A fast method for multipath range and velocity estimation of millimeter wave radar

Abstract

The present invention relates to a millimeter-wave radar multipath range and velocity fast estimation method, characterized by comprising: 1) obtaining a beat frequency signal; 2) obtaining a multipath component range-velocity spectrum using a fast Fourier transform; 3) preliminarily estimating multipath component parameters using a peak detection algorithm and determining a pre-estimated parameter input set; 4) estimating the multipath component parameters using a high-precision channel parameter estimation algorithm, generating a reconstructed signal, and calculating a residual signal; 5) determining whether parameter estimation has been performed for all paths in the input set; if so, executing step 6); otherwise, switching to the next path and repeating step 4 using the residual signal as the input signal; and 6) determining whether a power difference is less than a threshold; if so, outputting an estimation result; otherwise, using the residual signal as the input of a peak detection algorithm and re-executing steps 3) and 6). Compared with the prior art, the present invention has the advantages of ensuring high-precision estimation while reducing the time complexity of the algorithm.

Description

Technical Field

[0001] The present invention relates to the fields of radar signal processing and environment perception based on vehicle-mounted millimeter-wave radar, and in particular to a method for fast multipath distance and speed estimation of millimeter-wave radar. Background Art

[0002] Millimeter-wave radar technology is a critical component of environmental perception systems for intelligent and autonomous driving. Currently, facing complex environments and challenging intelligent applications, automotive millimeter-wave radar technology faces multiple bottlenecks, including low detection accuracy, low parameter estimation resolution, and difficulty handling multipath effects. This makes it difficult to achieve low-ambiguity positioning of surrounding objects and accurately capture their motion. Therefore, it is necessary to adopt higher-precision channel parameter estimation algorithms to achieve in-depth environmental perception and multipath analysis.

[0003] However, due to the limitations of hardware computing power and application scenarios, high-precision channel parameter estimation algorithms are computationally complex and difficult to apply to the real-time requirements of automotive millimeter-wave radar environments. To address this issue, there is an urgent need to improve current high-precision channel parameter estimation algorithms so that they can adapt to real-time scenarios while maintaining comparable estimation accuracy. Summary of the Invention

[0004] The purpose of the present invention is to provide a millimeter wave radar multipath range and velocity fast estimation method in order to overcome the defects of the above-mentioned prior art.

[0005] The purpose of the present invention can be achieved by the following technical solutions:

[0006] A millimeter wave radar multipath range and velocity fast estimation method comprises the following steps:

[0007] Step 1) obtaining a beat frequency signal of a millimeter wave radar;

[0008] Step 2) using the beat frequency signal as input, a two-dimensional fast Fourier transform is used to obtain the multipath component range-velocity spectrum;

[0009] Step 3) Preliminarily estimating multipath component parameters using a peak detection algorithm, selecting some or all of the multipath components in the output of the peak detection algorithm as a pre-estimated parameter input set based on the pre-configured number of paths;

[0010] Step 4) Using a path from the pre-estimated parameter input set as input, a high-precision channel parameter estimation algorithm is used to estimate the multipath component parameters. When the algorithm converges, the estimated multipath component distance and velocity parameters are used to generate a reconstructed signal for the current path. The reconstructed signal is subtracted from the original signal to obtain a residual signal.

[0011] Step 5) determining whether parameter estimation has been performed on all paths in the pre-estimated parameter input set. If so, proceed to step 6). Otherwise, switch to the next path and repeat step 4) using the residual signal as the input signal until all paths have been traversed.

[0012] Step 6) determines whether the difference between the residual signal range-velocity spectrum power level and the original signal power level is less than a preconfigured threshold. If so, the multipath component parameter estimation is completed and the estimation result is output. Otherwise, the residual signal is used as the input of the peak detection algorithm and steps 3) to 6) are re-executed until the parameter estimation is completed.

[0013] The millimeter-wave radar is a linear frequency modulation radar. The transmitted signal is received by the receiving antenna after being transmitted through the channel. The echo signal undergoes mixing, anti-aliasing filtering and other steps to obtain a beat frequency signal. The beat frequency signal includes fast time and slow time. The fast time corresponds to distance dimension information, and the slow time corresponds to speed dimension information.

[0014] The multipath component expression of the beat signal at fast and slow times is:

[0015]

[0016] Among them, t s and t f They represent slow time and fast time respectively, L represents the total number of paths, l represents the current corresponding path index, α l is the complex amplitude of the beat signal of the lth path, f R,l It represents the frequency change caused by the time delay during signal propagation, records the multipath distance information, and is proportional to the propagation distance. V,l It represents the frequency change caused by Doppler during signal propagation, records the multipath velocity information, and is proportional to the propagation speed.

[0017] The multipath component parameters include time delay, Doppler and power, wherein time delay corresponds to distance dimension information and Doppler corresponds to speed dimension information.

[0018] In the step 2), the beat frequency signal is subjected to two-dimensional fast Fourier transforms to obtain a range-velocity spectrum.

[0019] In step 3), the peak detection results are sorted from large to small according to the power, and the pre-estimation parameter input set is obtained as Θ = [θ1, θ2, ..., θ L ], where θ l =[r l ,v l ].

[0020] The peak detection algorithm is a two-dimensional constant false alarm detection algorithm.

[0021] The high-precision channel parameter estimation algorithm is based on a spatial iterative generalized expectation maximization algorithm.

[0022] The high-precision channel parameter estimation algorithm first estimates the distance parameter and then estimates the speed parameter.

[0023] In step 4), the signal complex amplitude is obtained according to the multipath component parameter estimation result, and then the reconstructed signal of the current path is obtained according to the reconstructed signal expression. The signal complex amplitude is:

[0024]

[0025] in, is the original signal, represents the current path estimation signal steering vector, H is the conjugate transpose symbol, t s and t f They represent slow time and fast time respectively, l represents the current corresponding path index, k is the frequency modulation slope, f c is the carrier frequency, c is the speed of light, is the current path parameter estimation set, where r est With v est They are the estimated distance and speed results of the current path respectively;

[0026] The current path reconstruction signal expression is:

[0027]

[0028] in, Estimate the complex amplitude for the current path, Estimate the signal steering vector for the current path.

[0029] Compared with the prior art, the present invention has the following beneficial effects:

[0030] (1) Based on the prior information of the range-velocity spectrum, the present invention performs high-precision channel parameter estimation on the basis of peak detection, narrows the algorithm parameter search range, and reduces the time complexity of the algorithm.

[0031] (2) The present invention ensures rapid convergence of iteration by setting the step size coefficient, ensuring rapid and accurate estimation of parameters.

[0032] (3) The present invention sets the parameter input set and the residual signal detection threshold, and can autonomously adjust the number of paths of interest and the tolerable noise level according to different scenarios, thereby improving the robustness and environmental adaptability of the algorithm.

[0033] (4) The present invention utilizes the residual signal for re-estimation, which not only solves the problems of missed detection and dependence on parameter settings in traditional peak detection, but also ensures that the method can converge quickly under limited path estimation. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 is the original signal distance-velocity spectrum.

[0035] Figure 2 To achieve peak detection results using a two-dimensional constant false alarm detection algorithm.

[0036] Figure 3 The normalized root mean square error of the distance estimation under different numbers of iterations.

[0037] Figure 4 The normalized root mean square estimation error results of the velocity at different numbers of iterations.

[0038] Figure 5 The signal result is reconstructed for high-precision channel parameter estimation after single detection.

[0039] Figure 6 Calculation result of the residual signal after a single detection.

[0040] Figure 7 The signal is reconstructed for high-precision channel parameter estimation after re-detection.

[0041] Figure 8 is the result of residual signal calculation after re-detection.

[0042] Figure 9 Schematic diagram of the method of the present invention. DETAILED DESCRIPTION

[0043] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to the following embodiments.

[0044] A millimeter wave radar multipath range and velocity fast estimation method, such as Figure 9 ,include:

[0045] Step 1) obtaining a beat frequency signal of a millimeter wave radar;

[0046] Step 2) using the beat frequency signal as input, a two-dimensional fast Fourier transform is used to obtain the multipath component range-velocity spectrum;

[0047] Step 3) Preliminarily estimating multipath component parameters using a peak detection algorithm, selecting some or all of the multipath components in the output of the peak detection algorithm as a pre-estimated parameter input set based on the pre-configured number of paths;

[0048] Step 4) Using a path from the pre-estimated parameter input set as input, a high-precision channel parameter estimation algorithm is used to estimate the multipath component parameters. When the algorithm converges, the estimated multipath component distance and velocity parameters are used to generate a reconstructed signal for the current path. The reconstructed signal is subtracted from the original signal to obtain a residual signal.

[0049] Step 5) determining whether parameter estimation has been performed on all paths in the pre-estimated parameter input set. If so, proceed to step 6). Otherwise, switch to the next path and repeat step 4) using the residual signal as the input signal until all paths have been traversed.

[0050] Step 6) determines whether the difference between the residual signal range-velocity spectrum power level and the original signal power level is less than a preconfigured threshold. If so, the multipath component parameter estimation is completed and the estimation result is output. Otherwise, the residual signal is used as the input of the peak detection algorithm and steps 3) to 6) are re-executed until the parameter estimation is completed.

[0051] Step 1) includes:

[0052] Taking the most widely used linear frequency modulation radar (LFM) as an example, the transmitted signal is received by the receiving antenna after passing through the channel. The echo signal undergoes mixing and anti-aliasing filtering to generate a beat signal. The beat signal consists of fast and slow times, representing information in the range and velocity dimensions, respectively.

[0053] The multipath component expression of the beat frequency signal at fast and slow times is:

[0054]

[0055] Among them, t s and t f They represent slow time and fast time respectively, L represents the total number of paths, l represents the current corresponding path index, α l is the complex amplitude of the beat signal of the lth path. R,l It indicates the frequency change caused by the time delay during signal propagation, records the multipath distance information, and is proportional to the propagation distance. V,l It represents the frequency change caused by Doppler during signal propagation, records the multipath velocity information, and is proportional to the propagation speed.

[0056] Specifically, f R,l With f V,l The expression is

[0057]

[0058]

[0059] where f cis the carrier frequency, k is the frequency modulation slope, c is the speed of light, r l With v l are the distance and speed corresponding to the beat frequency signal of the lth path respectively. Substituting equations (2) and (3) into equation (1), the beat frequency signal expression can be obtained:

[0060]

[0061] Step 2) includes taking the beat frequency signal as input and performing two fast Fourier transforms to obtain the range-velocity spectrum of the multipath component. V, Often smaller than f R,l In the distance-velocity spectrum, the fast-time distance-velocity coupling term is often directly regarded as the distance result.

[0062] During the specific implementation, a linear frequency modulation radar test platform was built based on a laboratory environment. The measurement location was a wide outdoor road. The radar was fixed by a bracket at a height of 1m from the ground. There was a pedestrian target facing the radar in the scene. It moved from near to far, from a position 4m away from the radar to a position 16m away. The moving speed was about a uniform speed of 1.5m / s. The total measurement time was about 8s. In addition, there were no other moving targets. In the parameter configuration, the measurement data block has a total of 64 frames, each frame contains M, which is a chirp sequence of 128, and each sequence has 512 fast time sampling points N. The radar transmit carrier frequency f c is 77.6GHz, the bandwidth BW is 400MHz, and the sampling rate f s is 10MHz, the pulse repetition interval T cp The frame repetition interval is 80μs and 128ms.

[0063] Distance resolution refers to the minimum distance between two adjacent points that can be distinguished in the fast Fourier transform. Its calculation formula is:

[0064]

[0065] Velocity resolution refers to the minimum velocity that can be distinguished between two adjacent points in the fast Fourier transform. Its calculation formula is:

[0066]

[0067] Using equations (5) and (6), we can calculate the range resolution and velocity resolution in this example to be 0.375m and 0.189m / s respectively. The range-velocity spectrum obtained by measurement is as follows: Figure 1 shown.

[0068] Step 3) includes:

[0069] A peak detection algorithm is used to preliminarily estimate the multipath components. Here, the peak detection algorithm is a two-dimensional constant false alarm detection algorithm. The protection cell refers to the cell in close proximity to the cell under test and is not included in the threshold calculation to prevent the majority of the actual signal from leaking into the reference cell. The reference cell refers to the data portion involved in the threshold calculation and is used to estimate the threshold value of the cell under test. The threshold value calculation uses the cell averaging algorithm as an example. The threshold is determined by estimating the mean value of the cells surrounding each point. The level of the cell under test is compared with the threshold value. If the level of the cell under test is greater than the threshold value, a target is determined to be present.

[0070] For the threshold value T(i, j) of the unit point to be tested in the i-th row and j-th column, it can be expressed as the average value of the reference unit train(i, j) corresponding to the point multiplied by the compensation coefficient γ, where the size of γ is mainly affected by the environmental signal-to-noise ratio.

[0071]

[0072] In this example, the constant false alarm detection algorithm parameter configuration sets the distance and speed protection units to 4 and 8, respectively, the distance and speed reference units to 2 and 4, and the compensation coefficient γ to 12dB. The calculation of the reference unit mean can be accelerated using a two-dimensional prefix sum algorithm. The key idea of ​​the two-dimensional prefix sum algorithm is to first calculate the two-dimensional prefix sum corresponding to all points in the matrix, then accumulate the units within the range block by adding and subtracting the prefix sums between different endpoints.

[0073] The peak detection results in multiple distance-velocity units, such as Figure 2 As shown. It is used as the pre-estimation result and as the parameter input set for the subsequent high-precision channel parameter estimation algorithm. To ensure the performance of the subsequent algorithm, it is necessary to sort the detection results from large to small according to the power size, and finally obtain the parameter input set Θ = [θ1,θ2,...,θ L ]. Where θ l =[r l ,v l ].

[0074] In this example, after peak detection is completed, the result is a parameter input set sorted from largest to smallest, with a total of 164 paths. Here, the number of paths estimated by the subsequent algorithm is set to 150.

[0075] Step 4) includes:

[0076] After obtaining the parameter input set Θ, the space-based iterative generalized expectation maximization algorithm is used to accurately estimate the parameters. The parameter input set Θ is traversed using the path as the iterative variable, and the parameter value r of the lth item is traversed each time. l ,v l As the pre-input of the beat signal model r pre,v pre . A high-precision channel parameter estimation algorithm is used to accurately estimate the distance and speed of each path.

[0077] First, perform distance estimation and select the distance estimation range as r est ∈[r pre -δr,r pre +δr], where δr = Δr / 2, and the number of estimated points is set to 20. Δr is the range resolution. In this case, different range cells do not overlap, and analysis of one cell does not affect other areas.

[0078] At this time, according to formula (4), the current estimated path can be expressed as

[0079]

[0080] In the expectation step, a two-dimensional expression of the distance parameter in the fast and slow time dimensions is obtained. By calculating the conditional expectation of the log-likelihood function and solving the maximum value in the maximization step, r est Approximate the exact value. In order to reduce the computational complexity, some slow time series can be used to estimate the distance parameters during the conditional expectation calculation process.

[0081] Then the speed is estimated and the distance parameter r is updated. est Then, the speed estimation range is selected as v est ∈[v pre -δv,v pre +δv], δv = Δv / 2, and the number of estimation points is set to 20. Δv is the velocity resolution. In this case, different velocity units do not overlap, and the analysis of one unit will not affect other areas.

[0082] At this time, according to formula (8), the current estimated path can be expressed as

[0083]

[0084] In the same way, we can get an accurate estimate of the velocity v est In order to reduce the computational complexity, some fast time series can be used to estimate the velocity parameters during the conditional expectation calculation process.

[0085] Since there is coupling between distance and speed, after one iteration, the estimated distance and speed are estimated once. est ,v est As the input r of the new round of iteration pre ,v pre Continue estimating until r est ,v estUntil convergence is achieved and no changes are made. In order to further improve the estimation accuracy, the step size is updated after each iteration. Let the distance step size coefficient and the speed step size coefficient be α, β (α>1, β>1), respectively. Then the parameter estimation range of the new round of iteration is updated to r est ∈[r pre -δr / α,r pre +δr / α],v est ∈[v pre -δv / β,v pre +δv / β].

[0086] In the simulation test, we assume that a path is estimated and 10 iterations are performed. The number of Monte Carlo simulations is set to 200 for each iteration. The distance step coefficient α is set to 10, the speed step coefficient β is set to 5, the number of estimation points is set to 20, and the signal-to-noise ratio is set to 20dB. The distance and speed estimation results are as follows: Figure 3 and 4 As shown in the figure, the result of iteration 0 is the result of the fast Fourier transform pre-estimation parameters. As can be seen from the figure, when the number of iterations is 2, the distance and velocity estimation can basically converge. At this time, the normalized RMS error of the distance estimation is approximately -44dB, which is a 23dB reduction compared to the pre-estimation error of -21dB. The normalized RMS error of the velocity estimation is approximately -40dB, which is a 16dB reduction compared to the pre-estimation error of -24dB. This shows that this algorithm can achieve better resolution than the traditional Fourier transform.

[0087] After the distance-velocity estimation of a single path reaches convergence, the obtained parameter estimation results can be used to estimate the complex amplitude of the signal. The specific expression is:

[0088]

[0089] in

[0090]

[0091] is the original signal. H is the conjugate transpose symbol, is the current path parameter estimation set, where r est With v est are the estimated distance and speed results of the current path respectively. According to formula (12), the reconstructed signal of the current path can be calculated as

[0092]

[0093] in, Estimate the complex amplitude for the current path, Estimate the signal steering vector for the current path.

[0094] The reconstructed signal is subtracted from the original signal to obtain the residual signal.

[0095] Step 5) includes determining whether all paths in the pre-estimated parameter input set have been parameter estimated. If so, proceed to step 6). Otherwise, switch to the next path and repeat step 4) with the residual signal as the input signal until all paths in the parameter input set have been traversed. 150 estimated paths are selected, the number of iterations is 2, and the final reconstructed signal result is as follows: Figure 5 shown.

[0096] Step 6) includes:

[0097] After traversing the parameter input set, if the difference between the residual signal's range-velocity spectrum power level and the original signal's power level exceeds a preconfigured threshold, it indicates that some multipath paths are still incompletely estimated. Peak detection is then performed on the residual signal for the next round of estimation. Otherwise, all paths are considered completely estimated. The threshold is an empirical parameter, primarily influenced by the current signal-to-noise ratio (SNR) and the peak detection compensation factor.

[0098] After a single peak detection, Figure 6 The residual signal obtained by subtracting the original signal from the reconstructed signal is shown. The peak power of the residual signal is reduced to -50 dB compared to the peak power of the original signal, almost reaching the noise floor. This indicates that the high-precision channel parameter estimation algorithm has essentially achieved signal reconstruction. However, some paths are still not fully estimated. This is because some targets were missed during the initial detection, resulting in unsuccessful signal reconstruction. Therefore, the residual signal needs to be re-detected and a secondary estimation performed on the missed targets.

[0099] After the residual signal is subjected to peak detection and high-precision channel parameter estimation algorithm estimation, the re-detection estimation result is obtained as follows: Figure 7 As shown, the residual signal calculation results are as follows Figure 8 As shown. Figure 5 and Figure 6 It can be found that the missed detection path has been effectively re-estimated, and the residual signal power has been suppressed to the noise level, which can be regarded as the completion of reconstruction.

[0100] When all estimation processes are completed, the algorithm outputs the reconstructed signal, the residual signal, and the distance, speed, and complex amplitude results of all estimated paths.

[0101] Estimation accuracy analysis. Compared with the traditional range-velocity spectrum peak detection, this method can improve the estimation accuracy by about 20dB. Figure 3 and Figure 4 Therefore, the problems of low fast Fourier transform estimation accuracy, spectrum leakage, and obvious sidelobe effect are overcome.

[0102] Estimation accuracy and residual rate analysis. Because peak detection inevitably suffers from missed detections and false detections, and its parameters require constant adjustment for different environments, traditional range-velocity spectrum analysis cannot achieve accurate and reliable target detection and positioning, with significant missed and false alarms. This method uses residual signal re-detection and re-estimation. Even if there are missed detections in the first estimation, subsequent detections can further estimate the remaining path, ensuring accurate and effective detection, ultimately bringing the residual signal power level close to the noise level.

[0103] Time complexity analysis. For traditional high-precision channel parameter estimation algorithms, taking the estimation of a single distance parameter as an example, assuming a single frame contains 128 chirped pulses, and each chirp estimates 50 paths, a total of 128 × 50 = 6,400 paths must be estimated per frame. This includes a large number of paths that are repeatedly estimated, such as for stationary targets, resulting in a large number of repeated estimation processes. Each path distance parameter requires a full search. If two rounds of estimation are performed, with 500 estimation points in the first round and 100 points in the second, the number of single estimation executions is 6,400 × 600 = 3,840,000. This number of executions increases significantly if multiple parameters are jointly estimated. This method, based on prior information about the range-velocity spectrum, determines the number of paths based on peak detection results. For example, in this example, 300 paths are used. The estimation range for a single parameter is limited to the peak detection results, significantly reducing the number of estimation points. For example, in this example, 20 single parameter estimation points are used, the number of iterations is 2, and both range and velocity parameters are estimated, resulting in a total of 80 estimation points. The final estimated execution times are 300×80=24000, which greatly reduces the complexity compared with the traditional channel parameter estimation algorithm. The peak detection algorithm can use the two-dimensional prefix sum algorithm to calculate the peak value from o(n 2 ) complexity is reduced to o(1), which has little impact on the overall algorithm time complexity.

[0104] The preferred specific embodiments of the present invention are described in detail above. In addition, it should be noted that the specific embodiments described in this specification may be named differently, and the above content described in this specification is merely an example of the structure of the present invention. Any equivalent changes or simple changes made based on the structure, features and principles of the present invention are included in the protection scope of the present invention. Those skilled in the art of the present invention may make various modifications or additions to the described specific examples or adopt similar methods, as long as they do not deviate from the structure of the present invention or exceed the scope defined by the claims, they should all fall within the protection scope of the present invention.

Claims

1. A millimeter wave radar multipath range and velocity fast estimation method, characterized in that: The following steps are involved: Step 1) obtaining a beat frequency signal of a millimeter wave radar; Step 2) using the beat frequency signal as input, a two-dimensional fast Fourier transform is used to obtain the multipath component range-velocity spectrum; Step 3) Preliminarily estimating multipath component parameters using a peak detection algorithm, selecting some or all of the multipath components in the output of the peak detection algorithm as a pre-estimated parameter input set based on the pre-configured number of paths; Step 4) Using a path from the pre-estimated parameter input set as input, a high-precision channel parameter estimation algorithm is used to estimate the multipath component parameters. When the algorithm converges, the estimated multipath component distance and velocity parameters are used to generate a reconstructed signal for the current path. The reconstructed signal is subtracted from the original signal to obtain a residual signal. Step 5) determining whether parameter estimation has been performed on all paths in the pre-estimated parameter input set. If so, proceed to step 6). Otherwise, switch to the next path and repeat step 4) using the residual signal as the input signal until all paths have been traversed. Step 6) Determine whether the difference between the residual signal range-velocity spectrum power level and the original signal power level is less than a preconfigured threshold. If so, multipath component parameter estimation is complete and the estimation result is output. Otherwise, use the residual signal as the input of the peak detection algorithm and re-execute steps 3) to 6) until parameter estimation is complete. In step 4), the complex amplitude of the signal is obtained according to the multipath component parameter estimation result, and then the reconstructed signal of the current path is obtained according to the reconstructed signal expression. The complex amplitude of the signal estimated for the current path is: in, is the original signal, represents the current path estimation signal steering vector, H is the conjugate transpose symbol, t s and t f They represent slow time and fast time respectively, l represents the current corresponding path index, k is the frequency modulation slope, f c is the carrier frequency, c is the speed of light, is the current path parameter estimation set, where r est With v est They are the estimated distance and speed results of the current path respectively; The current path reconstruction signal expression is: in, Estimate the complex amplitude for the current path, Estimate the signal steering vector for the current path.

2. The method for rapid estimation of multipath range and velocity of millimeter wave radar according to claim 1, characterized in that: The millimeter-wave radar is a linear frequency modulation radar. The transmitted signal is received by the receiving antenna after being transmitted through the channel. The echo signal undergoes mixing and anti-aliasing filtering to obtain a beat frequency signal. The beat frequency signal includes fast time and slow time. The fast time corresponds to distance dimension information, and the slow time corresponds to speed dimension information.

3. The method for rapid estimation of multipath range and velocity of millimeter wave radar according to claim 2, characterized in that: The multipath component expression of the beat signal at fast and slow times is: Among them, t s and t f They represent slow time and fast time respectively, L represents the total number of paths, l represents the current corresponding path index, α l is the complex amplitude of the beat signal of the lth path, f R,l It represents the frequency change caused by the time delay during signal propagation, records the multipath distance information, and is proportional to the propagation distance. V,l It represents the frequency change caused by Doppler during signal propagation, records the multipath velocity information, and is proportional to the propagation speed.

4. The method for rapid estimation of millimeter wave radar multipath range and velocity according to claim 1, wherein: The multipath component parameters include time delay, Doppler and power, wherein time delay corresponds to distance dimension information and Doppler corresponds to speed dimension information.

5. The millimeter wave radar multipath range and velocity fast estimation method according to claim 1, characterized in that: In the step 2), the beat frequency signal is subjected to two-dimensional fast Fourier transforms to obtain a range-velocity spectrum.

6. The millimeter wave radar multipath range and velocity fast estimation method according to claim 1, characterized in that: In step 3), the peak detection results are sorted from large to small according to the power, and the pre-estimation parameter input set is obtained as Θ = [θ1, θ2, ..., θ L ], where θ l =[r l ,v l ].

7. The method for rapid estimation of millimeter wave radar multipath range and velocity according to claim 1, characterized in that: The peak detection algorithm is a two-dimensional constant false alarm detection algorithm.

8. The millimeter wave radar multipath range and velocity fast estimation method according to claim 1, characterized in that: The high-precision channel parameter estimation algorithm is based on a spatial iterative generalized expectation maximization algorithm.

9. The method for rapid estimation of millimeter wave radar multipath range and velocity according to claim 8, characterized in that: The high-precision channel parameter estimation algorithm first estimates the distance parameter and then estimates the speed parameter.

Images