable for high-speed wireless communication scenarios such as microwave communication, mobile communication and scattering communication.

Claims

1. A transceiver based on negative entropy statistics and compressed sensing, comprising a channel encoder (1), a modulator (2), a sparse encoder (3), a multi-channel transmitter (4), a multi-channel receiver (5), a digital preprocessor (6), a symbol synchronizer (9), a multi-channel demodulator (10), a channel decoder (11) and a power supply (12), characterized in that: It also includes an interference canceller (7) based on negative entropy statistics and a signal reconstructor (8) based on compressed sensing; The channel encoder (1) performs channel coding on the received service information and then sends the coded signal to the modulator (2); The modulator (2) digitally modulates the coded signal and sends the modulated signal to the sparse encoder (3); The sparse encoder (3) performs sparse encoding on the modulated signal and then sends the digital sparse modulated signal to the multi-channel transmitter (4); The multi-channel transmitter (4) performs digital / analog conversion, up-conversion, power amplification and filtering on the digital sparse modulation signal, and then outputs a radio frequency transmission signal to the multi-channel receiver (5); The multi-channel receiver (5) performs low-noise amplification, filtering, down-conversion and analog / digital conversion on the received radio frequency signal, and outputs the digital sampling signal to the digital preprocessor (6); The digital preprocessor (6) performs digital down-conversion and filtering on the digital sampling signal, and sends the processed digital baseband signal to the interference canceller (7) based on negative entropy statistics; The interference canceller (7) based on negative entropy statistics performs de-averaging, whitening, multi-dimensional negative entropy statistics, parallel iterative extraction and interference elimination on the received digital baseband signal to obtain a de-interference signal, and sends it to a signal reconstructor (8) based on compressed sensing; A signal reconstructor (8) based on compressed sensing performs sparse orthogonal optimization, compressed sampling, tensor analysis, nanofiltering and nonlinear reconstruction processing on the interference-removing signal, and outputs the processed purified reconstructed signal to a symbol synchronizer (9) and a multi-channel demodulator (10); The symbol synchronizer (9) extracts synchronization information and sends the synchronization information to the multi-channel demodulator (10); The multi-channel demodulator (10) performs synchronous demodulation on the purified reconstructed signal according to the synchronization information, and sends the processed demodulated signal to the channel decoder (11); The channel decoder (11) performs efficient channel decoding according to the demodulated signal to obtain decoded service information.

2. A transceiver based on negative entropy statistics and compressed sensing according to claim 1, characterized in that: The interference eliminator (7) based on negative entropy statistics includes a de-averaging device (13), a signal whitening device (14), a multi-dimensional negative entropy statistician (15), a multi-dimensional threshold generator (16), a multi-dimensional signal parallel iterative extractor (17), an interference signal eliminator (18) and a signal anti-whitening device (19); The de-averaging device (13) is used to perform de-averaging preprocessing on the input digital baseband signal, and send the de-averaged signal to the signal whitening device (14); The signal whitener (14) performs whitening processing on the signal after the mean value is removed, and sends the obtained whitened signal to the multidimensional negative entropy statistician (15), the multidimensional threshold generator (16) and the multidimensional signal parallel iterative extractor (17); The multidimensional negative entropy statistics device (15) is used to perform multidimensional negative entropy statistics on the whitening signal, and send the statistical results to the multidimensional signal parallel iterative extractor (17); The multidimensional threshold generator (16) obtains the characteristic value of the whitening signal, adaptively generates a multidimensional threshold according to the characteristic value, and sends the multidimensional threshold to the multidimensional signal parallel iterative extractor (17); The multidimensional signal parallel iterative extractor (17) performs parallel iterative extraction processing on the whitened signal input by the signal whitening device (14) according to the negative entropy statistical result input by the multidimensional negative entropy statistic device (15) and the multidimensional threshold value input by the multidimensional threshold value generator (16), and sends the processed independent multidimensional signal to the interference signal eliminator (18); The interference signal eliminator (18) separates the interference signal from the independent multi-dimensional signal, and sends the purified signal after the interference is eliminated to the signal de-whitening device (19); The signal de-whitening device (19) performs de-whitening processing on the purified signal, and then sends the processed interference-free signal to the signal reconstructor (8) based on compressed sensing.

3. The transceiver device based on negative entropy statistics and compressed sensing according to claim 1, characterized in that: A signal reconstructor (8) based on compressed sensing includes a low-pass filter bank (20), a sparse orthogonal optimizer (21), a compressed sampler (22), an orthogonal matching tracker (23), a signal tensor analyzer (24), a Wiener filter (25) and a nonlinear reconstructor (26); The low-pass filter group (20) is used to perform multiple groups of filtering processes on the interference removal signal output by the interference eliminator (7) based on negentropy statistics, and output the filtered signal to the sparse orthogonal optimizer (21); The sparse orthogonal optimizer (21) performs sparse orthogonal processing on the filtered signal sent by the low-pass filter group (20), and sends the sparse signal to the compression sampler (22) and the signal tensor analyzer (24); The compression sampler (22) is used to compress and sample the sparse signal from the sparse orthogonal optimizer (21), highlight the dominant information in the sparse signal, and send the compressed sampled signal to the orthogonal matching tracker (23); The orthogonal matching tracker (23) performs orthogonal matching and dominant information tracking on the compressed sampling signal from the compressed sampler (22), and sends the purified dominant information to the nonlinear reconstructor (26); The signal tensor analyzer (24) performs tensor analysis on the sparse signal from the sparse orthogonal optimizer (21), obtains the signal feature vector, and sends it to the Wiener filter (25); The Wiener filter (25) performs optimal Wiener filtering on the signal feature vector from the signal tensor analyzer (24), and sends the optimized feature vector to the nonlinear reconstructor (26); The nonlinear reconstructor (26) is used to perform nonlinear reconstruction processing on the purified advantage information from the orthogonal matching tracker (23) and the optimized feature vector from the Wiener filter (25), obtain a purified reconstructed signal, and send it to the symbol synchronizer (9) and the multi-channel demodulator (10).

Images