A cyclic compressed sensing multi-path channel estimation method based on TDS-OFDM

Abstract

The embodiment of the application discloses a kind of loop compressed sensing (LCS) multipath channel estimation methods based on TDS-OFDM, it is related to satellite channel estimation technical field, it can improve the channel estimation precision for TDS-OFDM system for satellite communication, reliable recovery of signal is completed simultaneously.The application comprises: the data after being modulated and framing after being influenced by multipath fading channel is extracted after being improved after the preamble sequence of the receiving end, initial LS channel estimation is carried out using preamble sequence, data part is reconstructed and equalized;All PN sequences in the guard interval of separation receiving end are used as measurement vector, use local PN sequence and receiving end equalization book data to construct sensing matrix, and carry out compressed sensing channel estimation;Channel estimation value after compressed sensing channel estimation is used to carry out data equalization again, and sensing matrix and channel estimation are constructed again, and multiple loop compressed sensing channel estimation is carried out.

Description

Technical Field

[0001] This invention relates to the field of satellite channel estimation technology, and in particular to a cyclic compressed sensing multipath channel estimation method based on TDS-OFDM. Background Technology

[0002] Small satellite constellations offer advantages such as low construction cost, minimal susceptibility to ground-based interference, and wide coverage, making them an important direction for future aerospace technology development. TDS-OFDM is a special type of OFDM system that utilizes all subcarriers for information transmission, significantly improving spectral efficiency and making it suitable for high-speed transmission scenarios. However, under multipath channel conditions, PN sequences will alias with the data, causing a deterioration in the system's bit error rate and limiting the widespread application of TDS-OFDM. Therefore, scholars both domestically and internationally have been dedicated to researching channel estimation algorithms to reduce the impact of multipath channels on the system.

[0003] Currently, channel estimation techniques for TDS-OFDM (Time Domain Synchronous Orthogonal Frequency Division Multiplexing) systems have proposed training sequence-based methods, significantly reducing system complexity. However, due to the multipath fading and sparsity of wireless channels, channel estimation relying solely on training sequences is insufficient to capture all channel information. Therefore, this method still suffers from low accuracy, affecting signal recovery. Improved approaches exist, such as compressed sensing-based methods to address the sparsity of wireless channels. These methods analyze the impact of channel delay and amplitude fading on TDS-OFDM symbols, constructing a sensing matrix using training sequences and data. This improves the recovery accuracy of sparse wireless channels and reduces system resource consumption. However, when the difference between channel delay and guard interval length is small, constructing a sensing matrix solely from the non-aliasing regions of sequences and data results in insufficient measurement samples to encompass all channel information, leading to persistently low signal recovery accuracy. In addition, the entire training sequence from the receiver is used to construct the perception matrix. The overlapping parts of the sequence and data will introduce errors, so data balancing is particularly important.

[0004] Therefore, how to improve the channel estimation accuracy of compressed sensing technology when used for channel estimation while ensuring the communication transmission rate of the TDS-OFDM system, and at the same time achieve reliable signal recovery, has become a key area that needs further research. Summary of the Invention

[0005] This invention provides a TDS-OFDM-based cyclic compressed sensing satellite multipath channel estimation method. Addressing the problems in the background art where wireless multipath channels cause training sequence and data aliasing in TDS-OFDM systems, resulting in low accuracy when relying solely on training sequences for channel recovery, and insufficient data accuracy requirements when using the entire region for compressed sensing channel estimation, the proposed method improves the channel estimation accuracy of TDS-OFDM systems. It also solves the system's anti-interference capability in infinite multipath channel environments while achieving accurate channel recovery.

[0006] To achieve the above objectives, the embodiments of the present invention adopt the following technical solutions:

[0007] S01. The receiving end extracts the improved preamble sequence from the data after modulation and framing and the influence of multipath fading channel. The preamble sequence is used to perform initial LS channel estimation, and the data part is reconstructed and equalized.

[0008] S02. Separate all PN sequences in the guard interval of the receiver and use them as measurement vectors. Construct a sensing matrix using the local PN sequences and the equalization data of the receiver and perform compressed sensing channel estimation.

[0009] S03. Use the channel estimation value after compressed sensing channel estimation to perform data equalization again, and reconstruct the sensing matrix and channel estimation again, and perform compressed sensing channel estimation multiple times.

[0010] S11. The data transmitted at the transmitting end is sequentially modulated with 64QAM, then modulated with OFDM and a local sequence PN sequence is added to form a TDS-OFDM symbol, and an improved preamble sequence is added to frame the data.

[0011] S12. The data frame is passed through a multipath sparse channel, which includes multipath delay and amplitude fading.

[0012] The TDS-OFDM-based cyclic compressed sensing satellite multipath channel estimation method provided in this invention aims to improve the system's anti-interference capability and communication performance by mitigating inter-symbol interference caused by the wireless channel. It combines preamble sequence LS channel estimation with compressed sensing channel estimation, and improves the cyclic estimation equalization between the frame header preamble sequence and the locally received PN sequence, achieving more accurate wireless channel estimation performance and enhancing the system's anti-interference capability and communication performance. Since the initial LS estimation equalization of the frame header preamble sequence is fed back to the subsequent receiver PN sequence measurement matrix for cyclic compressed sensing channel estimation, the data aliasing problem caused by time delay is mitigated, further improving the system's channel estimation and communication performance. Attached Figure Description

[0013] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0014] Figure 1 A block diagram of a TDS-OFDM system based on cyclic compressed sensing provided for embodiments of the present invention;

[0015] Figure 2 This is a diagram of the overall data frame structure in a specific example provided in the embodiments of the present invention;

[0016] Figure 3 A flowchart of cyclic compressed sensing in a specific example provided in the embodiments of the present invention;

[0017] Figure 4 A comparison chart of the performance of cyclic compressed sensing channel estimation and LS estimation in a specific example provided in the embodiments of the present invention;

[0018] Figure 5 A comparison chart of cyclic compressed sensing performance based on different reconstruction algorithms in specific examples provided in this embodiment of the invention;

[0019] Figure 6 A comparison chart of the bit error rate performance of LS estimation and cyclic compressed sensing channel estimation systems in specific examples provided in this embodiment of the invention;

[0020] Figure 7 A comparison chart of the bit error rate performance of cyclic compression sensing based on different reconstruction algorithms in specific examples provided in this embodiment of the invention;

[0021] Figure 8 This is a schematic diagram of the method flow provided in an embodiment of the present invention. Detailed Implementation

[0022] To enable those skilled in the art to better understand the technical solutions of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. Embodiments of the present invention will be described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention. Those skilled in the art will understand that, unless specifically stated otherwise, the singular forms “a,” “an,” “the,” and “the” used herein may also include the plural forms. It should be further understood that the term “comprising” as used in the specification of the present invention means the presence of the stated features, integers, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof. It should be understood that when we say an element is “connected” or “coupled” to another element, it can be directly connected or coupled to the other element, or there may be intermediate elements. Furthermore, “connected” or “coupled” as used herein can include wireless connections or couplings. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items. It will be understood by those skilled in the art that, unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. It should also be understood that terms such as those defined in general dictionaries should be understood to have the meaning consistent with their meaning in the context of the prior art, and should not be interpreted in an idealized or overly formal sense unless defined as herein.

[0023] The design concept of this embodiment is to establish a TDS-OFDM satellite communication system, in which the following are designed: Figure 1 The cyclic compressed sensing channel estimation module shown encodes the input bit data channel at the transmitter, performs constellation mapping and serial-to-parallel conversion on the generated binary data bit stream, loads the modulated constellation symbols onto N subcarriers, converts the frequency domain signal into OFDM symbols in the time domain through IFFT transformation, inserts the same PN sequence of length M before each OFDM symbol, and finally transmits the data through parallel-to-serial conversion.

[0024] After the data passes through a multipath channel superimposed with Gaussian white noise, the receiver employs compressed sensing for channel estimation due to the sparsity of the channel. To improve the accuracy of compressed sensing, the measurement vector is selected from all received PN sequences. By analyzing the impact of time delay on the data, a compressed sensing matrix is ​​constructed, and the data at the receiver is reconstructed to obtain more accurate channel information. The serial data is then equalized. Subsequently, serial-to-parallel conversion, FFT transformation, and constellation demapping are performed to obtain the data before channel decoding.

[0025] This invention provides a cyclic compressed sensing multipath channel estimation method based on TDS-OFDM, such as... Figure 8 As shown, it includes:

[0026] S01. The receiver extracts the improved preamble sequence from the data after modulation and framing, which has been affected by multipath fading channels. The preamble sequence is used to perform initial LS channel estimation, and the data part is reconstructed and equalized.

[0027] In practical applications, unlike the static and predictable nature of wired channels, the changes in the wireless channel environment are dynamic and difficult to predict. Wireless channel factors can directly affect the performance of a communication system, requiring precise analysis. This embodiment adds an improved preamble sequence before the data frame to obtain prior information about the channel for initial channel estimation. After obtaining the channel estimate, signal compensation is needed, i.e., at the receiver, the signal is compensated using a characteristic function opposite to that of the channel to eliminate the time or frequency selectivity of the channel, i.e., channel equalization. Due to the delay spread of the wireless channel, the biggest drawback of this system is that the PN sequence will alias with the data, significantly degrading the system performance. Therefore, at the receiver, it is necessary to analyze the interference caused by the wireless multipath channel, and then perform inter-block interference cancellation and reconstruction on the data to separate the interference between the PN sequence and the data, thus ensuring the system's communication performance.

[0028] S02. All PN sequences in the receiver guard interval are separated and used as measurement vectors. A sensing matrix is ​​constructed using the local PN sequences and receiver equalization data to perform compressed sensing channel estimation.

[0029] In compressed sensing channel estimation, the sample matrix needs to be set to be larger than the number of parameters to be estimated. When the channel discrete sampling period is the same as the system discrete sampling period, the guard interval length can be set to be greater than the channel impulse response vector length.

[0030] S03. Use the channel estimation value after compressed sensing channel estimation to perform data equalization again, and reconstruct the sensing matrix and channel estimation again, and perform compressed sensing channel estimation multiple times.

[0031] Each iteration yields a more accurate channel estimate, which in turn improves data accuracy and creates positive feedback.

[0032] S11. The data from the transmitting end is sequentially modulated with 64QAM, then OFDM modulated and a local sequence PN sequence is added to form a TDS-OFDM symbol, and an improved preamble sequence is added to frame the data.

[0033] S12. The data frame is passed through a multipath sparse channel, which includes multipath delay and amplitude fading.

[0034] Specifically, in the channel environment, this includes: channel modeling as a multi-delay tap filter, with the mathematical expression as follows:

[0035]

[0036] Where h(n) represents the total channel impulse response of all paths at time n, a i f is the amplitude gain of the impulse response of the subpath. d To fix the Doppler frequency shift, τ i For time delay. When the multipath component is Rayleigh fading, a i The probability density function of the Rayleigh distribution is:

[0037]

[0038] Where σ 2 It is the variance of the channel envelope. When the multipath component is Rician fading, a i Following a Rice distribution, the probability density function is as follows:

[0039]

[0040] Where I0() represents the zeroth-order Bessel function of the first kind, and A is the peak value of the envelope amplitude of the main signal.

[0041] In this embodiment, the compressed sensing module includes: sparse representation of the signal, measurement matrix design, and signal reconstruction algorithm. Since the satellite wireless channel itself has sparse characteristics, it is necessary to analyze the design of the measurement matrix and the signal reconstruction algorithm. The design of the sensing matrix includes:

[0042] The signal, after passing through a multipath channel and being superimposed with Gaussian noise, is expressed at the receiver as follows:

[0043]

[0044] Where n is a subset of the vectors with a mean of 0 and a variance of . Gaussian white noise, due to the inter-symbol interference caused by fading channel delay and the data tailing phenomenon caused by multipath channels, requires that the sample matrix be larger than the number of parameters to be estimated in compressed sensing channel estimation. When the channel discrete sampling period is consistent with the system discrete sampling period, the guard interval length can be set to be greater than the channel impulse response vector length, and M≤L, i.e., MT, can be set. s ≤LT s In traditional TDS-OFDM systems with rapid time variations, only the channel coefficient changes in the inter-block interference (IBI) region within the guard interval can be obtained. This means that only the beginning of the channel tap changes can be estimated from the PN sequence. Since the receiver does not know the data samples before channel estimation, it cannot obtain the rest of the channel tap distribution. Due to missing the channel impulse response at other times, the channel estimation results are inaccurate, resulting in poor symbol detection performance. To overcome this problem in traditional TDS-OFDM systems, compressed sensing channel estimation is performed at the receiver using the entire PN sequence. Under slow fading channels, the system performance is improved by recovering from inter-block interference.

[0045] At this point, the symbol vector expression at the receiving end is:

[0046]

[0047] Where ψ(i) and y(i) are the trailing parts of PN and time domain after passing through the multipath channel, respectively. Due to the influence of multipath delay, the L sampling points at the end of the currently received (i-1)th domain data are mixed with the first L time domain sampling samples of the currently received i-th PN sequence, which causes inter-block interference. However, the part of the PN sequence with a length greater than the channel impulse response length is not interfered with, that is, there is no inter-block interference.

[0048] The first M sampling points of the received symbol constitute the received PN sequence, expressed as follows:

[0049]

[0050] in, y(i)=x(i)*h, using PN sequences for compressed sensing channel estimation, then d i Let A be the measurement vector and A be the perception matrix. The expression for A is:

[0051]

[0052] In A, x represents the equalized time-domain data, which in this system is a 64QAM symbol. The first equalized data is estimated by LS using the frame header preamble sequence to obtain the first channel impulse response value. After the compressed sensing estimates the channel, the channel frequency response matrix is ​​constructed to equalize the data. Then, the data is iterated and the sensing matrix is ​​used for secondary equalization. This process is repeated three times until convergence.

[0053] In this embodiment, the first channel impulse response value is estimated using LS with the frame header preamble sequence, including: the frame header is also divided into the frame body and the preceding and following cyclic prefixes, and the frame body vector of the frame header is [p1 p2 p1 * p2 * ], where p1 is a 64QAM symbol modulated by a random sequence of length N / 2, p2 is a symmetric sequence of p1, p1 * Let p1 be the conjugate sequence, and p2 be the conjugate sequence. * This represents the conjugate sequence of p2. The large frame body consists of multiple OFDM symbols. At this point, the frame header sequence has prefixes and suffixes. After removing the prefixes and suffixes, the influence of multipath effects can be minimized, resulting in a more accurate initial channel estimate. The principle expression for the first LS estimation is:

[0054]

[0055] The frequency domain expression after data reconstruction at the receiving end is:

[0056] Y = FHx + N

[0057] =FHF H X+N

[0058] =GX+N

[0059] Where Y is the frequency domain representation of the received data, H is the cyclic Topulitz matrix, F is the normalized DFT matrix, and G = FHF H It is the channel frequency response (CFR) matrix, and the cyclic Topplitz matrix expression is:

[0060]

[0061] Here, the element h(n,l) represents the amplitude gain of the nth subcarrier on the lth path. When the channel is a slow fading channel, the gain on each subcarrier remains unchanged. At this time, h(n,l)=h(0,l)=h(1,l)=…=h(N-1,l).

[0062] The expression for the balanced data at this point is:

[0063]

[0064] The data from the last equalization is demodulated and decoded to obtain the output bits at the receiver, which can be compared with the bits at the transmitter to obtain the bit error rate.

[0065] In this embodiment, the reconstruction algorithm includes:

[0066] For the reconstruction after sampling, the reconstruction algorithm adopts the Compressed Sampling Matching Pursuit Algorithm (CoSaMP), which is a type of greedy algorithm such as OMP and ROMP. It represents sparse signals by selecting the most relevant atoms in the sensing matrix.

[0067] In summary, the cyclic compressed sensing channel estimation algorithm is as follows:

[0068]

[0069]

[0070] This embodiment also includes: a simulation comparison between the proposed cyclic compressed sensing method and the traditional LS estimation TDS-OFDM. Additionally, to verify the impact of the reconstruction module on the cyclic compressed sensing method, the performance of the proposed LCS-CoSaMP method was compared with that of LCS-OMP and LCS-ROMP methods. The initial parameters of the satellite TDS-OFDM system based on cyclic compressed sensing mainly include the TDS-OFDM system parameters and channel parameter settings, as shown in Tables 1 and 2, respectively.

[0071] Table 1

[0072]

[0073] Table 2

[0074]

[0075]

[0076] This embodiment proposes a cyclic compressed sensing multipath channel estimation method based on TDS-OFDM, which is a cyclic channel estimation method based on LS channel estimation of the frame header preamble sequence and compressed sensing channel estimation of the subsequent received PN sequence. By proposing a cyclic compressed sensing channel estimation method that equally feeds back the initial LS estimation of the frame header to the measurement matrix of the subsequent received PN sequence, the data aliasing problem caused by time delay is improved. The mean square error curve of its channel estimation under multipath Rayleigh sparse channels is compared with the performance of traditional LS channel estimation, and the bit error rate is compared as follows: Figure 4 and Figure 6 As shown in the figure, the cyclic compressed sensing channel estimation method has better performance in both MSE and BER compared to the traditional LS estimation method of TDS-OFDM system.

[0077] Cyclic compressed sensing selects the mean square error of channel estimation as the optimization objective, establishes an optimization model, and introduces the CoSaMP reconstruction algorithm for iterative calculation to obtain the optimal mean square error performance, thereby achieving multipath channel estimation. Furthermore, to verify the impact of the reconstruction module on the cyclic compressed sensing method, the proposed LCS-CoSaMP method is compared with other reconstruction algorithms such as LCS-OMP and LCS-ROMP in terms of channel estimation performance and system bit error rate. Figure 5 and Figure 7 As shown in the figure, the proposed LCS-CoSaMP algorithm has the best MSE performance under different signal-to-noise ratio conditions.

[0078] In summary, the advantages of this embodiment compared to existing technologies are: it can simultaneously achieve two effects: 1) By analyzing the key characteristics of the multipath channel's impact on the signal, an optimized model is established, and a cyclic compressed sensing channel estimation method is introduced, which feeds back the initial LS estimation equalization of the frame header preamble sequence to the subsequent receiver's PN sequence measurement matrix, thereby achieving accurate estimation of the multipath channel. 2) It improves the signal's anti-interference capability and enhances the system's communication performance.

[0079] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on its differences from other embodiments. In particular, the device embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments. The above descriptions are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A multipath channel estimation method based on TDS-OFDM loop compressed sensing (LCS), characterized in that, include: S01. The receiving end extracts the improved preamble sequence from the data after modulation and framing, which has been affected by multipath fading channels. In the improved preamble sequence, the frame header is divided into the frame body and the preceding and following cyclic prefixes. The frame body vector of the frame header is [p1 p2 p1...]. * p2 * ], where p1 is a 64QAM symbol modulated by a random sequence with a length of 512 sampling points, p2 is a symmetric sequence of p1, p1 * Let p1 be the conjugate sequence, and p2 be the conjugate sequence. * The conjugate sequence of p2 is used to perform initial LS channel estimation using the preamble sequence, and the data portion is reconstructed and equalized. S02. Use all the PN sequences of the receiver in the guard interval of the receiver as a measurement vector, and construct a sensing matrix using the local PN sequence of the transmitter and the equalization data of the receiver to perform compressed sensing channel estimation. S03. Use the channel estimation value after compressed sensing channel estimation to perform data equalization again, and reconstruct the sensing matrix and channel estimation again, and perform compressed sensing channel estimation multiple times.

2. The method according to claim 1, characterized in that, Also includes: S11. The data from the transmitting end is sequentially modulated with 64QAM, then modulated with OFDM and the local sequence PN sequence from the transmitting end is added to form a TDS-OFDM symbol, and the improved preamble sequence is added to frame the data. S12. The data frame is passed through a multipath sparse channel, which includes multipath delay and amplitude fading.

3. The method according to claim 2, characterized in that, In the process of performing OFDM modulation and adding the local PN sequence of the transmitter to form a TDS-OFDM symbol, the 64QAM modulated data is converted from serial to parallel, and then converted by 1024-point IFFT to form an OFDM symbol. The guard interval between each subcarrier of OFDM is filled with a 256-point local PN sequence of the transmitter to form a TDS-OFDM symbol. Multiple consecutive TDS-OFDM symbols form a data frame, and an improved preamble sequence is added to the frame header.

4. The method according to claim 1, characterized in that, Initial LS channel estimation using the receiver preamble includes: The transmitted signal passes through a multipath channel, which is a linear convolution relationship, resulting in an N-point long signal. Channel with length L The data after convolution has a length of N + L - 1, and at this point, the data produces a tail of length L - 1, expressed as: y = x * h The signal, after passing through a multipath channel and being superimposed with Gaussian noise, is expressed at the receiver as follows: Where n is a subset of the vectors with a mean of 0 and a variance of . Gaussian white noise M is the local PN sequence at the transmitter, and M is the length of the local PN sequence at the transmitter.

5. The method according to claim 1, characterized in that, Also includes: Based on the influence of the channel on symbols, the expression for the i-th symbol vector at the receiver is: Where ψ(i) and y(i) are the trailing parts of PN and time domain after passing through the multipath channel, respectively. Due to the influence of multipath delay, the L sampling points at the end of the currently received (i-1)th time domain data are mixed with the first L time domain sampling samples of the currently received PN sequence. The first M sampling points of the received symbol constitute the received PN sequence, expressed as: in, y(i) = x(i)*h, using the received PN sequence for compressed sensing channel estimation, then d i Let A be the measurement vector and A be the perception matrix. The expression for A is: In A, x represents the equalized time-domain data, which is the 64QAM symbol in this system.

6. The method according to claim 1, characterized in that, Also includes: Local PN sequence at the transmitter After passing through the channel, the received sequence is d, and the LS estimation principle expression is: After equalizing the data with the initial channel estimate, compressed sensing channel estimation can be performed to obtain a new channel estimate, and the data can be equalized again.

7. The method according to claim 6, characterized in that, Also includes: After performing multiple equalizations on the data to obtain the final channel estimate, it can be compared with the true value of the channel model used to obtain the mean square error.

Images