A method for suppressing multi-path targets of millimeter wave radar in a tunnel environment

Abstract

The application belongs to the technical field of millimeter wave radar target detection and signal processing, and specifically discloses a millimeter wave radar multipath target suppression method in a tunnel environment. The method establishes a geometric position correlation relationship between a real target and each order of multipath targets in a range-Doppler diagram by using millimeter wave radar echo signals in the tunnel. On this basis, a candidate signal point set meeting a geometric constraint condition is subjected to coherence analysis and energy consistency verification in sequence, so that the real target echo and the multipath interference signal are effectively distinguished, and the discrimination accuracy of the multipath signal is significantly improved. At the same time, after distinguishing the real target and the multipath signal, the spatial position and amplitude information of the identified multipath signal point are used to construct a multipath signal space distribution estimation diagram. According to the diagram and in accordance with a reduction ratio, the multipath signal component is removed from the original range-Doppler diagram, so that the multipath interference is directly suppressed in the signal processing stage.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of millimeter wave radar target detection and signal processing, and specifically discloses a millimeter wave radar multipath target suppression method in a tunnel environment. BACKGROUND

[0002] Accurate positioning of personnel during tunnel construction can provide timely warning when personnel enter dangerous areas, thereby ensuring personnel safety. However, the tunnel construction environment is narrow, dark, and has a large amount of dust, which poses great challenges to personnel positioning.

[0003] Millimeter wave radar has strong anti-dust interference and high ranging accuracy, and thus becomes an optimal sensor for tunnel personnel positioning. However, the electromagnetic waves emitted by the millimeter wave radar are reflected by the tunnel wall to form multipath false targets, which seriously interfere with target detection, and thus it is necessary to suppress multipath targets.

[0004] To solve this problem, existing technologies have provided solutions, such as the millimeter wave radar tunnel scene multipath ghost suppression method provided in Chinese Patent Publication No. CN120103276B. This solution processes in the radar point cloud domain, calculates the slope of the median line of the dynamic target point pair, and matches it with the slope of the reflection surface to identify the second-order multipath “ghost” belonging to the mirror relationship, and supplements the confidence evaluation of multidimensional features such as signal-to-noise ratio and speed constraints to screen and suppress the preliminary identification results.

[0005] The above solution suppresses multipath interference in the rear end of the radar signal processing link, and its effectiveness is highly dependent on the quality of the early point cloud clustering and the accurate fitting of the tunnel wall surface. Once the clustering is wrong or the wall surface modeling is biased, the subsequent mirror discrimination will lose reliable basis, and the suppression performance will be significantly reduced.

[0006] Further, the above solution intervenes at the end of the radar signal processing link, i.e., after the point cloud is formed. At this time, the multipath energy has been deeply integrated with the real echo, which not only increases the computational overhead, but also makes it difficult to accurately separate the false trajectories that have been formed, and it is easy to mistakenly delete adjacent real targets.

[0007] In addition, the above solution for identifying multipath targets only relies on the spatial mirror symmetry as the core geometric correlation feature, which is easy to misjudge the independent targets with occasional similar positions as multipath pairs in complex tunnel scenes, resulting in high false and missed detection rates. SUMMARY

[0008] To solve the above technical problems or at least partially solve the above technical problems, the present application provides a millimeter wave radar multipath target suppression method in a tunnel environment.

[0009] The purpose of the present application can be realized by the following technical solutions: a tunnel environment millimeter wave radar multi-path target suppression method, comprising: signal processing of millimeter wave radar echo signals in the tunnel to generate a range-Doppler map.

[0010] Based on the tunnel geometry, the geometric position correlation between the real target and each order of multi-path target in the range-Doppler map is established.

[0011] According to the geometric position correlation, all candidate signal point sets that meet the geometric constraint conditions are identified in the range-Doppler map, and the candidate target signal groups belonging to the same scattering source are screened out through coherence analysis.

[0012] For the candidate target signal group, energy consistency verification is performed using the energy attenuation relationship between the real target and the multi-path target to mark the real target signal point and the multi-path signal point.

[0013] Based on the spatial position and amplitude of the marked multi-path signal point, a multi-path signal spatial distribution estimation map is generated.

[0014] According to the multi-path signal spatial distribution estimation map, the corresponding signal components are subtracted from the original range-Doppler map according to the subtraction ratio.

[0015] The enhanced range-Doppler map after subtraction processing is subjected to target detection, and the real target point set is output.

[0016] The above-mentioned all technical solutions are combined, and the present application has the following positive effects: 1. The present application establishes the geometric position correlation between the real target and each order of multi-path target in the range-Doppler map by using the millimeter wave radar echo signals in the tunnel, and on this basis, the coherence analysis and energy consistency verification are successively implemented for the candidate signal point set that meets the geometric constraint conditions, thereby effectively distinguishing the real target echo and the multi-path interference signal, realizing high confidence recognition of the real target in the complex tunnel scene, significantly improving the discrimination accuracy of the multi-path signal, and greatly reducing the false positive rate and the missed detection rate in target detection.

[0017] 2. After distinguishing the real target and the multi-path signal, the present application uses the spatial position and amplitude information of the identified multi-path signal point to construct a multi-path signal spatial distribution estimation map, and according to the map and according to the subtraction ratio, the multi-path signal components are removed from the original range-Doppler map, so as to directly suppress the multi-path interference in the signal processing stage. Since it does not depend on the previous target point cloud data, the accumulation of early data errors is reduced, the real-time processing capability is improved, and the deletion of adjacent real targets is largely avoided. BRIEF DESCRIPTION OF DRAWINGS

[0018] The application is further described with the accompanying drawings, but the embodiments in the drawings do not constitute any limitation to the application, and other embodiments can be obtained by those of ordinary skill in the art without creative labor on the basis of the following drawings.

[0019] Figure 1 The method of the application is implemented by the step diagram.

[0020] Figure 2 The implementation flowchart of the candidate target signal group belonging to the same scattering source screened out by the coherence analysis in the application.

[0021] Figure 3 The implementation flowchart of the energy consistency verification in the application. DETAILED DESCRIPTION

[0022] The technical solutions in the embodiments of the application will be described clearly and completely below with the accompanying drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all other embodiments obtained by those of ordinary skill in the art without creative labor are within the protection scope of the application.

[0023] Referring to Figure 1 As shown in the drawings, the application proposes a millimeter wave radar multipath target suppression method in a tunnel environment, including the following steps: S1, generating a range-Doppler map by signal processing on the millimeter wave radar echo signal in the tunnel.

[0024] In a long and narrow closed environment of a tunnel, electromagnetic waves emitted by a radar will be reflected by structural surfaces such as walls, ground and vaults during propagation, forming a multipath propagation effect, and then producing a false target response in radar detection. In order to suppress the false targets produced due to the multipath effect, it is necessary to distinguish the real target echo from the multipath false echo.

[0025] The distinguishing operation needs to consider the radial distance of the target relative to the radar and the radial velocity of the target relative to the radar. The range-Doppler map can present the radial distance of the target relative to the radar and the radial velocity of the target relative to the radar at the same time as a two-dimensional signal representation. Therefore, the application uses the range-Doppler map as an analysis tool to represent the target information in the echo signal, thereby providing a basis for subsequent multipath target identification and suppression.

[0026] The distance-Doppler map is a signal representation in the form of a two-dimensional matrix, the horizontal axis of which corresponds to the radial distance of the target from the radar, divided into a plurality of consecutive distance units, and the vertical axis of which corresponds to the radial velocity of the target relative to the radar, divided into a plurality of consecutive Doppler units. The coordinates of each pixel point (i.e. distance-Doppler unit) in the map uniquely determine the corresponding distance value and radial velocity value, and the numerical amplitude of the pixel point represents the strength of the echo signal received by the radar at the specific distance and velocity.

[0027] The distance-Doppler map not only clearly shows the spatial position and motion state of the real target, but also visually presents the false responses generated by multipath propagation in the form of discrete signal points. Therefore, by analyzing the distribution, amplitude and mutual relationship of the signal points in the distance-Doppler map, a basis for distinguishing real targets from multipath interference can be provided.

[0028] In specific embodiments, the process of generating a distance-Doppler map based on millimeter wave radar echo signals is as follows: first, signal transmission and reception are performed. Specifically, the millimeter wave radar transmits a high-frequency electromagnetic wave signal. The signal is reflected after encountering a target (including a real target and an environmental structure) in the propagation path, forming an echo signal and being received by the radar antenna. The echo signal encodes the distance and radial velocity information of the target.

[0029] Then, the received echo signal is mixed with the original signal transmitted by the radar to generate an intermediate frequency signal containing target information. In this process, if the target has radial motion relative to the radar, the frequency of the echo signal will shift based on the Doppler effect, and the shift is proportional to the radial velocity.

[0030] Then, the intermediate frequency signal obtained after mixing is sampled and analog-to-digital converted to convert it into a discrete digital signal for subsequent digital signal processing.

[0031] Secondly, Fourier transform processing is performed, and the digitized signal is subjected to two Fourier transforms. The first Fourier transform is performed along the fast time dimension, converting the signal from the time domain to the distance frequency domain, thereby separating the signal components of different time delays (corresponding to different distances) and obtaining the distance information of the target. The second Fourier transform is performed along the slow time dimension (i.e. within each distance unit), converting the signal from the time domain to the Doppler frequency domain, thereby extracting the radial velocity information of the target within each distance unit.

[0032] Finally, the data obtained after the above two Fourier transform processes are subjected to amplitude calculation and arrangement to form the distance-Doppler map. In the map, the horizontal and vertical coordinates of each unit (i.e. pixel point) correspond to a determined distance index and velocity index, respectively, and the amplitude value of the unit represents the signal energy at the unit.

[0033] The distance-Doppler graph generated by the above process constitutes the data basis for subsequent multi-path target geometric correlation analysis, coherence discrimination, energy verification and interference suppression processing. In the present application, the coordinate unit or unit refers to the basic resolution unit in the distance-Doppler graph.

[0034] S2, based on the tunnel geometry, the geometric position correlation between the real target and each order multi-path target in the distance-Doppler graph is established.

[0035] In a tunnel environment, the multi-path echoes formed by the reflection of electromagnetic waves on the surface of walls, ground or vaults, etc. will appear as signal points similar to the characteristics of real targets in the distance-Doppler graph. It is difficult to effectively distinguish them only according to the amplitude information of each signal point. Therefore, the distance-Doppler graph itself cannot directly identify the physical source of the signal points.

[0036] To solve the above problems, the present application uses the prior knowledge of tunnel geometry. Under the premise of knowing the tunnel width, height, curvature and other geometric structures, there is a definite geometric relationship between the real target and the multi-path virtual image formed by its single or multiple mirror reflection. This spatial geometric relationship can be converted into the correlation of its coordinate position in the distance-Doppler graph. By establishing this geometric position correlation, the tunnel structure information can be converted into a clue for identifying multi-path interference in the Doppler domain, so as to predict the area where the multi-path signal may appear, and facilitate accurate distinction between real targets and multi-path false targets.

[0037] The specific process of establishing the geometric position correlation will be described below in conjunction with the preferred embodiments: S201, define the spatial coordinate system: take the phase center of the millimeter wave radar as the coordinate origin to establish a three-dimensional space rectangular coordinate system; and define the main direction of the radar field of view as the positive direction of the Z axis, the horizontal transverse direction as the X axis, and the vertical direction as the Y axis.

[0038] S202, establish a tunnel wall geometry model: based on the tunnel design drawing, in the three-dimensional space rectangular coordinate system, use mathematical equations to model the main reflecting surfaces in the tunnel. The reflecting surfaces at least include the two side walls, the vault and the ground.

[0039] For example, for a rectangular cross-section tunnel, the right side wall can be modeled as a plane equation, the equation is The left side wall equation is wherein represents the tunnel width, the ground equation is y=0, and if the vault is a flat top, the equation is wherein H represents the tunnel height.

[0040] By wall geometry modeling, according to the mirror reflection principle, the real target is mirrored about each reflecting wall, thereby constructing a corresponding first-order, second-order, and multi-order virtual mirror source. Based on this, the multipath signal propagation path originally via wall reflection to the radar can be equivalently regarded as a straight line path directly propagated to the radar by the virtual mirror source. The analysis of the complex reflection path is converted into the geometric calculation of the straight line path between the virtual point source and the radar, thereby simplifying the theoretical derivation and modeling of the influence of the multipath signal in the range and velocity dimensions.

[0041] S203, define the type of multipath propagation path: define the path of electromagnetic wave directly propagating from the radar to the real target as a direct path; define the path of electromagnetic wave reaching the real target after single reflection of any wall of the tunnel as a first-order multipath path; define the path of electromagnetic wave reaching the real target after two wall reflections as a second-order multipath path; higher-order multipath paths can be similarly defined.

[0042] S204, calculate path parameters and establish correlation based on ray tracing, based on the ray tracing principle in geometric optics, the equivalent propagation parameters of the direct path, each first-order and second-order multipath path corresponding to the real target are calculated, mainly including propagation distance and equivalent radial velocity.

[0043] Among them, the propagation distance refers to the equivalent total length of the radar signal propagating along the path. For the direct path, its propagation distance is the Euclidean distance between the radar and the real target. For the first-order or second-order multipath path, its propagation distance can be obtained by calculating the Euclidean distance between the radar and the corresponding virtual image source.

[0044] The equivalent radial velocity refers to the velocity component corresponding to the Doppler shift caused by the target motion in the direction of the specific propagation path. For the direct path, its radial velocity is the projection of the real target motion speed in the radar line-of-sight direction (Z-axis). For the multipath path, the equivalent radial velocity needs to project the real target's motion speed vector onto the equivalent propagation direction of the multipath path (i.e. the direction of the radar pointing to the corresponding virtual image source).

[0045] S205, parameter mapping to range-Doppler map: for each path, after obtaining its propagation distance and equivalent radial velocity, according to the resolution and range adopted when generating the range-Doppler map by the millimeter wave radar signal processing, through linear scaling transformation and rounding operation, its corresponding row index and column index in the range-Doppler map are determined. Thus, the theoretical position correlation set of the real target point and each order of multipath target point generated by it in the range-Doppler map is established.

[0046] It should be noted that the signal amplitude of each cell in the range-Doppler map is essentially the coherent accumulation of the energy of the echo signals of all propagation path lengths falling within the distance interval and radial velocity falling within the velocity interval. Therefore, for a target with a certain position and velocity, the echo generated by any one reflection path will produce a response on a certain cell in the range-Doppler map in an ideal case without noise and interference. The mapping established in this step is based on this physical principle, which converts the complex spatial reflection geometry into a predictable, searchable coordinate correlation rule in the range-Doppler map.

[0047] The geometric position correlation relationship established by S2 realizes the conversion of the multi-path reflection geometry constraint in the physical space into a quantifiable coordinate correlation rule in the signal processing domain (range-Doppler map). This provides an accurate geometric prediction basis for the selection of candidate signal point sets in the subsequent steps, and improves the directivity and reliability of multi-path target recognition.

[0048] S3, according to the geometric position correlation relationship, all candidate signal point sets satisfying the geometric constraint condition are identified in the range-Doppler map, and the candidate target signal groups belonging to the same scattering source are selected through coherence analysis.

[0049] After establishing the geometric position correlation relationship between the real target and the multi-path target, considering that the range-Doppler map contains a large number of multi-path signal points generated by environmental reflection, if all signal points are processed without distinction, a large number of false targets will inevitably be introduced. Further, since the multi-path signal and the real target echo may be highly similar in amplitude and position, it is still difficult to reliably distinguish between the two only by relying on geometric constraints.

[0050] To solve the above problems, the present application adopts a two-stage discrimination strategy: first, using the established geometric position correlation relationship, the candidate signal point sets satisfying the geometric constraint condition are selected from the range-Doppler map, thereby effectively narrowing the analysis range and focusing on physically reasonable signal combinations. Secondly, coherence analysis is performed on these candidate signal point sets, and whether they come from the same physical scattering source is judged by evaluating the phase consistency between signals, laying a foundation for accurately distinguishing real target points and multi-path signal points.

[0051] As a preferred embodiment of the above steps, in the first stage, all candidate signal point sets satisfying the geometric constraint condition are identified as follows: S301, in the range-Doppler map, according to the signal amplitude, the signal points with an amplitude exceeding a predetermined threshold are selected as potential real target points.

[0052] In one example, the preset threshold can be set as a signal-to-noise ratio threshold, for example, 8 dB. Specifically, for each signal point in the range-Doppler map, the ratio of the echo power to the corresponding local background noise power is calculated; if the ratio is greater than 8 dB, the point is retained as a potential real target point. Wherein, the background noise power can be obtained by the data collected by the radar in the idle period.

[0053] Considering that in the millimeter wave radar system, the propagation loss of the direct path is the smallest, therefore the echo signal of the real target is usually stronger than the multi-path signal reflected once or more times. By setting the amplitude threshold, the focus can be preferentially focused on the strong response point with high signal-to-noise ratio, which is used as the initial assumption of the real target, thereby greatly reducing the subsequent analysis range and improving the processing efficiency.

[0054] S302, according to the geometric position correlation relationship between the real target point and each order multi-path target point, for each potential real target point, the theoretical positions of the first order multi-path and the second order multi-path in the range-Doppler map are calculated, and a set of predicted coordinates are obtained.

[0055] S303, taking each predicted coordinate as the center, search whether there is a signal point in the preset neighborhood, if there is, record the point as a suspected multi-path target point associated with the potential real target point.

[0056] If at least one suspected multi-path target point is detected in the above neighborhood, it is determined that these points and the potential real target point satisfy the multi-path geometric constraint condition.

[0057] S304, the potential real target point and all suspected multi-path target points associated with it which satisfy the multi-path geometric constraint condition are combined to form a candidate signal point set.

[0058] The identification of the candidate signal point set in the application is based on the characteristics that the multi-path echo and the direct echo come from the same physical scatterer: although the multi-path echo and the direct echo appear as multiple separated points in the range-Doppler map, their spatial arrangement follows the specific constraints determined by the tunnel geometry. Therefore, by searching the measured signal points in the neighborhood of the theoretically predicted multi-path position, the signal cluster which satisfies the physical law can be effectively identified, thereby excluding irrelevant clutter.

[0059] Referring to Figure 2 As shown in the continued implementation of S3 step, the second stage, the coherence analysis is as follows: S305, extracting the radar echo complex signal corresponding to each signal point in each candidate signal point set.

[0060] S306, respectively calculating the cross-correlation coefficient between the complex signal of each suspected multi-path target point in the candidate signal point set and the complex signal of the potential real target point, as the coherence measurement set of the candidate signal point set.

[0061] S307, the coherence measure set in each candidate signal point set is clustered: if the coherence measure distribution of a certain candidate signal point set presents a single cluster, it is determined that the signals in the set are highly coherent and belong to the same scattering source, and is recorded as a candidate target signal group; otherwise, it is determined that the points in the candidate signal point set belong to independent scattering sources.

[0062] In an example implementation, the clustering can employ a K-means clustering algorithm, and when all cross-correlation coefficients are clustered into one class, it is determined as a single cluster.

[0063] The coherence analysis adopted by the present application is based on the complex characteristics of millimeter wave radar echo signals, i.e. containing amplitude and phase information. When the signals of the same real target return to the radar through different propagation paths, although the lengths of the paths are different, the phase relationship remains highly consistent after time alignment because they originate from the same transmitted pulse and scattering time, thereby showing strong cross-correlation in the complex signal domain.

[0064] In a specific coherence analysis implementation, each candidate signal point set will generate a set of cross-correlation coefficients, respectively representing the coherence degree between each suspected multipath target point and the potential real target point. To determine whether these signals are homologous, the present application employs a clustering method to analyze the structure of the set of coherence measures, and by examining the relative distribution between the coefficients rather than relying on a fixed threshold, the inherent consistency is intuitively revealed.

[0065] S4, for the candidate target signal group, energy consistency verification is performed using the energy attenuation relationship between the real target and the multipath target to mark the real target signal points and the multipath signal points.

[0066] After coherence analysis, considering that relying only on phase consistency may still misjudge some strong clutter or accidental coherent interference as multipath signals.

[0067] To solve the above problem, the present application further introduces energy consistency verification, which utilizes the rule that multipath propagation paths are longer and experience reflection loss, and the echo amplitude of the multipath should be significantly lower than that of the direct path in theory. By checking whether the amplitude ratio conforms to this physical law, real targets and multipath signal points can be effectively distinguished on the basis of coherence analysis, providing suppression objects for subsequent multipath suppression.

[0068] Referring to Figure 3 As an embodiment of the present application, the energy consistency verification is as follows: S401, for the candidate target signal group, the signal amplitude values of the potential real target points and each suspected multipath target point in the range-Doppler plot are extracted respectively.

[0069] S402, the ratio of the amplitude value of each suspected multipath target point to the amplitude value of the potential real target point is calculated.

[0070] The amplitude ratio reflects the degree of signal attenuation of the multipath path relative to the direct path.

[0071] S403, compare the calculated amplitude ratio with the set energy attenuation threshold: if all amplitude ratios are less than the energy attenuation threshold, the energy consistency verification is successful, the potential real target point is confirmed as a real target signal point, and the associated suspected multipath target point is marked as a multipath signal point.

[0072] If any amplitude ratio is greater than or equal to the energy attenuation threshold, it means that the multipath point energy is abnormally high, which does not meet the energy attenuation characteristics of typical multipath propagation, and the energy consistency verification fails.

[0073] In one specific embodiment, the energy attenuation threshold represents the maximum amplitude ratio allowed for multipath signals relative to direct path echoes in a tunnel environment, and is usually set to 0.3, i.e. the multipath signal amplitude should not exceed 30% of the real target signal. When the tunnel wall reflection is strong or the target distance from the radar is close, the multipath signal may be strong, and the threshold can be large; when the wall has strong wave absorption or the target distance is far, the multipath signal is weaker, and the threshold can be small to avoid misjudgment.

[0074] S5, based on the spatial position and amplitude of the marked multipath signal points, generating a multipath signal spatial distribution estimation map.

[0075] After distinguishing the real target signal points and multipath target points through energy consistency verification, subsequent multipath suppression operation needs to be performed, however, since the identified multipath points are distributed discretely, it is difficult to directly use them to continuously and globally eliminate the multipath interference components in the range-Doppler map.

[0076] To solve the above problems, the information of the identified multipath signal points is used to construct the distribution of the multipath signal, which is beneficial to global multipath suppression.

[0077] In a preferred embodiment of the present application, the multipath signal spatial distribution estimation map is generated according to the following process: S501, collect the spatial position coordinates and signal amplitudes in the range-Doppler map of all multipath signal points marked as multipath signal points after energy consistency verification.

[0078] S502, based on the spatial position of the multipath signal points, using a distance-weighted spatial interpolation algorithm to estimate the multipath signal amplitude of all positions within the tunnel space coverage range.

[0079] Since the amplitude of multipath echoes is affected by the length of the propagation path, it usually presents a continuous and locally smooth distribution characteristic in the tunnel space. The closer the position to the known multipath point, the more similar its multipath interference amplitude should be; otherwise, the difference is greater. Distance-weighted interpolation embodies the spatial correlation of near and far by giving higher weight to adjacent samples, which conforms to the physical law of electromagnetic wave propagation.

[0080] In an optional example, the distance-weighted spatial interpolation algorithm is implemented as follows: first, divide the tunnel space covered by radar detection into a regular three-dimensional grid.

[0081] Secondly, take a set of known discrete multipath signal points as input, each point containing its three-dimensional spatial coordinates and the corresponding signal amplitude.

[0082] Subsequently, for any grid node in the tunnel space coverage, i.e. the to-be-estimated point, take the reciprocal of its Euclidean distance to each known multipath signal point as the interpolation weight.

[0083] Finally, the multipath amplitude estimate of the point is obtained by linearly weighting and combining the signal amplitudes of all known multipath signal points according to the above weights.

[0084] S503, map the signal amplitude estimated by interpolation back to the corresponding coordinates of the range-Doppler plot to form a multipath signal space distribution estimation plot with the same size as the original range-Doppler plot.

[0085] When performing multipath suppression, the key step is to subtract the multipath interference component on the original range-Doppler plot. In order to ensure the accuracy of this subtraction operation, it is necessary to ensure the spatial alignment between the multipath distribution estimation plot and the original range-Doppler plot. Only under this alignment condition, the multipath amplitude to be removed can be calculated on each Doppler unit, so as to avoid the problem of actual target misdeletion or interference residue caused by image misalignment.

[0086] To achieve the above purpose, the signal amplitude estimated by interpolation needs to be mapped back to the corresponding coordinate position of the original range-Doppler plot, and the specific implementation process is as follows: first, for each reflection path, calculate the total propagation path length of the electromagnetic wave from the target to the radar receiver through the reflection surface based on the geometric optics principle, and deduce the corresponding round-trip time delay of the path; at the same time, according to the radial component of the relative motion speed between the target and the radar in the equivalent propagation direction, the Doppler shift amount corresponding to the path echo is calculated by using the Doppler effect formula.

[0087] Secondly, the round-trip time delay of each reflection path is converted into the discrete unit index of the distance axis in the range-Doppler map, and the corresponding Doppler shift is converted into the discrete unit index of the Doppler axis; based on this mapping relationship, the estimated multipath signal amplitude value on the spatial grid is assigned to the corresponding coordinate position in the range-Doppler map.

[0088] Finally, through the above mapping process, a multipath signal distribution estimation map with the same size and pixel alignment as the original range-Doppler map is generated.

[0089] S6, according to the multipath signal spatial distribution estimation map, subtract the corresponding signal component from the original range-Doppler map by a subtraction ratio.

[0090] After constructing the multipath signal spatial distribution estimation map strictly aligned with the original range-Doppler map, targeted multipath suppression can be implemented in the original range-Doppler map, and the specific suppression process is as follows: S601, map each spatial position in the multipath signal spatial distribution estimation map to the corresponding coordinate unit of the original range-Doppler map.

[0091] S602, for each coordinate unit, the ratio of the signal amplitude value of each unit in the multipath signal spatial distribution estimation map to the signal amplitude of the corresponding unit in the original range-Doppler map is taken as the subtraction ratio.

[0092] S603, according to the subtraction ratio, subtract the corresponding proportion of multipath signal components from the corresponding unit of the original range-Doppler map.

[0093] S604, after completing the unit-by-unit signal subtraction, output the enhanced range-Doppler map with multipath interference suppressed.

[0094] The application of the above operation is that, since the original range-Doppler map contains both the true reflection signal from the target and the multipath interference caused by environmental reflection, and the multipath signal spatial distribution estimation map is a spatial distribution containing only multipath components predicted by an algorithm. Therefore, each cell in the estimation map represents the expected multipath signal amplitude.

[0095] When calculating the ratio of the signal amplitudes of the corresponding cells in the two maps, it is actually quantifying the degree of contribution of multipath interference at a particular position to the total received signal. This ratio, as a correction factor, can be used to accurately measure the amount of multipath interference that needs to be subtracted from the original signal, and according to this ratio, the impact of multipath interference can be minimized without completely removing the target signal.

[0096] S7, target detection is performed on the enhanced range-Doppler map after the subtraction process, and the true target point set is output.

[0097] After the multipath interference suppression is completed, the false targets caused by strong reflections in the original range-Doppler map are significantly reduced, but residual interference can still cause false detection.

[0098] To solve the above problems, the application performs target detection on the enhanced range-Doppler map and outputs a high-purity real target point set after the multipath false target is removed, thereby improving the target resolution capability of radar positioning data.

[0099] In the specific implementation of the above steps, the target detection includes the following: selecting cells exceeding the threshold from the amplitude response data of the full map based on the preset threshold to form an initial target point set.

[0100] It should be noted that although potential real target points have been preliminarily screened out by the preset threshold based on the original range-Doppler map before multipath suppression, this process inevitably contains false responses caused by strong multipath interference.

[0101] After multipath suppression, the multipath components in the original range-Doppler map have been significantly weakened, and the resulting enhanced range-Doppler map more truly reflects the direct echo distribution of the target. Therefore, it is necessary to re-perform target detection on the enhanced range-Doppler map to extract a higher-purity candidate target point set from the suppressed interference signal, thereby avoiding bringing false points misdetected due to multipath dominance into the subsequent processing flow.

[0102] For each candidate point in the initial target point set, the multipath signal estimation amplitude value of its corresponding coordinate position in the multipath signal space distribution estimation map is queried.

[0103] The estimated amplitude value is compared with the preset residual limit value, and the candidate point whose multipath signal estimation amplitude value exceeds the residual limit value is removed, and the remaining points are constructed and output as a real target point set.

[0104] The setting of the residual limit value in the above is to distinguish between real target responses and multipath residuals that are not completely suppressed, and the limit value can be set to 5%-10% of the estimated amplitude of the corresponding position in the multipath signal space distribution estimation map.

[0105] Considering that even after multipath suppression processing, there can still be multipath interference that is not completely eliminated in strong reflection areas, by referring to the multipath signal space distribution estimation map, the expected multipath amplitude of each position can be known.

[0106] When a candidate target point is detected, its estimated value in the map is queried to determine whether it is located in a high multipath interference zone. If the estimated multipath amplitude of a point exceeds the residual limit, it means that multipath interference can dominate the signal response of the point, which indicates that the point is likely to be a false target caused by multipath effect. Therefore, these points need to be removed. This process is essentially a secondary verification of the preliminary detection result using multipath prior knowledge to ensure the accuracy of the final real target.

[0107] Through the above scheme, the present application realizes early identification and suppression of multipath interference in the signal domain of the range-Doppler map, and comprehensively utilizes the three physical constraints of geometry, coherence and energy. Compared with the prior art, the present application has the advantages of moving the source of suppression forward, having more reliable basis for discrimination, being insensitive to early processing errors of point clouds, and the like, and significantly improves the accuracy and reliability of target detection in a tunnel environment.

[0108] The above embodiments can be realized wholly or partially by software, hardware, firmware or any combination thereof. When realized by software, the above embodiments can be realized wholly or partially in the form of a computer program product.

[0109] Those skilled in the art can appreciate that the modules and algorithm steps of the examples described in conjunction with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solutions. Those skilled in the art can use different methods to realize the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0110] In addition, the functional modules in each embodiment of the present application can be integrated in one processing module, or each module can exist physically alone, or two or more modules can be integrated in one module.

[0111] The above is merely specific embodiments of the present application, but the protection scope of the present application is not limited thereto. Any modification or replacement within the technical scope disclosed in the present application can be easily thought of by those skilled in the art, and should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

[0112] Finally, the above is merely preferred embodiments of the present application, and is not intended to limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. A method for suppressing multipath targets using millimeter-wave radar in a tunnel environment, characterized in that, include: The millimeter-wave radar echo signal in the tunnel is processed to generate a range-Doppler map; Based on the tunnel geometry, establish the geometric positional relationship between the real target and multipath targets of various orders in the range-Doppler map; Based on the geometric position correlation, all candidate signal point sets that meet the geometric constraints are identified in the range-Doppler map, and candidate target signal groups belonging to the same scattering source are screened out through coherence analysis. For candidate target signal groups, the energy consistency is verified by using the energy attenuation relationship between real targets and multipath targets, so as to mark real target signal points and multipath signal points; Based on the spatial location and amplitude of the marked multipath signal points, a spatial distribution estimation map of the multipath signal is generated. Based on the spatial distribution estimation map of the multipath signal, the corresponding signal components are subtracted from the original range-Doppler map according to the subtraction ratio; Target detection is performed on the enhanced range-Doppler image after subtraction, and the true target point set is output.

2. The method for suppressing multipath targets in a tunnel environment using millimeter-wave radar as described in claim 1, characterized in that: The relationship between the geometric positions of real targets and multipath targets of various orders in the range-Doppler map is established as follows: Using the phase center of the millimeter-wave radar as the origin of the coordinate system, the radar line of sight is defined as the positive z-axis, the horizontal direction as the x-axis, and the vertical direction as the y-axis, thus constructing a three-dimensional spatial coordinate system. Based on the tunnel design drawings, model the geometric equations of the tunnel wall in a three-dimensional spatial coordinate system. The wall includes at least the two side walls, the vault, and the ground. In the geometric equations, the direct path between the radar and the target is defined as the line-of-sight path, the path of the electromagnetic wave after reflection from any single wall is defined as the first-order multipath path, and the path of the electromagnetic wave after reflection from two walls is defined as the second-order multipath path. Based on the principle of ray tracing, the propagation distance and equivalent radial velocity of the direct line-of-sight path, the first-order multipath path, and the second-order multipath path are calculated respectively, and mapped to the corresponding coordinates of the range-Doppler diagram to establish the geometric positional relationship between the real target point and the target points of each order of multipath.

3. The method for suppressing multipath targets in a tunnel environment using millimeter-wave radar as described in claim 2, characterized in that: The identified set of all candidate signal points that satisfy the geometric constraints is as follows: In the distance-Doppler image, signal points whose amplitude exceeds a preset threshold are selected as potential real target points based on the signal amplitude. Based on the geometric positional relationship between the real target point and the target points of each multipath, for each potential real target point, the theoretical positions of the corresponding first-order multipath and second-order multipath in the distance-Doppler map are calculated to obtain a set of predicted coordinates; Centered on each predicted coordinate, search for actual signal points within a preset neighborhood. If a signal point exists, record it as a suspected multipath target point associated with a potential real target point. If at least one suspected multipath target point is detected in the aforementioned neighborhood, then these points are determined to satisfy the multipath geometric constraint condition with the potential real target point. Potential real target points are combined with all suspected multipath target points associated with them that meet the multipath geometric constraints to form a candidate signal point set.

4. The method for suppressing multipath targets in a tunnel environment using millimeter-wave radar as described in claim 3, characterized in that: The coherence analysis is as follows: Extract the complex radar echo signal corresponding to each signal point in each candidate signal point set; Calculate the cross-correlation coefficient between the complex signal of each suspected multipath target point and the complex signal of the potential real target point in the candidate signal point set, and use it as the coherence metric set of the candidate signal point set.

5. The method for suppressing multipath targets in a tunnel environment using millimeter-wave radar as described in claim 4, characterized in that: The candidate target signal groups belonging to the same scattering source, as selected through coherence analysis, are as follows: Cluster the set of coherence metrics in each candidate signal point set; If the coherence metric distribution of a candidate signal point set shows a single cluster, then the signals at each point in the set are determined to be highly coherent and belong to the direct-view and multipath components generated by the same scattering source, and are denoted as the candidate target signal group. Conversely, if the candidate signal points are not found to be independent scattering sources, then the points in the candidate signal point set are determined to be independent scattering sources.

6. The method for suppressing multipath targets in a tunnel environment using millimeter-wave radar as described in claim 3, characterized in that: The energy consistency verification includes the following: For candidate target signal groups, the signal amplitude values ​​of potential real target points and each suspected multipath target point in the group are extracted in the range-Doppler map; Calculate the ratio of the amplitude value of each suspected multipath target point to the amplitude value of the potential true target point; The calculated amplitude ratios are compared with the set energy attenuation threshold. If all amplitude ratios are less than the energy attenuation threshold, the energy consistency verification is successful, the potential real target point is confirmed as the real target signal point, and the associated suspected multipath target point is marked as the multipath signal point. If any amplitude ratio is greater than or equal to the energy decay threshold, the energy consistency verification fails.

7. The method for suppressing multipath targets in a tunnel environment using millimeter-wave radar as described in claim 1, characterized in that: The multipath signal spatial distribution estimation map is generated as follows: Collect the spatial coordinates of all points marked as multipath signal points after passing energy consistency verification and their signal amplitude in the range-Doppler plot; Based on the spatial location of multipath signal points, a distance-weighted spatial interpolation algorithm is used to estimate the multipath signal amplitude at all locations within the tunnel's spatial coverage area. The signal amplitude obtained by interpolation is mapped back to the corresponding coordinates of the range-Doppler map, forming a multipath signal spatial distribution estimation map with the same size as the original range-Doppler map.

8. The method for suppressing multipath targets in a tunnel environment using millimeter-wave radar as described in claim 1, characterized in that: The reduction ratio is described in the following steps: Map each spatial location in the multipath signal spatial distribution estimation map to the corresponding coordinate cell in the original distance-Doppler map; For each coordinate cell, the ratio of the signal amplitude value of each cell in the multipath signal spatial distribution estimation map to the signal amplitude of the corresponding cell in the original distance-Doppler map is used as the subtraction ratio.

9. The method for suppressing multipath targets in a tunnel environment using millimeter-wave radar as described in claim 8, characterized in that: The following process is used to subtract the corresponding signal component from the original distance-Doppler image according to the subtraction ratio: Subtract the multipath signal components from the corresponding cells of the original distance-Doppler map according to the subtraction ratio of the coordinate cells; After completing the unit-by-unit signal subtraction, the output is an enhanced range-Doppler image that suppresses multipath interference.

10. The method for suppressing multipath targets in a tunnel environment using millimeter-wave radar as described in claim 1, characterized in that: The target detection based on the enhanced range-Doppler image after subtraction includes the following: For the enhanced range-Doppler map, cells exceeding the threshold are selected from the amplitude response data of the entire map based on a preset threshold to form an initial target point set; For each candidate point in the initial target point set, query the estimated amplitude value of the multipath signal at its corresponding coordinate position in the multipath signal spatial distribution estimation map; The estimated amplitude value is compared with the configured residual limit. Candidate points whose estimated amplitude value of the multipath signal exceeds the residual limit are eliminated, and the remaining points are used to form and output the true target point set.

Images