MIMO radar multi-target detection method and system based on strong target constraint

Abstract

The present application relates to a MIMO radar multi-target detection method and system based on strong target restriction, which designs a beam weight matrix solving method with strong target restriction on the basis of a MIMO radar multi-target detection method based on reinforcement learning, takes out possible strong targets from the detection results of the MIMO radar, calculates the power gain restriction values corresponding to each strong target, designs a beam optimization scheme with strong target restriction and solves the beam weight matrix by using a convex optimization toolbox. During detection, the radar waveform corresponding to the transmission beam weight matrix is used, the Wald type statistics of the radar echo is calculated, the state calculation and action selection are performed, the angle unit to which the beam pattern of the MIMO radar needs to be focused is selected, and finally the beam weight matrix of the next detection time is quickly solved according to the beam weight matrix solving method with strong target restriction, so as to improve the detection performance of the MIMO radar on weak targets.

Description

Technical Field

[0001] This invention belongs to the field of radar target detection technology, and relates to a MIMO radar multi-target detection method and system based on strong target constraints. Background Technology

[0002] With the increasing application of automotive radar, wireless communication (such as millimeter-wave radar in 5G networks), and drones in intelligent scenarios, MIMO (Multiple Input Multiple Output) radar is also being used more and more. Currently, reinforcement learning (RL) has been applied to multi-target detection tasks in MIMO radar. Through reinforcement learning, MIMO radar systems can adaptively adjust their configuration and operating parameters to cope with complex and dynamically changing environments, thereby achieving more accurate and efficient multi-target detection. This method not only improves detection performance but also effectively reduces human intervention and optimizes system resource usage. However, in existing beam optimization schemes, the power gain in the optimized beam pattern is basically consistent across the angular cells that need to be focused, resulting in insufficient detection performance for weak targets. Summary of the Invention

[0003] To address the problems existing in the above-mentioned traditional methods, this invention proposes a MIMO radar multi-target detection method based on strong target constraints, a MIMO radar multi-target detection system based on strong target constraints, a radar device, and a computer-readable storage medium, which can effectively improve the detection performance of weak targets.

[0004] To achieve the above objectives, the embodiments of the present invention adopt the following technical solutions:

[0005] On the one hand, a multi-target detection method for MIMO radar based on strong target constraints is provided, including the following steps:

[0006] Initialize the current detection time and the beam weight matrix of the MIMO radar;

[0007] Transmit the radar waveform corresponding to the beam weight matrix and receive the radar echoes of each angle unit, and calculate the Wald-type statistics of the radar echoes.

[0008] Calculate the state of the Markov decision process based on the Wald-type statistics of the radar echo;

[0009] Choose the action of the Markov decision process based on the state;

[0010] When the detection time is 0, the angle element that the MIMO radar beam pattern needs to focus on is selected according to the action;

[0011] Based on the method for solving the beam weight matrix with strong target constraints, the beam weight matrix at the next detection time is solved based on the angle element; the beam weight matrix at the next detection time is used for MIMO radar multi-target detection at the next detection time.

[0012] On the other hand, a MIMO radar multi-target detection system based on strong target constraint is also provided, including:

[0013] The initialization module is used to initialize the current detection time and the beam weight matrix of the MIMO radar.

[0014] The echo calculation module is used to transmit the radar waveform corresponding to the beam weight matrix and receive the radar echoes of each angle unit, and calculate the Wald-type statistics of the radar echoes.

[0015] The state calculation module is used to calculate the state of the Markov decision process based on the Wald-type statistics of the radar echo;

[0016] The action selection module is used to select actions for the Markov decision process based on the state.

[0017] Angle selection module is used to select the angle unit that the MIMO radar beam pattern needs to focus on when the detection time is 0, based on the action.

[0018] The weighting solution module is used to solve the beam weight matrix at the next detection time based on the beam weight matrix solution method with strong target constraints, using angle elements; the beam weight matrix at the next detection time is used for MIMO radar multi-target detection at the next detection time.

[0019] On another front, a radar device is also provided, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps of the above-mentioned MIMO radar multi-target detection method based on strong target constraints.

[0020] Furthermore, a computer-readable storage medium is also provided, on which a computer program is stored, which, when executed by a processor, implements the steps of the aforementioned MIMO radar multi-target detection method based on strong target constraints.

[0021] One of the above technical solutions has the following advantages and beneficial effects:

[0022] The aforementioned MIMO radar multi-target detection method and system based on strong target constraints further explores reinforcement learning-based MIMO radar multi-target detection methods. A beam weight matrix solution method with strong target constraints is designed. This method extracts potential strong targets from the MIMO radar detection results, calculates the power gain constraint value for each strong target, designs a beam optimization scheme with strong target constraints, and uses a convex optimization toolbox to obtain the beam weight matrix. For each detection time, the radar waveform corresponding to the beam weight matrix is ​​transmitted, and radar echoes from each angle cell are received. After calculating the Wald-type statistics of the radar echoes, state calculation and action selection are performed. Then, the angle cells that need to be focused on in the MIMO radar beam pattern are selected. Finally, based on the beam weight matrix solution method with strong target constraints, the beam weight matrix for the next detection time is quickly obtained from the angle cells, thereby improving the MIMO radar's detection performance for weak targets. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention or the conventional technology, the drawings used in the description of the embodiments or the conventional technology 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.

[0024] Figure 1 This is a flowchart illustrating a MIMO radar multi-target detection method based on strong target constraints in one embodiment;

[0025] Figure 2 The flowchart shows an existing MIMO radar multi-target detection method based on reinforcement learning.

[0026] Figure 3 This is a flowchart illustrating a method for solving the beam weighting matrix with strong target constraints in one embodiment.

[0027] Figure 4 A comparison of beam patterns obtained by the improved optimization scheme based on the internal convex approximation in one embodiment and the original optimization scheme based on the internal convex approximation is shown.

[0028] Figure 5 This is a schematic diagram illustrating the application process of a MIMO radar multi-target detection method based on strong target constraints in one embodiment;

[0029] Figure 6 This is a schematic diagram comparing the beam patterns obtained by two beam optimization schemes in one embodiment.

[0030] Figure 7 Here is the normalized power spectral density function of clutter in one embodiment;

[0031] Figure 8 This is a schematic diagram comparing the performance of ablation experiments in one embodiment, where (a) is the detection probability change curve with and without prior information assistance, and (b) is the detection probability change curve with and without strong target restriction;

[0032] Figure 9 This is a schematic diagram comparing detection performance in one embodiment, where (a) represents the detection performance of the method proposed in this invention, and (b) represents the detection performance of a classic method in the art.

[0033] Figure 10 This is a schematic diagram of the module architecture of a MIMO radar multi-target detection system based on strong target constraints in one embodiment. Detailed Implementation

[0034] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention.

[0035] It should be noted that, in this document, the reference to "embodiment" means that a specific feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The presentation of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. Those skilled in the art will understand that the embodiments described herein can be combined with other embodiments. The term "and / or" as used in this specification and the appended claims refers to any combination of one or more of the associated listed items, and all possible combinations thereof.

[0036] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0037] Collocated MIMO radar is a type of multiple-input multiple-output (MIMO) radar system with multiple co-located (i.e., antennas located in the same or close proximity) transmit and receive antenna arrays. Unlike traditional radar systems, MIMO radar utilizes multi-antenna technology to enhance target detection, localization, and identification capabilities by transmitting different signals in multiple directions. Collocated MIMO radar is widely used in automotive radar (e.g., vehicle collision avoidance systems), wireless communication (e.g., millimeter-wave radar in 5G networks), and drone detection and tracking. By utilizing collocated MIMO technology, radar systems can provide higher resolution, stronger multi-target detection capabilities, and greater system robustness. The MIMO radar used in this invention is a collocated MIMO radar, meaning that the transmit and receive antennas are closely arranged, and the element spacing is set to half a wavelength.

[0038] In one embodiment, the target's azimuth angle relative to the MIMO radar can be denoted as ψ, then the transmission steering vector a T (ψ), receiving the guiding vector a R (ψ) can be defined as:

[0039]

[0040] Where, N T N is the number of transmitting antennas in a MIMO radar. R This refers to the number of receiving antennas in a MIMO radar. T This represents the transpose operation of a matrix or vector. Let's denote the radar transmit signal of a MIMO radar. It is generated by s(p) = WΦ(p), where, It is N T A vector composed of orthogonal signals. It is the beam weighting matrix of the MIMO radar. It is the set of radar transmitted signals. The transmitted beam pattern B(ψ) of a MIMO radar can be defined as:

[0041]

[0042] in,(·) * The conjugate transpose operation represents a matrix or vector. (·) H This represents the conjugate transpose operation of a matrix or vector. Furthermore, the trace of matrix R, tr(R), is equal to P. T P T This represents the transmit power of the MIMO radar.

[0043] Assume that the azimuth vector θ = {θ1, θ2, ..., θ} of the target (detected by radar) is known. K The traditional optimization scheme can then be described as follows:

[0044]

[0045] Here, max represents the maximum value, min represents the minimum value, and st represents the constraint condition. By introducing auxiliary variables... The above traditional optimization scheme can be equivalently transformed into:

[0046]

[0047] Where (i) represents constraint one, (ii) represents constraint two, (iii) represents constraint three, and K represents the total number of azimuth angles in the azimuth vector. The observation area of ​​the MIMO radar is uniformly divided into L angular units, denoted as ψ1, ψ2, ..., ψ L .

[0048] In one embodiment, such as Figure 1 As shown, a multi-target detection method for MIMO radar based on strong target constraint is provided, which may include the following processing steps S10 to S20:

[0049] S10, Initialize the current detection time and the beam weight matrix of the MIMO radar;

[0050] S12, transmit the radar waveform corresponding to the beam weight matrix and receive the radar echo of each angle unit, and calculate the Wald type statistics of the radar echo.

[0051] S14, calculate the state of the Markov decision process based on the Wald-type statistics of the radar echo;

[0052] S16, Select the action of the Markov decision process based on the state;

[0053] S18, when the detection time is 0, select the angle unit that the MIMO radar beam pattern needs to focus on according to the action;

[0054] S20, based on the beam weight matrix solution method with strong target constraint, the beam weight matrix at the next detection time is solved based on the angle element; the beam weight matrix at the next detection time is used for MIMO radar multi-target detection at the next detection time.

[0055] Understandably, to more intuitively demonstrate the improvements of this invention, the following is a brief introduction to the existing reinforcement learning-based MIMO radar multi-target detection method. This process is actually a Markov Decision Process (MDP), as follows: Figure 2 It is mainly constructed from four elements: state, policy, action, and reward. Furthermore, MDP is a sequential process, therefore... Figure 2 Subscripts are used to represent the time corresponding to variable t.

[0056] Suppose that at time t, the radar receives echoes from L angular units, denoted as follows: l = 1, 2, ..., L. Then, perform the following steps one by one:

[0057] Step (1), calculate Wald-type statistics for l = 1, 2, ..., L are denoted as

[0058]

[0059] in, Representing Kronecker product, W t This represents the beam weighting matrix used by the radar at time t. The definition can be as follows:

[0060]

[0061] represent The element corresponding to the i-th row and j-th column. This represents rounding down to the nearest integer. Representative vector The i-th element, and:

[0062]

[0063] Where ||·||2 represents the 2-norm of the orientation quantity.

[0064] Step (2), calculate state s according to the following formula. t :

[0065]

[0066] Where min{·} represents taking the minimum value of the elements in the sequence, and H is the maximum number of detectable targets for the radar. It is a threshold value, which is based on the pre-set false alarm rate P. FA get, u(·) represents the step response, if Returns 1 if the value is 1, otherwise returns 0. State space. Defined as

[0067] Step (3): Select action a according to the ε-greedy strategy. t :

[0068]

[0069] in, Let be the action space, where an action represents the number of targets that may be present in the radar detection area. Q represents the state-value function, which is modeled here as a matrix of size (H+1)×H, and Q(s,a) represents the value of taking action a in state s.

[0070] Step (4), select the angle unit that needs to be focused on the beam pattern according to the following formula:

[0071]

[0072] in, a t argmax(Λ t ) represents taking Λ t a t The angle element corresponding to the maximum value.

[0073] Step (5), according to θ t Design a beam optimization scheme as shown in equation (5), and use the convex optimization toolbox to solve for W. t+1 .

[0074] Step (6), calculate the reward r t :

[0075]

[0076] Where ψ={ψ1,ψ2,…,ψ L}, It is the angle element detected by the radar at time t, and:

[0077]

[0078] Here, Q1 represents the first-order Markum function.

[0079] Step (7), jump to step (1), and obtain s following the same steps. t+1 a t+1 r t+1 .

[0080] Step (8), update Q according to the following formula:

[0081] Q(s t ,a t )←Q(s t ,a t )+β·(r t+1 +γ·Q(s t+1 ,a t+1 )-Q(s t ,a t))(14)

[0082] Where β, γ∈[0,1] are called the learning rate and discount factor, respectively.

[0083] Based on this, this embodiment designs a method for solving the beam weighting matrix with strong target constraints, such as... Figure 3 As shown, the process may include the following steps S01 to S05:

[0084] S01, extract possible strong targets from the detection results of the MIMO radar.

[0085] It is understandable that when searching for possible strong targets at time t, the set of angle cells containing the strong targets is initialized in the detection results of the MIMO radar. Let k = 1; find θ k The index value in ψ is denoted as l. k , that is, θ k =ψ(l k );like This is a pre-set detection probability threshold, set here. If k = K, end the process; otherwise, set k = k + 1 and proceed to find θ. k The steps for indexing values ​​in ψ.

[0086] S03, calculate the power gain limit value corresponding to each strong target.

[0087] It is understandable that when calculating the power gain limit value corresponding to each strong target at time t, it is generally possible to denote the value as follows: M≤K. Let m=1, find θ m The index value in ψ is denoted as c. m , that is, θ m =ψ(c m ); θ is calculated according to the following formula m Power gain limit

[0088]

[0089] in,

[0090]

[0091]

[0092] also, Alternatively, it can be obtained by looking up a table; in this embodiment, we take...

[0093] If m = M, end the process; otherwise, set m = m + 1 and proceed to find θ.m The steps for indexing values ​​in ψ.

[0094] S05, Design an improved beam optimization scheme with strong target constraints and solve it using the convex optimization toolbox.

[0095] The original beam optimization scheme takes the following form:

[0096]

[0097] in,

[0098] The above scheme (i) is a non-convex problem, which can be improved by the internal convex approximation method to the following form:

[0099]

[0100] in,

[0101]

[0102] It is of size N T ×N T The identity matrix, and:

[0103]

[0104] Solution (ii) can be solved directly using the CVX toolbox, but due to f k For W to be a quadratic convex function, we must have: Therefore, the power gain of the beam pattern obtained by scheme (ii) on strong targets must be higher than the preset value. To obtain a more reasonable beam pattern, the following is a feasible beam optimization scheme with strong target constraints, improved from equation (ii):

[0105]

[0106] Suppose that at a certain time t, we obtain θ t ={ν=-0.4,-0.25,0.05,0.25}, N T =32, P T =1. For example... Figure 4 As shown, the beam patterns obtained by the improved optimization scheme based on the convex approximation (Equation (18)) and the original optimization scheme based on the convex approximation (Equation (ii)) are compared. The unit of power gain is dB, and the unit of spatial frequency is (1 / m) or no unit (after normalization). The same applies to the other figures below. Obviously, the beam optimization scheme corresponding to Equation (18) is more reasonable.

[0107] Next, the beam weight matrix W is obtained by solving equation (18) using the convex optimization toolbox (i.e., the CVX toolbox in MATLAB). t .

[0108] Based on the aforementioned method for solving the beam weighting matrix with strong target constraints, the MIMO radar multi-target detection method based on strong target constraints provided in this embodiment can be specifically described as follows: Figure 5 As shown, the processing steps may include the following:

[0109] S10, Initialize the current detection time t=0 and the beam weight matrix of the MIMO radar.

[0110] S12, Transmit Beam Weighting Matrix W t The corresponding radar waveforms are received, and radar echoes from each angle unit are also received. l = 1, 2, ..., L, calculate the Wald type statistic of the radar echo according to equation (6).

[0111] S14, calculate the state s of the Markov decision process according to equation (9) based on the Wald-type statistics of the radar echo. t ;

[0112] S16, select action a of the Markov decision process according to equation (10) based on the state. t ;

[0113] S18, if the detection time t = 0, then select the angle element θ that the beam pattern needs to be focused on according to equation (11). t .

[0114] S20, Solve the beam weight matrix W for the next detection time according to the beam weight matrix solution method with strong target constraint. t+1 .

[0115] The aforementioned MIMO radar multi-target detection method based on strong target constraints further explores reinforcement learning-based MIMO radar multi-target detection methods. It proposes a beam weight matrix solution method with strong target constraints. This method extracts potential strong targets from the MIMO radar detection results, calculates the power gain constraints for each strong target, designs a beam optimization scheme with strong target constraints, and uses a convex optimization toolbox to solve for the beam weight matrix. For each detection time, the radar waveform corresponding to the beam weight matrix is ​​transmitted, and radar echoes from each angle cell are received. After calculating the Wald-type statistics of the radar echoes, state calculations and action selection are performed. Then, the angle cells that need to be focused on in the MIMO radar beam pattern are selected. Finally, based on the beam weight matrix solution method with strong target constraints, the beam weight matrix for the next detection time is quickly solved using the angle cells, thereby improving the MIMO radar's detection performance for weak targets.

[0116] In one embodiment, prior to step S20, the aforementioned MIMO radar multi-target detection method based on strong target constraints may further include the following processing steps:

[0117] When the detection time is greater than 0, the angle element that the MIMO radar beam pattern needs to focus on is selected based on the action and the prior information of the previous time.

[0118] Calculate the reward of the Markov decision process based on the Wald-type statistics of the radar echo;

[0119] Update the state-value function based on the action.

[0120] It is understandable that if the detection time t is not t=0, then according to the following formula (21), combined with the prior information of the previous time (i.e. Λ), t-1 To select the angle element that needs to be focused on in the beam pattern:

[0121]

[0122] In this way, the proposed method (i.e., the aforementioned MIMO radar multi-target detection method based on strong target constraint) can be prevented from degrading the detection performance of strong targets.

[0123] If the detection time t > 0, then the reward r of the Markov decision process is calculated according to equation (12). t .

[0124] If the detection time t > 0, then update the state-value function Q according to equation (14).

[0125] Then proceed to step S20 as described above.

[0126] In one embodiment, after step S20, the above-described MIMO radar multi-target detection method based on strong target constraints may further include the following processing steps:

[0127] The process ends when the detection time equals the number of time steps for MIMO radar multi-target detection.

[0128] It can be understood that if the detection time t = T, where T is the time step of MIMO radar multi-target detection, the process ends; otherwise, the detection time moves to the next detection time, i.e., t ← t+1, and proceeds to step S11. In this way, accurate detection process control can be achieved, avoiding errors in the calculation process and saving computing resources.

[0129] In one embodiment, to more intuitively illustrate the effectiveness of the aforementioned MIMO radar multi-target detection method based on strong target constraints, an experimental example is provided. Those skilled in the art should understand that this experimental example is merely illustrative and intended to aid in understanding the aforementioned MIMO radar multi-target detection method based on strong target constraints, and is not the sole limitation thereof.

[0130] The observation area of ​​the MIMO radar is uniformly divided into L = 20 angular units, and the element spacing d is set to half wavelength λ2. The spatial frequency ν is defined as follows:

[0131]

[0132] Each angular unit can then be represented by a unique spatial frequency, namely {-0.5,-0.45,…,0.45}.

[0133] Suppose that at a certain time t, we obtain θ t ={ν=-0.4,-0.25,0.05,0.25}, N T =32. Figure 6 By comparing the beam patterns obtained from the two beam optimization schemes, it can be seen that the beam optimization scheme with strong target limitation can effectively limit the power gain on strong targets, thereby improving the power gain on weak targets.

[0134] Assume there are four targets located at {ν = -0.4, -0.2, -0.05, 0.35}, all with a signal-to-clutter ratio of -9 dB. β = 0.5, γ = 0.5, H = 6, ε = 0.5, P... T =1. N T =N R =32, P FA =10 -4 The normalized power spectral density function of clutter is as follows: Figure 7 As shown.

[0135] like Figure 8 The diagram showing the performance comparison of ablation experiments illustrates the detection probability changes of the proposed method with and without prior information, and with and without strong target constraints. It can be observed that the detection performance of the proposed method significantly decreases without prior information. While the detection performance of the proposed method slightly improves for strong targets without strong target constraints, it significantly decreases for weak targets. (a) shows the detection probability change curve with and without prior information, and (b) shows the detection probability change curve with and without strong target constraints. The unit of time can be milliseconds.

[0136] The experimental parameters are the same as above, such as... Figure 9 The detection performance comparison shown also compares the detection performance of the method proposed in this invention with that of classical methods in the field. (a) shows the detection performance of the method proposed in this invention, and (b) shows the detection performance of classical methods in the field. The unit of time can be milliseconds. It can be seen that the performance of the method proposed in this invention is better than that of classical methods.

[0137] It should be understood that, although Figure 1 The steps are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order in which these steps are executed; they can be performed in other orders. Figure 1 At least some of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.

[0138] In one embodiment, such as Figure 10As shown, a MIMO radar multi-target detection system 100 based on strong target constraints is provided, which may include an initialization module 11, an echo calculation module 13, a state calculation module 15, an action selection module 17, an angle selection module 19, and a weighting solution module 21. The initialization module 11 is used to initialize the current detection time and the beam weight matrix of the MIMO radar. The echo calculation module 13 is used to transmit the radar waveform corresponding to the beam weight matrix and receive the radar echoes from each angle element, calculating the Wald-type statistics of the radar echoes. The state calculation module 15 is used to calculate the state of the Markov decision process based on the Wald-type statistics of the radar echoes. The action selection module 17 is used to select the action of the Markov decision process based on the state. The angle selection module 19 is used to select the angle element that the MIMO radar beam pattern needs to focus on when the detection time is 0, based on the action. The weighting solution module 21 is used to solve the beam weight matrix at the next detection time based on the angle element according to the beam weight matrix solution method with strong target constraints; the beam weight matrix at the next detection time is used for MIMO radar multi-target detection at the next detection time.

[0139] The aforementioned MIMO radar multi-target detection system 100 based on strong target constraints further studies the reinforcement learning-based MIMO radar multi-target detection method by designing a beam weight matrix solution method with strong target constraints. This method extracts potential strong targets from the MIMO radar detection results, calculates the power gain constraint value corresponding to each strong target, designs a beam optimization scheme with strong target constraints, and uses a convex optimization toolbox to obtain the beam weight matrix. For each detection time, the system transmits the radar waveform corresponding to the beam weight matrix and receives radar echoes from each angle element. After calculating the Wald-type statistics of the radar echoes, it performs state calculations and action selection, then selects the angle elements that the MIMO radar beam pattern needs to focus on. Finally, based on the beam weight matrix solution method with strong target constraints, the system quickly solves for the beam weight matrix at the next detection time, thereby improving the MIMO radar's detection performance for weak targets.

[0140] In one embodiment, the angle selection module 19 is further configured to select the angle element that the MIMO radar beam pattern needs to focus on, based on the action and prior information from the previous moment, when the detection time is greater than 0. The aforementioned MIMO radar multi-target detection system 100 based on strong target constraints further includes: a reward calculation module for calculating the reward of the Markov decision process based on the Wald-type statistics of the radar echo; and a function update module for updating the state-value function based on the action.

[0141] In one embodiment, the angle selection module 19 terminates the process when the detection time equals the number of time steps for MIMO radar multi-target detection.

[0142] In one embodiment, the MIMO radar is a co-located MIMO radar.

[0143] For specific limitations regarding the MIMO radar multi-target detection system 100 based on strong target constraints, please refer to the corresponding limitations of the MIMO radar multi-target detection method based on strong target constraints mentioned above, which will not be repeated here. Each module in the aforementioned MIMO radar multi-target detection system 100 based on strong target constraints can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in hardware or independently of a device with data processing capabilities, or stored in software in the memory of the aforementioned device, so that the processor can call and execute the operations corresponding to each module. The aforementioned device can be, but is not limited to, various types of MIMO radar devices already existing in the art.

[0144] In one embodiment, a radar device is provided, including a memory and a processor. The memory stores a computer program, and the processor, when executing the computer program, performs the following processing steps: initializing the current detection time and the beam weight matrix of the MIMO radar; transmitting the radar waveform corresponding to the beam weight matrix and receiving radar echoes from each angle element, and calculating the Wald-type statistics of the radar echoes; calculating the state of the Markov decision process based on the Wald-type statistics of the radar echoes; selecting the action of the Markov decision process based on the state; when the detection time is 0, selecting the angle element that the MIMO radar beam pattern needs to focus on based on the action; solving the beam weight matrix for the next detection time based on the angle elements using a beam weight matrix solution method with strong target constraints; and using the beam weight matrix for the next detection time for MIMO radar multi-target detection at the next detection time.

[0145] In one embodiment, when the processor executes the computer program, it can also implement the steps or sub-steps added in the various embodiments of the MIMO radar multi-target detection method based on strong target constraints.

[0146] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When executed by a processor, the computer program performs the following processing steps: initializing the current detection time and the beam weight matrix of the MIMO radar; transmitting the radar waveform corresponding to the beam weight matrix and receiving the radar echoes from each angle cell, and calculating the Wald-type statistics of the radar echoes; calculating the state of the Markov decision process based on the Wald-type statistics of the radar echoes; selecting the action of the Markov decision process based on the state; when the detection time is 0, selecting the angle cell that the MIMO radar beam pattern needs to focus on based on the action; solving the beam weight matrix for the next detection time based on the angle cells using the beam weight matrix solution method with strong target constraints; the beam weight matrix for the next detection time is used for MIMO radar multi-target detection at the next detection time.

[0147] In one embodiment, when the computer program is executed by the processor, it can also implement the steps or sub-steps added in the various embodiments of the MIMO radar multi-target detection method based on strong target constraints.

[0148] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided by this invention can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), memory bus DRAM (RDRAM), and interface DRAM (DRDRAM), etc.

[0149] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0150] The above embodiments merely illustrate several implementation methods of the present invention, and their descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of protection of the invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and all such modifications and improvements fall within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the appended claims.

Claims

1. A multi-target detection method for MIMO radar based on strong target constraint, characterized in that, Including the following steps: Initialize the current detection time and the beam weight matrix of the MIMO radar; Transmit the radar waveform corresponding to the beam weight matrix and receive the radar echoes of each angle unit, and calculate the Wald-type statistics of the radar echoes. Calculate the state of the Markov decision process based on the Wald-type statistics of the radar echo; Choose the action of the Markov decision process based on the state; When the detection time is 0, the angle element that the MIMO radar beam pattern needs to focus on is selected according to the action; Based on the beam weight matrix solution method with strong target power gain limitation, the beam weight matrix at the next detection time is solved based on the angle element; the beam weight matrix at the next detection time is used for MIMO radar multi-target detection at the next detection time.

2. The MIMO radar multi-target detection method based on strong target constraint according to claim 1, characterized in that, According to the beam weighting matrix solution method with strong target power gain limitation, before the step of solving the beam weighting matrix at the next detection time based on the angle element, the following step is also included: When the detection time is greater than 0, the angle element that the MIMO radar beam pattern needs to focus on is selected based on the action and the prior information of the previous time. Calculate the reward of the Markov decision process based on the Wald-type statistics of the radar echo; Update the state-value function based on the action.

3. The MIMO radar multi-target detection method based on strong target constraint according to claim 1 or 2, characterized in that, According to the method for solving the beam weight matrix with strong target power gain limitation, after the step of solving the beam weight matrix at the next detection time based on the angle element, the following steps are also included: The process ends when the detection time equals the number of time steps for MIMO radar multi-target detection.

4. A MIMO radar multi-target detection system based on strong target constraint, characterized in that, include: The initialization module is used to initialize the current detection time and the beam weight matrix of the MIMO radar. The echo calculation module is used to transmit the radar waveform corresponding to the beam weight matrix and receive the radar echoes of each angle unit, and calculate the Wald-type statistics of the radar echoes. The state calculation module is used to calculate the state of the Markov decision process based on the Wald-type statistics of the radar echo; The action selection module is used to select actions for the Markov decision process based on the state. Angle selection module is used to select the angle unit that the MIMO radar beam pattern needs to focus on when the detection time is 0, based on the action. The weighting solution module is used to solve the beam weight matrix at the next detection time based on the angle element according to the beam weight matrix solution method with strong target power gain limitation; the beam weight matrix at the next detection time is used for MIMO radar multi-target detection at the next detection time.

5. The MIMO radar multi-target detection system based on strong target constraint according to claim 4, characterized in that, The angle selection module is also used to select the angle unit that the MIMO radar beam pattern needs to focus on when the detection time is greater than 0, based on the action and the prior information of the previous time. The MIMO radar multi-target detection system based on strong target constraints also includes: The reward calculation module is used to calculate the reward of the Markov decision process based on the Wald-type statistics of the radar echo. The function update module is used to update the state-value function based on actions.

6. The MIMO radar multi-target detection system based on strong target constraint according to claim 4 or 5, characterized in that, The angle selection module terminates the process when the detection time equals the number of time steps for MIMO radar multi-target detection.

7. The MIMO radar multi-target detection system based on strong target constraint according to claim 6, characterized in that, The MIMO radar is a co-located MIMO radar.

8. A radar device comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the MIMO radar multi-target detection method based on strong target constraints as described in any one of claims 1 to 3.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the MIMO radar multi-target detection method based on strong target constraints as described in any one of claims 1 to 3.

Images