Communication-sensing integrated multi-parameter joint estimation method and device

Abstract

The embodiment of the present invention discloses a multi-parameter joint estimation method and device for integrated communication perception, which relates to the field of wireless communication and radar technology, wherein the method comprises: performing down-conversion, cyclic prefix removal, discrete time sampling and DFT on a time-domain OFDM signal to obtain a discrete three-dimensional signal observation in the frequency domain, then performing smoothing processing on the signal to obtain a plurality of sub-signal observations, and then estimating the angles, distances and speeds of a plurality of targets by a one-dimensional root-finding MUSIC algorithm and matching them to obtain an initial parameter pair, and based on the initial parameter pair, performing subspace iteration and Levenberg-Marquardt convex optimization on the local convex region of each target in the two-dimensional space spectrum to obtain a plurality of two-dimensional parameter pairs, and finally matching the parameter pairs using the minimum distance principle to obtain the optimal three-dimensional parameter pair. The present invention accurately obtains three-dimensional parameters by parallel two-dimensional parameter estimation and multi-target three-dimensional parameter matching, thereby solving the problem that the prior art cannot break through the Rayleigh resolution, has high complexity and insufficient real-time performance.

Description

Technical Field

[0001] The present invention relates to the field of wireless communication and radar technology, and in particular to a multi-parameter joint estimation method and device integrating communication and perception. Background Art

[0002] Integrated Sensing and Communication (ISAC) has been formally listed as a key research area of ​​the sixth generation wireless communication system. On the one hand, the significance of ISAC is to integrate and share the two functions of communication and sensing into the same system by sharing most of the hardware, signal processing modules, and typical transmission signals. The close cooperation of communication and sensing functions can significantly improve spectrum efficiency and reduce equipment size, cost and power consumption. On the other hand, the communication function enhances the perception function, and the perception function assists the communication function, achieving the effect of enhancing the performance of both. In the design of the ISAC system, advanced multi-dimensional parameter joint estimation technology is the key to improving integration efficiency and enhancing system functions. Compared with one-dimensional parameter joint estimation, multi-dimensional parameter joint estimation has significant advantages. On the one hand, it can make full use of multi-dimensional information to perceive the target and improve the target perception ability. On the other hand, multi-dimensional perception solves the matching problem between multiple parameters of the target.

[0003] However, it is still a challenge to maintain high-resolution multi-parameter joint estimation while reducing complexity, increasing system real-time performance and reducing computing power. The three-dimensional discrete Fourier transform (3D-DFT) algorithm based on 3D-DFT has simple circuit implementation and low complexity, but its resolution is strictly limited by the Rayleigh resolution determined by the antenna, bandwidth and signal time length of the perception system. The existing three-dimensional multiple signal classification algorithm and three-dimensional subspace rotation invariant algorithm in super-resolution algorithms can break through the Rayleigh resolution, but their complexity is very high, especially for the multi-parameter joint estimation of large-scale antennas and multi-carrier systems. The complexity is too high, resulting in insufficient real-time performance.

[0004] Therefore, there is an urgent need for a multi-parameter joint estimation method that can reduce complexity, improve mean square error performance, and be more accurate and efficient in integrating communication and perception. Summary of the invention

[0005] The embodiments of the present invention provide a multi-parameter joint estimation method integrating communication perception to solve the problem that the existing technology cannot break through the Rayleigh resolution, is too complex, and lacks real-time performance. The technical solution is as follows:

[0006] According to one aspect of the present invention, a multi-parameter joint estimation method for integrating communication and perception includes: receiving a time-domain OFDM signal, down-converting, removing cyclic prefixes, discrete-time sampling, and discrete Fourier transforming the time-domain OFDM signal to obtain a discrete three-dimensional signal observation in the frequency domain; smoothing the discrete three-dimensional signal observation in the frequency domain by constructing a reconstructed rotation matrix and a reflection coefficient matrix to obtain a plurality of sub-signal observations; processing the plurality of sub-signal observations by a one-dimensional root-finding MUSIC algorithm to estimate the angles, distances, and speeds of a plurality of targets and matching them to obtain initial parameter pairs; based on the initial parameter pairs, performing subspace iteration and Levenberg-Marquardt convex optimization on the local convex region of each target in the two-dimensional spatial spectrum to obtain a plurality of two-dimensional parameter pairs; matching the initial parameter pair with the two-dimensional parameter pair using the minimum distance principle to obtain an optimal three-dimensional parameter pair that matches the plurality of targets one by one.

[0007] In one embodiment, a discrete three-dimensional signal observation in the frequency domain is obtained. The model is as follows:

[0008] ;

[0009] in, Represent antenna, subcarrier and symbol index respectively; Indicates the total number of targets; Indicates The reflection coefficient of a target indicates the reflection intensity of the target to the signal; is the phase difference between adjacent antennas, and The arrival angle of each target is related to is the phase difference between adjacent subcarriers, The distance to the target is related to is the phase difference between adjacent symbols, and The speed of the target is related to The mean is 0 and the variance is of complex Gaussian noise.

[0010] In one embodiment, the frequency domain discrete three-dimensional signal observation is smoothed by constructing a reconstruction rotation matrix and a reflection coefficient matrix to obtain multiple sub-signal observations. The calculation formula is as follows:

[0011] ;

[0012] in, is the reconstructed rotation matrix, is the reflection coefficient matrix, is the phase rotation vector of the sub-signal observation, is the spatially smoothed noise term.

[0013] In one embodiment, the multiple sub-signal observations are processed by a one-dimensional root-finding MUSIC algorithm, and the angles, distances and speeds of multiple targets are estimated and matched to obtain initial parameter pairs by the following steps: calculating the covariance matrix of the multiple sub-signal observations, performing singular value decomposition on the covariance matrix to obtain a noise subspace; constructing a one-dimensional root function based on the orthogonality of the subspace and the steering vector, quickly searching for the root value of the one-dimensional root function through Pisarenko decomposition, and obtaining the angle, distance and speed based on the root value; using the minimum mean square error criterion, matching the angle, distance and speed to obtain an initial parameter pair.

[0014] In one embodiment, the distance is obtained according to the root value. The calculation formula is as follows:

[0015] ,

[0016] in, is the root value, is the frequency of the signal.

[0017] In one embodiment, the speed is obtained according to the root value. and angle The calculation formula is as follows:

[0018] ,

[0019] in, is the root value, is the propagation speed of the signal.

[0020] In one embodiment, based on the initial parameter pair, subspace iteration and Levenberg-Marquardt convex optimization are performed on the local convex region of each target in the two-dimensional spatial spectrum to obtain multiple two-dimensional parameter pairs, which is achieved by the following steps: spatial and time domain smoothing of the discrete three-dimensional signal observations in the frequency domain to obtain multiple sub-signal observations; constructing an objective function for jointly estimating two-dimensional parameters by minimizing the least squares function of the two-dimensional variables; and optimizing the objective function using the subspace updating Levenberg-Marquardt method to obtain multiple two-dimensional parameter pairs.

[0021] In one embodiment, the least squares function of the two-dimensional variables is minimized. The formula is as follows:

[0022]

[0023] in, , There are three sets of two-dimensional variables, which are calculated by pairing them together. is the noise subspace.

[0024] In one embodiment, the calculation formula for obtaining the optimal three-dimensional parameter pair that matches multiple targets one by one by matching the initial parameter pair with the two-dimensional parameter pair using the minimum distance principle is as follows:

[0025] ,

[0026] Among them, for the three sequences and Re-pair to get the final parameter pair .

[0027] According to one aspect of the present invention, a multi-parameter joint estimation device integrating communication and perception includes: a signal preprocessing module, which is used to receive a time-domain OFDM signal, down-convert the time-domain OFDM signal, remove the cyclic prefix, perform discrete time sampling and discrete Fourier transform on the time-domain OFDM signal, and obtain a discrete three-dimensional signal observation in the frequency domain; a spatial smoothing module, which is used to smooth the discrete three-dimensional signal observation in the frequency domain by constructing a reconstructed rotation matrix and a reflection coefficient matrix to obtain a plurality of sub-signal observations; a parameter initialization module, which is used to process the plurality of sub-signal observations by a one-dimensional root-finding MUSIC algorithm, estimate the angles, distances and speeds of a plurality of targets and match them with each other to obtain an initial parameter pair; a parallel two-dimensional parameter estimation module, which is used to perform subspace iteration and Levenberg-Marquardt convex optimization on the local convex region of each target in the two-dimensional spatial spectrum based on the initial parameter pair, and obtain a plurality of two-dimensional parameter pairs; and a multi-target three-dimensional parameter matching module, which is used to match the initial parameter pair with the two-dimensional parameter pair using the minimum distance principle to obtain an optimal three-dimensional parameter pair that matches the plurality of targets one by one.

[0028] The beneficial effects brought by the technical solution provided by the present invention are:

[0029] In the above technical scheme, the present invention first receives a time-domain OFDM signal, performs down-conversion, cyclic prefix removal, discrete time sampling and discrete Fourier transform on the time-domain OFDM signal to obtain a discrete three-dimensional signal observation in the frequency domain, then performs smoothing processing on the signal to obtain a plurality of sub-signal observations, and then processes the plurality of sub-signal observations through a one-dimensional root-finding MUSIC algorithm, estimates the angles, distances and speeds of a plurality of targets and matches them with each other to obtain initial parameter pairs, and based on the initial parameter pairs, performs subspace iteration and Levenberg-Marquardt convex optimization on the local convex region of each target in the two-dimensional spatial spectrum to obtain a plurality of two-dimensional parameter pairs, and finally uses the minimum distance principle to compare the initial parameter pairs with the two-dimensional parameter pairs. The optimal three-dimensional parameter pairs that match multiple targets one by one are matched. By combining multiple stages such as signal preprocessing, spatial smoothing, initialization, parallel two-dimensional parameter estimation, and multi-target three-dimensional parameter matching, the three-dimensional parameters of angle, distance and speed of each perceived target can be accurately obtained. The existing cellular network signal resources can be directly used to achieve high real-time and high-precision perception, saving system resources while reducing system computing power requirements. Under the premise of not affecting cellular communication performance, low-complexity and high-precision joint angle, distance and speed estimation is achieved, which is easy to realize efficient integration of positioning perception and cellular communication system, thereby effectively solving the problems that the existing technology cannot break through the Rayleigh resolution, is too complex, and lacks real-time performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in describing the embodiments of the present invention. Obviously, the drawings described below are only some embodiments of the present invention, and those skilled in the art can obtain other drawings based on these drawings without creative work.

[0031] Figure 1 is a flow chart of a multi-parameter joint estimation method for integrating communication and perception according to an exemplary embodiment;

[0032] Figure 2 It is a schematic diagram of the implementation environment of a SIMO-OFDM ISAC system application scenario in an application scenario;

[0033] Figure 3 yes Figure 2 A flowchart of a multi-parameter joint estimation method for integrated communication and perception in corresponding application scenarios;

[0034] Figure 4 yes Figure 3 Schematic diagram of spatial smoothing in step S2 in the corresponding embodiment;

[0035] Figure 5is a schematic diagram of performance comparison results of an exemplary embodiment and the prior art;

[0036] Figure 6 yes Figure 5 A schematic diagram showing the complexity comparison between the corresponding embodiment and the prior art;

[0037] Figure 7 It is a block diagram of a multi-parameter joint estimation device integrating communication and perception according to an exemplary embodiment. DETAILED DESCRIPTION

[0038] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be interpreted as limiting the present invention.

[0039] It will be understood by those skilled in the art that, unless expressly stated, the singular forms "one", "said", and "the" used herein may also include plural forms. It should be further understood that the term "comprising" used in the specification of the present disclosure refers to the presence of the 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 refer to an element as being "connected" or "coupled" to another element, it may be directly connected or coupled to the other element, or there may be intermediate elements. In addition, the "connection" or "coupling" used herein may include wireless connection or wireless coupling. The term "and / or" used herein includes all or any unit and all combinations of one or more associated listed items.

[0040] The present invention provides a multi-parameter joint estimation method for integrated communication perception. By combining multiple stages such as signal preprocessing, spatial smoothing, initialization, parallel two-dimensional parameter estimation, and multi-target three-dimensional parameter matching, the three-dimensional parameters of angle, distance, and speed of each perceived target are finally accurately obtained, solving the problem that the prior art cannot break through the Rayleigh resolution, and the complexity is too high and the real-time performance is insufficient. The multi-parameter joint estimation method for integrated communication perception is suitable for a multi-parameter joint estimation device for integrated communication perception, and the multi-parameter joint estimation device for integrated communication perception can be an electronic device. The multi-parameter joint estimation method for integrated communication perception in the embodiment of the present invention can be applied to a variety of scenarios, such as multi-parameter joint estimation for integrated communication perception, etc.

[0041] See also Figure 1 , an embodiment of the present invention provides a multi-parameter joint estimation method integrating communication and perception, which is applicable to electronic devices.

[0042] In the following method embodiments, for ease of description, the execution subject of each step of the method is taken as an electronic device as an example for illustration, but this does not constitute a specific limitation.

[0043] like Figure 1 As shown, the method may include the following steps:

[0044] Step 110, receiving a time-domain OFDM signal, down-converting, removing cyclic prefix, discrete time sampling and discrete Fourier transforming the time-domain OFDM signal to obtain a discrete three-dimensional signal observation in the frequency domain.

[0045] In one possible implementation, the signal observable in the frequency domain is obtained as a discrete three-dimensional signal The model is as follows:

[0046] ;

[0047] in, Represent antenna, subcarrier and symbol index respectively; Indicates the total number of targets; Indicates The reflection coefficient of a target indicates the reflection intensity of the target to the signal; is the phase difference between adjacent antennas, and The arrival angle of each target is related to is the phase difference between adjacent subcarriers, The distance to the target is related to is the phase difference between adjacent symbols, and The speed of the target is related to The mean is 0 and the variance is of complex Gaussian noise.

[0048] Specifically, the time-domain OFDM signal received by the ISAC base station is first down-converted, that is, the carrier frequency of the signal is reduced to the baseband or intermediate frequency, and then the cyclic prefix (CP) is removed to combat the multipath effect. Discrete-time sampling is then performed and the sampled signal is subjected to DFT discrete Fourier transform to obtain a signal in the frequency domain. At this time, the signal is three-dimensional, corresponding to the dimensions of the antenna, subcarrier, and symbol, respectively.

[0049] Step 130 , smoothing the discrete three-dimensional signal observation in the frequency domain by constructing a reconstruction rotation matrix and a reflection coefficient matrix to obtain a plurality of sub-signal observations.

[0050] In one possible implementation, multiple sub-signal observations are obtained. The calculation formula is as follows:

[0051] ;

[0052] in, is the reconstructed rotation matrix, is the reflection coefficient matrix, is the phase rotation vector of the sub-signal observation, is the spatially smoothed noise term.

[0053] Specifically, by constructing a reconstructed rotation matrix and a reflection coefficient matrix, the original frequency domain signal observation is smoothed into multiple sub-signal observations to reduce the mutual influence between different target signals, making the subsequent parameter estimation more accurate.

[0054] Step 150, processing the multiple sub-signal observations by a one-dimensional root-finding MUSIC algorithm, estimating the angles, distances and velocities of the multiple targets and matching them to obtain initial parameter pairs.

[0055] In one possible implementation, the covariance matrix of multiple sub-signal observations is calculated, and the covariance matrix is ​​subjected to singular value decomposition to obtain a noise subspace. A one-dimensional root function is constructed based on the orthogonality of the subspace and the steering vector. The root value of the one-dimensional root function is quickly searched through Pisarenko decomposition, and the angle, distance and speed are obtained based on the root value. The angle, distance and speed are matched with each other using the minimum mean square error criterion to obtain an initial parameter pair.

[0056] In one possible implementation, the distance is obtained according to the root value. The calculation formula is as follows:

[0057] ,

[0058] in, is the root value, is the frequency of the signal.

[0059] In one possible implementation, the speed is obtained according to the root value. and angle The calculation formula is as follows:

[0060] ,

[0061] in, is the root value, is the propagation speed of the signal.

[0062] Step 170 , based on the initial parameter pairs, subspace iteration and Levenberg-Marquardt convex optimization are performed on the local convex region of each target in the two-dimensional spatial spectrum to obtain a plurality of two-dimensional parameter pairs.

[0063] In one possible implementation, the discrete three-dimensional signal observation in the frequency domain is smoothed in the spatial and time domains to obtain multiple sub-signal observations. The objective function for jointly estimating the two-dimensional parameters is constructed by minimizing the least squares function of the two-dimensional variables. The objective function is optimized using the subspace updating Levenberg-Marquardt method to obtain multiple two-dimensional parameter pairs.

[0064] In one possible implementation, the least squares function of the two-dimensional variables is minimized The formula is as follows:

[0065]

[0066] in, , There are three sets of two-dimensional variables that are calculated by pairing them together. is the noise subspace.

[0067] Step 190, using the minimum distance principle, the initial parameter pair is matched with the two-dimensional parameter pair to obtain an optimal three-dimensional parameter pair that matches the multiple targets one by one.

[0068] In a possible implementation, the calculation formula for obtaining the optimal three-dimensional parameter pair that matches multiple targets one by one by matching the initial parameter pair with the two-dimensional parameter pair using the minimum distance principle is as follows:

[0069] ,

[0070] Among them, for the three sequences and Re-pair to get the final parameter pair .

[0071] Through the above process, the present invention first receives a time-domain OFDM signal, performs down-conversion, cyclic prefix removal, discrete time sampling and discrete Fourier transform on the time-domain OFDM signal to obtain a discrete three-dimensional signal observation in the frequency domain, then performs smoothing processing on the signal to obtain a plurality of sub-signal observations, and then processes the plurality of sub-signal observations through a one-dimensional root-finding MUSIC algorithm, estimates the angles, distances and speeds of a plurality of targets and matches them with each other to obtain initial parameter pairs, performs subspace iteration and Levenberg-Marquardt convex optimization on the local convex region of each target in the two-dimensional spatial spectrum based on the initial parameter pairs, obtains a plurality of two-dimensional parameter pairs, and finally uses the minimum distance principle to compare the initial parameter pairs with the two-dimensional parameter pairs. The optimal three-dimensional parameter pairs that match multiple targets one by one are matched. By combining multiple stages such as signal preprocessing, spatial smoothing, initialization, parallel two-dimensional parameter estimation, and multi-target three-dimensional parameter matching, the three-dimensional parameters of angle, distance and speed of each perceived target can be accurately obtained. The existing cellular network signal resources can be directly used to achieve high real-time and high-precision perception, saving system resources while reducing system computing power requirements. Under the premise of not affecting cellular communication performance, low-complexity and high-precision joint angle, distance and speed estimation is achieved, which is easy to realize efficient integration of positioning perception and cellular communication system, thereby effectively solving the problems that the existing technology cannot break through the Rayleigh resolution, is too complex, and lacks real-time performance.

[0072] In one application scenario, Figure 2 A schematic diagram of the implementation environment of a SIMO-OFDM ISAC system application scenario is shown. Figure 3 A flowchart of the multi-parameter joint estimation method that integrates application communication and perception in this application scenario is shown.

[0073] Among them, the SIMO-OFDM ISAC system is a system that combines single-input multiple-output (SIMO) orthogonal frequency division multiplexing (OFDM) technology with integrated communication and perception (ISAC). This system integrates communication and perception functions, aiming to improve spectrum utilization, reduce equipment costs, and improve system performance. It can also adapt to the needs of future 6G and higher-level communication systems. While transmitting communication signals, the system can also perform environmental perception and target detection, and realize the integration of communication and perception functions by sharing hardware and spectrum resources.

[0074] like Figure 2As shown in the figure, the system includes an ISAC base station equipped with 1 transmitting antenna and L receiving antennas and U sensing targets. ISAC Transmit Antennas are the transmitting antennas of the system. In the figure, there are multiple transmitting antennas connected to Target 1-Target U and User Users respectively. These transmitting antennas are responsible for transmitting ISAC transmission signals, which contain both communication data and radar signals for sensing and detecting targets. By using the same frequency band and hardware resources to simultaneously transmit communication and sensing signals, efficient resource utilization is achieved; L Receive Antenna is the receiving antenna of the system, which is responsible for receiving signals from the transmitting antennas. These signals include target echo signals (Target Echo Signal) and user uplink signals (User Uplink Signal). The receiving antenna transmits the received signals to the subsequent processing unit for signal separation, decoding and target detection.

[0075] Target 1 in the figure is connected to the transmitter through an antenna, indicating that these targets are the receiving ends of communication and also the target objects of perception. The targets reflect the ISAC signals sent by the transmitting antenna to generate target echo signals, which are received by the receiving antenna and used for target detection and positioning; users send uplink signals (User UplinkSignal) to the receiving antenna for data transmission or service request.

[0076] The signal types in the implementation environment include ISAC transmission signal: the signal sent by the transmitting antenna, which contains both communication data and perception radar signals; target echo signal: the signal generated after the target reflects the ISAC signal, which is received by the receiving antenna for target detection and positioning; user uplink signal: the uplink signal sent by the user, which is used for data transmission or communication with other devices.

[0077] This ISAC wireless communication system can be applied to a variety of scenarios, such as intelligent transportation systems, drone communication and navigation, target tracking and positioning in the Internet of Things, etc. In these applications, the system can transmit a large amount of data in real time, while achieving accurate perception of the surrounding environment and target detection, providing support for intelligent decision-making. The system realizes the dual functions of communication and perception through efficient resource utilization and signal processing technology, and has broad application prospects in modern communication and radar technology.

[0078] like Figure 3 As shown, the multi-parameter joint estimation method integrating communication and perception may include the following steps:

[0079] Step S1, signal preprocessing.

[0080] The preprocessing includes down-conversion, cyclic prefix removal, discrete time sampling and discrete Fourier transform DFT.

[0081] Specifically, the signal sent by the ISAC system is a 5G NR reference signal. M OFDM symbols and N subcarriers, Δ f is the subcarrier spacing. i The angle, distance and speed of the target are θ i , r i and v i , then in l Antenna, m OFDM symbol n The received signal observations on the subcarriers for:

[0082]

[0083] in, l= 0,..., L -1 ,n= 0,..., N -1, m= 0,..., M -1 is the antenna, subcarrier and symbol index, For the i The target reflection coefficient, are the phase differences between adjacent antennas, adjacent subcarriers, and adjacent symbols, respectively. The mean is 0 and the variance is of complex Gaussian noise.

[0084] Furthermore, the signal observations for all antennas, subcarriers and symbols are concatenated into a column vector:

[0085]

[0086] in,

[0087]

[0088] Step S2: spatial smoothing of signal observation quantities.

[0089] like Figure 4 As shown, the original signal observation matrix symbol is segmented and reorganized on the antenna threshold antenna to generate multiple sub-signal observations (i.e., sub-matrices), each of which can be regarded as a snapshot.

[0090] Specifically, use represents the size of the new sub-signal observation in the spatial domain, frequency domain and time domain. Therefore, the number of spatial smoothing sub-signal observations is and , the number of snapshots ,for , No. The sub-signal observations are generated as follows:

[0091]

[0092] in, ,

[0093] .

[0094] Furthermore, the new signal observation can be expressed as .

[0095] Among them, the reconstructed rotation matrix , , compared with the 0th sub-signal observation, the The phase rotation vector of the sub-signal observation is , is a spatially smooth noise term, which is reshaped in the same way as Consistent, use , and Respectively and .

[0096] Step S3, preliminarily estimate and match the angle, distance and speed parameters of the target.

[0097] Specifically, Prompt is tokenized by Tokenizer and converted into a token sequence that the model can understand. Then, the embedding model is used to convert these token sequences into embedded vectors prompt_embeddings, which contain the semantic information of Prompt.

[0098] Step S4, using the initial parameters to perform subspace iterative convex optimization to obtain three sets of precise estimation parameter pairs.

[0099] Specifically, for the signal observation value after frequency domain spatial smoothing, S=LS f M snapshots, denoted as ,in .

[0100] Therefore the covariance matrix can be calculated as:

[0101]

[0102] in , , .

[0103] Perform singular value decomposition to obtain the noise subspace :

[0104]

[0105] By utilizing the orthogonality of the noise subspace and the steering vector, the following one-dimensional root function can be constructed:

[0106] ,

[0107] Using Pisarenko decomposition, we can quickly search for U roots , and get .

[0108] Similarly, let and , the angle and speed can be obtained as follows:

[0109]

[0110] Then, in order to match the angles, distances and velocities of multiple targets, the target reflection coefficients are first estimated, and then the minimum mean square error criterion is used for the signal observations.

[0111] Will The estimated parameters and Match to The possible combination of targets, the corresponding reconstructed rotation matrix is ​​used , by traversing each Can get Target backscatter coefficient The maximum likelihood estimate of :

[0112]

[0113] use The best pairing combination represented by can be obtained by the minimum mean square error criterion of the following signal observation z:

[0114]

[0115] Then, a two-dimensional precise estimation is performed, and the signal observations after spatial and temporal spatial smoothing have S=SaNSt snapshots. , .

[0116] Therefore, the covariance matrix can be calculated as

[0117] ,

[0118] Through singular value decomposition, we get The noise subspace , and then jointly estimate the angle and velocity parameters by minimizing the least squares function of the following two-dimensional variables:

[0119]

[0120] in, and .

[0121] In order to avoid the extremely high complexity of exhaustive search in traditional 2D-MUSIC algorithm and improve the performance, the objective function is optimized using the In-subspace Update (ISU) Levenberg-Marquardt method.

[0122] Specifically, the initial parameters are set first, including the iteration counter q, the initial optimization point and other parameters. In each iteration, the relevant parameters or matrices are updated according to the current estimated value and the noise subspace, and the Jacobian matrix of the target vector function is calculated. This matrix reflects the sensitivity of the target function to each parameter and is the key information in the optimization process. Then, according to the Levenberg-Marquardt algorithm, the current iteration step is calculated to decide how to move in the parameter space to approach the optimal solution. If a certain convergence condition is met, the iteration is stopped and it is considered that a good enough solution has been found. If the maximum number of iterations is not reached, the iteration continues. If the current iteration step does not meet the convergence condition, the gain ratio is adjusted according to the Levenberg-Marquardt algorithm to adjust the calculation method of the iteration step, and try to search for the optimal solution in a more appropriate direction. When the convergence condition is met or the maximum number of iterations is reached, the final precise estimation result is output: three sequences and , .

[0123] In the above process, the method of optimizing the objective function by using subspace updating and Levenberg-Marquardt algorithm in the embodiment of the present invention aims to improve the performance and accuracy of the traditional 2D-MUSIC algorithm. By iteratively updating parameters and calculating the Jacobian matrix, a more accurate estimation result can be obtained without increasing too much computational complexity.

[0124] Step S5, multi-target three-dimensional parameter matching stage, uses the minimum distance principle to obtain the optimal parameter pair.

[0125] Specifically, according to the minimum distance criterion, the sequence and Re-pair:

[0126] ,

[0127] Get the final parameter pair .

[0128] Through the above steps, the embodiment of the present invention combines multiple stages such as signal preprocessing, spatial smoothing, initialization, parallel two-dimensional parameter estimation, and multi-target three-dimensional parameter matching to finally accurately obtain the three-dimensional parameters of angle, distance and speed of each perceived target. It can directly use the existing cellular network signal resources to achieve high real-time and high-precision perception, save system resources while reducing system computing power requirements, and achieve low-complexity and high-precision joint angle, distance and speed estimation without affecting cellular communication performance. It is easy to achieve efficient integration of positioning perception and cellular communication systems, thereby effectively solving the problems that the existing technology cannot break through the Rayleigh resolution, is too complex, and lacks real-time performance.

[0129] In another application scenario, simulation experiments are used to verify the beneficial effects of the present invention, and the parameters are set as follows: the system carrier frequency is 25 GHz, the subcarrier spacing is 120 kHz, the OFDM data cycle length is 8.33 us, the OFDM cyclic prefix length is 0.59 us, the number of carriers N=88:10:138, and the number of OFDM symbols is M=48:8:88. The number of receiving antennas L=8:2:36, and the number of perceived targets U=3. The angle, distance, and speed of the three targets are (20°, 39.73m, -10m / s), (-23.16°, 60.5m, 29.61m / s), and (-10.6°, 80.21m, 10.11m / s).

[0130] Figure 5 In the figure, L = 16, N = 128 and M = 80 are fixed, which correspond to the comparison results of the joint angle, distance and speed estimation of the present invention and the existing 3D-DFT, 3D-MUSIC and 3D-ESPRIT algorithms, respectively. Compared with the 3D-DFT algorithm, the RMSE of the present invention is reduced by at least 2 orders of magnitude. The performance of the 3D-ESPRIT algorithm and the 3D-MUSIC are almost the same, while the performance of the present invention is better than the two existing super-resolution algorithms and is closer to RCRB. Specifically, when the RMSE is 10-2 standard units, the signal-to-noise ratio gains of the proposed algorithm for angle measurement, distance measurement and speed measurement are 1.25dB, 0.7dB and 0.75dB, respectively.

[0131] Figure 6The average running time of the present invention and the existing 3D-DFT, 3D-MUSIC and 3D-ESPRIT algorithms is compared with the number of antennas, subcarriers and symbols. When the number of resources increases, the complexity of the present invention is more significantly improved than that of the 3D-ESPRIT algorithm and the 3D-MUSIC algorithm. Specifically, when the number of antennas, the number of subcarriers and the number of symbols are 32, 128 and 80 respectively, the time consumption of the present invention is reduced by 2242 times and 149 times compared with the 3D-MUSIC algorithm and the 3D-ESPRIT algorithm respectively.

[0132] Through the above process, the RMSE of the present invention is reduced by at least 2 orders of magnitude compared with the commonly used 3D-DFT algorithm. Compared with the mainstream super-resolution 3D-MUSIC and 3D-ESPRIT algorithms, the present invention improves the performance while reducing the complexity by about 3 orders of magnitude and 2 orders of magnitude respectively. At the same time, the present invention can use the reference signal of the cellular system as a perception signal, and can also use the modulated communication signal as a perception signal. Under the premise of not affecting the cellular communication performance, high-precision perception can be achieved. The existing cellular network signal resources can be directly used to achieve high real-time and high-precision perception, saving system resources while reducing the system computing power requirements. Under the premise of not affecting the cellular communication performance, low-complexity and high-precision joint angle, distance and speed estimation are achieved, and it is easy to realize the efficient integration of positioning perception and cellular communication system, so as to effectively solve the problem that the existing technology cannot break through the Rayleigh resolution, and the complexity is too high and the real-time performance is insufficient.

[0133] The following is an embodiment of the device of the present invention, which can be used to execute the communication perception integrated multi-parameter joint estimation method involved in the present invention. For details not disclosed in the embodiment of the device of the present invention, please refer to the method embodiment of the communication perception integrated multi-parameter joint estimation method involved in the present invention.

[0134] See also Figure 7 In an embodiment of the present invention, a multi-parameter joint estimation device 800 integrating communication and perception is provided.

[0135] The communication-sensing integrated multi-parameter joint estimation device 800 includes but is not limited to: a signal preprocessing module 810 , a spatial smoothing module 830 , a parameter initialization module 850 , a parallel two-dimensional parameter estimation 870 and a multi-target three-dimensional parameter matching module 890 .

[0136] The signal preprocessing module 810 is used to receive a time-domain OFDM signal, perform down-conversion, cyclic prefix removal, discrete time sampling and discrete Fourier transformation on the time-domain OFDM signal, and obtain a discrete three-dimensional signal observable in the frequency domain.

[0137] The spatial smoothing module 830 is used to smooth the discrete three-dimensional signal observation in the frequency domain by constructing a reconstructed rotation matrix and a reflection coefficient matrix to obtain a plurality of sub-signal observations.

[0138] The parameter initialization module 850 is used to process multiple sub-signal observations through a one-dimensional root-finding MUSIC algorithm, estimate the angles, distances and speeds of multiple targets, and match them to obtain initial parameter pairs.

[0139] The parallel two-dimensional parameter estimation module 870 is used to perform subspace iteration and Levenberg-Marquardt convex optimization on the local convex region of each target in the two-dimensional spatial spectrum based on the initial parameter pair to obtain multiple two-dimensional parameter pairs.

[0140] The multi-target three-dimensional parameter matching module 890 is used to match the initial parameter pair with the two-dimensional parameter pair using the minimum distance principle to obtain the optimal three-dimensional parameter pair that matches multiple targets one by one.

[0141] It should be noted that, the multi-parameter joint estimation with integrated communication perception provided in the above embodiment is only illustrated by the division of the above-mentioned functional modules. In actual applications, the above-mentioned functions can be assigned to different functional modules as needed, that is, the internal structure of the multi-parameter joint estimation device with integrated communication perception will be divided into different functional modules to complete all or part of the functions described above.

[0142] In addition, the embodiments of the multi-parameter joint estimation device with integrated communication perception and the multi-parameter joint estimation method with integrated communication perception provided in the above embodiments belong to the same concept, and the specific manner in which each module performs operations has been described in detail in the method embodiments and will not be repeated here.

[0143] Compared with the related art, the present invention has the following beneficial effects:

[0144] 1. The present invention first receives a time-domain OFDM signal, performs down-conversion, cyclic prefix removal, discrete time sampling and discrete Fourier transform on the time-domain OFDM signal to obtain a discrete three-dimensional signal observation in the frequency domain, then performs smoothing processing on the signal to obtain a plurality of sub-signal observations, and then processes the plurality of sub-signal observations through a one-dimensional root-finding MUSIC algorithm, estimates the angles, distances and speeds of a plurality of targets and matches them to obtain initial parameter pairs, performs subspace iteration and Levenberg-Marquardt convex optimization on a local convex region of each target in a two-dimensional spatial spectrum based on the initial parameter pairs, obtains a plurality of two-dimensional parameter pairs, and finally matches the initial parameter pairs with the two-dimensional parameter pairs using the minimum distance principle to obtain The optimal three-dimensional parameter pairs that match multiple targets one by one are obtained. By combining multiple stages such as signal preprocessing, spatial smoothing, initialization, parallel two-dimensional parameter estimation, and multi-target three-dimensional parameter matching, the three-dimensional parameters of angle, distance, and speed of each perceived target can be accurately obtained. The existing cellular network signal resources can be directly used to achieve high real-time and high-precision perception, saving system resources while reducing system computing power requirements. Under the premise of not affecting cellular communication performance, low-complexity and high-precision joint angle, distance, and speed estimation is achieved, which is easy to realize efficient integration of positioning perception and cellular communication systems, thereby effectively solving the problems that the existing technology cannot break through the Rayleigh resolution, is too complex, and lacks real-time performance.

[0145] 2. The present invention can realize the joint estimation of angle, distance and speed under the requirements of high real-time performance and high precision, thereby improving the perception accuracy. The present invention avoids the exhaustive search in the traditional 2D-MUSIC algorithm and significantly reduces the algorithm complexity. Compared with the prior art, the complexity is reduced by several orders of magnitude.

[0146] 3. The present invention has significant advantages in accuracy and complexity in ultra-large-scale MIMO arrays and ultra-high bandwidth scenarios from 5G to 6G, and is suitable for the needs of future wireless communication systems. It is not only suitable for SIMO-OFDM ISAC systems, but can also be expanded and optimized as needed to adapt to different communication and perception needs.

[0147] It should be understood that, although the steps in the flowchart of the accompanying drawings are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.

[0148] The above descriptions are only some embodiments of the present invention. It should be pointed out that, for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A multi-parameter joint estimation method integrating communication and perception, characterized in that: The method comprises: Receiving a time-domain OFDM signal, down-converting, removing cyclic prefixes, discrete-time sampling, and discrete Fourier transforming the time-domain OFDM signal to obtain a discrete three-dimensional signal observation in the frequency domain; The discrete three-dimensional signal observation in the frequency domain is smoothed by constructing a reconstructed rotation matrix and a reflection coefficient matrix to obtain a plurality of sub-signal observations; The multiple sub-signal observations are processed by a one-dimensional root-finding MUSIC algorithm to estimate the angles, distances and velocities of multiple targets and match them to obtain initial parameter pairs; Based on the initial parameter pairs, subspace iteration and Levenberg-Marquardt convex optimization are performed on the local convex region of each target in the two-dimensional spatial spectrum to obtain multiple two-dimensional parameter pairs; The initial parameter pair is matched with the two-dimensional parameter pair using the minimum distance principle to obtain an optimal three-dimensional parameter pair that matches multiple targets one by one.

2. The communication-aware integrated multi-parameter joint estimation method according to claim 1, characterized in that: The obtained frequency domain discrete three-dimensional signal observation quantity The model is as follows: ； in, Represent antenna, subcarrier and symbol index respectively; Indicates the total number of targets; Indicates The reflection coefficient of a target indicates the reflection intensity of the target to the signal; is the phase difference between adjacent antennas, and The arrival angle of each target is related to is the phase difference between adjacent subcarriers, The distance to the target is related to is the phase difference between adjacent symbols, and The speed of the target is related to The mean is 0 and the variance is of complex Gaussian noise.

3. The communication-aware integrated multi-parameter joint estimation method according to claim 1, characterized in that: The method smoothes the discrete three-dimensional signal observation quantity in the frequency domain by constructing a reconstructed rotation matrix and a reflection coefficient matrix to obtain a plurality of sub-signal observation quantities. The calculation formula is as follows: ； in, is the reconstructed rotation matrix, is the reflection coefficient matrix, is the phase rotation vector of the sub-signal observation, is the spatially smoothed noise term.

4. The communication-aware integrated multi-parameter joint estimation method according to claim 1, characterized in that: The method of processing the plurality of sub-signal observations by a one-dimensional root-finding MUSIC algorithm, estimating the angles, distances and velocities of the plurality of targets and matching them to obtain initial parameter pairs includes: Calculating a covariance matrix of the plurality of sub-signal observations, and performing singular value decomposition on the covariance matrix to obtain a noise subspace; constructing a one-dimensional root function according to the orthogonality of the subspace and the steering vector, quickly searching for the root value of the one-dimensional root function through Pisarenko decomposition, and obtaining the angle, distance and speed according to the root value; The angle, distance and speed are matched with each other to obtain an initial parameter pair using a minimum mean square error criterion.

5. The communication-aware integrated multi-parameter joint estimation method according to claim 4, characterized in that: The calculation formula for obtaining the distance according to the root value is as follows: ， in, is the root value of the one-dimensional root function, is the frequency of the signal, For the The estimated distance to the target.

6. The communication-aware integrated multi-parameter joint estimation method according to claim 1, characterized in that: Based on the initial parameter pair, subspace iteration and Levenberg-Marquardt convex optimization are performed on the local convex region of each target in the two-dimensional spatial spectrum to obtain multiple two-dimensional parameter pairs, including: Performing spatial and temporal smoothing on the discrete three-dimensional signal observation in the frequency domain to obtain a plurality of sub-signal observations; The objective function of jointly estimating two-dimensional parameters is constructed by minimizing the least square function of two-dimensional variables; The objective function is optimized using the subspace updating Levenberg-Marquardt method to obtain a plurality of two-dimensional parameter pairs.

7. A multi-parameter joint estimation device integrating communication and perception, characterized in that: The device comprises: A signal preprocessing module is used to receive a time-domain OFDM signal, perform down-conversion, remove cyclic prefix, discrete time sampling and discrete Fourier transform on the time-domain OFDM signal, and obtain a discrete three-dimensional signal observation in the frequency domain; A spatial smoothing module, used for smoothing the discrete three-dimensional signal observation in the frequency domain to obtain a plurality of sub-signal observations by constructing a reconstructed rotation matrix and a reflection coefficient matrix; A parameter initialization module is used to process the multiple sub-signal observations through a one-dimensional root-finding MUSIC algorithm, estimate the angles, distances and speeds of multiple targets and match them to obtain initial parameter pairs; A parallel two-dimensional parameter estimation module is used to perform subspace iteration and Levenberg-Marquardt convex optimization on the local convex region of each target in the two-dimensional spatial spectrum based on the initial parameter pair to obtain multiple two-dimensional parameter pairs; The multi-target three-dimensional parameter matching module is used to match the initial parameter pair with the two-dimensional parameter pair using the minimum distance principle to obtain the optimal three-dimensional parameter pair that matches multiple targets one by one.

Images