A method for robust geometry reconstruction of millimeter wave radar room boundaries in the presence of multipath

Abstract

The application belongs to the technical field of millimeter wave radar. The application provides a millimeter wave radar room boundary robust geometric reconstruction method for multipath suppression. The embodiments of the present disclosure realize stable extraction of the real wall structure, effectively suppress the multipath echo and outlier point interference, and complete room boundary reconstruction and size accurate estimation by combining geometric constraints through antenna calibration, construction of a spatial point cloud geometric model, introduction of a robust straight line fitting and sequential model separation strategy. Through the sequential fusion and sequential robust geometric fitting strategy, the room wall boundary can be stably and accurately extracted, and the real wall and false echo can be effectively separated. In terms of wall detection integrity, anti-outlier point capability and size estimation accuracy, the embodiments of the present disclosure have obvious advantages, the room length and width estimation error is low, the boundary recognition success rate is significantly improved, and the embodiments of the present disclosure have good robustness and engineering applicability.

Description

Technical Field

[0001] This disclosure relates to the field of millimeter-wave radar technology, and more particularly to a robust geometric reconstruction method for millimeter-wave radar room boundaries for multipath suppression. Background Technology

[0002] Millimeter-wave radar possesses advantages such as all-weather, all-day operation, strong resistance to changes in lighting conditions, and good privacy protection, making it a promising candidate for applications in indoor sensing. By sensing the spatial structure and dimensions of a room using radar, it can provide fundamental environmental priors for tasks such as personnel positioning, behavior recognition, fall detection, robot navigation, and intelligent security.

[0003] Currently, most mainstream millimeter-wave radars adopt the TDM-MIMO system, which achieves high angular resolution in the horizontal direction through virtual array expansion, and combines range-angle (Range-Angle) images with range-dimensional FFT and direction-of-arrival (DOA) estimation algorithms to obtain the spatial distribution of indoor scatterers.

[0004] In room size estimation applications, walls are typically large, stable, and highly reflective surfaces, so wall echoes can be used to achieve boundary detection and size calculation.

[0005] Existing methods typically follow this process: first, range-dimensional FFT processing is performed on the radar echo to obtain range information; then, beamforming is used to estimate the azimuth angle and generate a range-angle distribution map; next, the CFAR algorithm is used for target detection; finally, the wall position is inferred through histogram statistics and scatter distribution, and the room length and width are calculated.

[0006] The aforementioned existing technical solutions have the following shortcomings in practical applications: 1) Lack of multipath suppression capability: In the indoor environment, there are multipath propagation phenomena such as wall reflection, ground reflection, ceiling reflection and secondary reflection of objects. Multipath echoes and real wall echoes are superimposed in the distance-angle domain, which can be easily misjudged as effective boundaries, resulting in increased size estimation errors.

[0007] 2) Projection interference problem: When the radar only has the ability to resolve azimuth angle, the echoes at different elevation heights are superimposed and projected in the two-dimensional plane, and the point cloud on the wall is distributed in a banded pattern rather than an ideal straight line. Traditional statistical or least squares fitting methods are difficult to accurately extract the true boundary.

[0008] 3) Poor ability to resist outliers: Clutter targets such as furniture, appliances and human bodies generate a large number of discrete points. Least square fitting or peak detection is easily affected by outliers, resulting in straight line deviation or misidentification.

[0009] 4) Reliance on empirical thresholds and insufficient robustness: Histogram methods require manual setting of threshold parameters, and repeated debugging is required in different scenarios, making it difficult to achieve adaptive and stable operation.

[0010] 5) Lack of a clear geometric modeling mechanism: Existing methods mostly stay at the level of signal statistics and do not model the wall from the perspective of spatial geometry, making it difficult to guarantee the physical consistency and accuracy of the reconstruction results.

[0011] Therefore, it is necessary to improve one or more of the problems existing in the above-mentioned related technical solutions.

[0012] It should be noted that this section is intended to provide background or context for the technical solutions of this disclosure as set forth in the claims. The description herein does not constitute an admission that it is prior art simply because it is included in this section. Summary of the Invention

[0013] The purpose of this disclosure is to provide a robust geometric reconstruction method for millimeter-wave radar room boundaries for multipath suppression, thereby overcoming, at least to some extent, one or more problems caused by the limitations and defects of related technologies.

[0014] According to embodiments of this disclosure, a robust geometry reconstruction method for millimeter-wave radar room boundaries oriented towards multipath suppression is provided, comprising: Step S1: Acquisition of raw radar data The millimeter-wave radar transmits linear frequency modulated continuous wave signals into the indoor space and receives echo signals. The echo signals are then mixed, filtered, and converted from analog to digital to obtain complex baseband data, forming multiple frames of raw radar data. Step S2: Distance-Angle Information Extraction Perform range-dimensional FFT on each frame of raw radar data to obtain the complex echo of each range cell; The range image is obtained based on the complex echoes of each range cell; Based on the range profile, spatial spectrum estimation is performed according to the array receiving vector on the same range cell to obtain the target direction, thereby generating a range-angle map; Constant false alarm rate detection is performed on the distance-angle map to obtain a set of effective target points in multiple frames; Step S3: Temporal fusion and multipath suppression The time domain is accumulated for the set of effective target points in multiple frames, the occurrence frequency of each target point is counted, and points that occur stably are selected according to a preset stability threshold to form a high-confidence candidate point cloud for the wall. Step S4: Spatial Point Cloud Modeling Convert the polar coordinates of the candidate point cloud on the wall to Cartesian coordinates to construct a two-dimensional spatial point cloud set; Step S5: Sequential Robust Geometric Fitting A robust line fitting algorithm is used to fit the two-dimensional point cloud set. Line models of multiple walls are extracted sequentially. After each wall line model is extracted, the interior points corresponding to the line are removed. The remaining point cloud is then fitted again until the required number of wall line models are obtained. Step S6: Determine geometric relationships and solve intersection points Calculate the angle between the direction vectors of the straight line models on each wall surface, and find the intersection point of the straight line pairs that satisfy the approximately perpendicular relationship to obtain multiple wall corner coordinates; Step S7: Room Dimension Reconstruction The length and width of the room are calculated based on the Euclidean distance between all corner coordinates, thus achieving geometric reconstruction of the room boundary.

[0015] Furthermore, the method also includes: A corner reflector is placed at a preset position, and raw data is acquired using TDMA-MIMO mode. After performing range dimension FFT processing on the raw data, frequency compensation and phase amplitude compensation are performed to complete the radar channel calibration.

[0016] Further, the step of generating a range-angle map by performing spatial spectrum estimation based on the range image and the array receive vectors on the same range cell to obtain the target direction includes: Based on the range image, the array reception vector is constructed using all array data elements on the same range cell:

[0017] In the formula, For the first Each antenna received a value. , This represents the total number of antennas. Construct the covariance matrix based on all array receive vectors:

[0018] In the formula, L is the accumulated number of Chirp. Represents conjugate transpose. For the first Each antenna received a value; Inverse eigenvalue decomposition of the covariance matrix yields the noise subspace matrix. ; The MUSIC spectrum is calculated based on the noise subspace matrix to generate a distance-angle map; where the MUSIC spectrum is:

[0019] In the formula, For azimuth angle variables, The spectral peak position corresponds to the direction in which the target is located, which is the array steering vector.

[0020] Furthermore, in step S3, the number of times each target point appears in the multi-frame constant false alarm rate detection results is counted:

[0021] in, Representing the The number of times each point appears. For the set of valid target points, For the first One effective target point; when If the condition is met, then the point is considered a stable point and retained; where, As the stability threshold, Total number of frames; Iterate through all valid target points in the set of valid target points to obtain a stable point cloud set. .

[0022] Furthermore, in step S5, the linear model is:

[0023] In the formula, For two-dimensional point cloud coordinates, a As the first coefficient, b As the second coefficient, c It is the third coefficient; The distance from the point to the line is:

[0024] when Then the point is determined to be an interior point; where; This is the distance tolerance.

[0025] Furthermore, robust line fitting algorithms include: The random sampling consensus algorithm is used to search for the line model with the most interior points in the point cloud as the first wall. Remove the interior points corresponding to the line; Repeat the fitting process in the remaining point cloud to extract the second and third walls in sequence.

[0026] Furthermore, in step S6, the intersection point of the straight lines whose angle between the direction vectors is within a preset range is obtained and used as the corner coordinates.

[0027] The technical solutions provided by the embodiments of this disclosure may include the following beneficial effects: In the embodiments of this disclosure, the robust geometric reconstruction method for room boundaries using millimeter-wave radar with multipath suppression, on the one hand, achieves stable extraction of the real wall structure by antenna calibration, constructing a spatial point cloud geometric model, and introducing robust linear fitting and sequential model separation strategies. This effectively suppresses multipath echoes and outlier interference, and combines geometric constraints to complete room boundary reconstruction and accurate size estimation. On the other hand, through temporal fusion and sequential robust geometric fitting strategies, the room wall boundaries can be extracted stably and accurately, achieving effective separation of the real wall surface from false echoes. It exhibits significant advantages in wall detection integrity, outlier resistance, and size estimation accuracy, with low room length and width estimation errors, significantly improved boundary recognition success rate, and good robustness and engineering applicability. Attached Figure Description

[0028] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure. It is obvious that the drawings described below are merely some embodiments of this disclosure, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0029] Figure 1 A step diagram illustrating a robust geometry reconstruction method for millimeter-wave radar room boundaries for multipath suppression in an exemplary embodiment of this disclosure is shown. Figure 2 A flowchart illustrating a robust geometry reconstruction method for millimeter-wave radar room boundaries for multipath suppression in an exemplary embodiment of this disclosure is shown. Figure 3 This diagram illustrates a room scene in an exemplary embodiment of the present disclosure. Figure 4 This diagram illustrates a simulated point cloud of a room scene in an exemplary embodiment of this disclosure. Figure 5 This diagram illustrates the results on simulated point cloud data in an exemplary embodiment of this disclosure. Figure 6 This document shows actual photographs of a measured room scene in an exemplary embodiment of this disclosure. Figure 7 This diagram illustrates a point cloud of a measured room scene in an exemplary embodiment of this disclosure. Figure 8 This diagram illustrates the results on measured data in an exemplary embodiment of this disclosure. Detailed Implementation

[0030] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that this disclosure will be more comprehensive and complete, and will fully convey the concept of the exemplary embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0031] Furthermore, the accompanying drawings are merely illustrative diagrams of embodiments of this disclosure and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities.

[0032] This example implementation provides a robust geometry reconstruction method for millimeter-wave radar room boundaries for multipath suppression. (Reference) Figure 1 As shown, this robust geometry reconstruction method for millimeter-wave radar room boundaries for multipath suppression may include: Step S1: Acquisition of raw radar data The millimeter-wave radar transmits linear frequency modulated continuous wave signals into the indoor space and receives echo signals. The echo signals are then mixed, filtered, and converted from analog to digital to obtain complex baseband data, forming multiple frames of raw radar data. Step S2: Distance-Angle Information Extraction Perform range-dimensional FFT on each frame of raw radar data to obtain the complex echo of each range cell; The range image is obtained based on the complex echoes of each range cell; Based on the range profile, spatial spectrum estimation is performed according to the array receiving vector on the same range cell to obtain the target direction, thereby generating a range-angle map; Constant false alarm rate detection is performed on the distance-angle map to obtain a set of effective target points in multiple frames; Step S3: Temporal fusion and multipath suppression The time domain is accumulated for the set of effective target points in multiple frames, the occurrence frequency of each target point is counted, and points that occur stably are selected according to a preset stability threshold to form a high-confidence candidate point cloud for the wall. Step S4: Spatial Point Cloud Modeling Convert the polar coordinates of the candidate point cloud on the wall to Cartesian coordinates to construct a two-dimensional spatial point cloud set; Step S5: Sequential Robust Geometric Fitting A robust line fitting algorithm is used to fit the two-dimensional point cloud set. Line models of multiple walls are extracted sequentially. After each wall line model is extracted, the interior points corresponding to the line are removed. The remaining point cloud is then fitted again until the required number of wall line models are obtained. Step S6: Determine geometric relationships and solve intersection points Calculate the angle between the direction vectors of the straight line models on each wall surface, and find the intersection point of the straight line pairs that satisfy the approximately perpendicular relationship to obtain multiple wall corner coordinates; Step S7: Room Dimension Reconstruction The length and width of the room are calculated based on the Euclidean distance between all corner coordinates, thus achieving geometric reconstruction of the room boundary.

[0033] The robust geometric reconstruction method for room boundaries using millimeter-wave radar, designed for multipath suppression, achieves two key improvements. First, by calibrating the antenna and constructing a spatial point cloud geometric model, it introduces robust linear fitting and sequential model separation strategies to stably extract the true wall structure, effectively suppressing multipath echoes and outlier interference. Geometric constraints are then used to reconstruct the room boundary and accurately estimate its dimensions. Second, through temporal fusion and sequential robust geometric fitting strategies, it can stably and accurately extract the room wall boundaries, effectively separating the true wall from spurious echoes. This method demonstrates significant advantages in wall detection integrity, outlier resistance, and dimension estimation accuracy. It exhibits low room length and width estimation errors, significantly improved boundary recognition success rate, and good robustness and engineering applicability.

[0034] Below, we will refer to Figures 1 to 8 The steps of the robust geometry reconstruction method for millimeter-wave radar room boundaries for multipath suppression described in this example embodiment will be explained in more detail.

[0035] In one embodiment, the method uses TDM-MIMO millimeter-wave radar as the sensing hardware, and achieves wall boundary extraction and room size estimation through spatial point cloud modeling and robust geometric fitting. The implementation process is as follows: Figure 2 As shown in the diagram, the experimental scenario is as follows: Figure 3 As shown. This application specifically includes the following steps: First, radar channel calibration is performed. A corner reflector is placed at a distance of 5m, and raw data is acquired using TDMA-MIMO mode. The collected data is then processed using range-dimension FFT, and frequency compensation and phase amplitude compensation are performed to ensure synchronization and accuracy between channels, thereby ensuring the accuracy of subsequent radar data.

[0036] Step S1: Radar Data Acquisition. The TDM-MIMO millimeter-wave radar periodically transmits linear frequency modulated continuous wave signals into the indoor space and receives the echoes. After mixing, filtering, and ADC sampling, complex baseband data is obtained, forming a multi-frame radar raw data cube:

[0037] in, The number of chirps per frame. Number of sampling points per chirp This represents the number of virtual array antennas.

[0038] Step S2: Distance-Angle Information Extraction. Perform a distance-dimensional FFT on each frame of acquired data. The Fourier transform result is:

[0039] in, For distance cell index, For the first Complex echoes of distance cells. The distance calculation relationship is:

[0040] in At the speed of light, This is the frequency modulation slope. From this, the range image is obtained.

[0041] The array receive vector is constructed from all array data elements in the same distance cell:

[0042] in, For the first Each antenna receives a certain number of values. Construct the covariance matrix:

[0043] Where L is the accumulated number of Chirp, This represents the conjugate transpose. (For) And the inverse eigenvalue decomposition yields the noise subspace matrix as follows: This leads to the MUSIC spectrum:

[0044] in, For azimuth angle variables, The spectral peak position corresponds to the direction of the target, which is the array steering vector. This yields the range-angle map. 2D-CFAR is then performed on the range-angle map to remove background noise and obtain the set of valid target points. , where N is the total number of frames.

[0045] Step S3: Temporal Fusion and Multipath Suppression. Utilizing the stable characteristics of wall echoes in the time dimension, cumulative statistics are performed on multi-frame CFAR results, i.e.:

[0046] in, Representing the The number of times a point appears, when The point is retained at that time, where Using the stability threshold, a set of stable point clouds is obtained. Only stable scattering points that persist in most frames are retained to suppress multipath spurious points, dynamic targets, and transient clutter interference, thereby obtaining a high-confidence candidate point cloud for the wall surface. .

[0047] Step S4: Spatial point cloud modeling. Map the polar coordinate data to a spatial Cartesian coordinate system: , Construct a two-dimensional spatial point cloud set This enables the conversion from the signal domain to the geometric domain.

[0048] Step S5: Sequential Robust Geometric Fitting. To address projection interference and point cloud outliers caused by only having orientation-dimension resolution, a robust linear fitting strategy is introduced. The model used to identify outliers is a linear model.

[0049] The distance from the point to the line is:

[0050] when Then the point is determined to be an interior point, where This is the distance tolerance.

[0051] The robust fitting strategy is as follows: (1) Use the stochastic consistency estimation algorithm to search for the line model with the most interior points in the point cloud as the first wall; (2) Remove the interior points corresponding to the line; (3) Repeat the fitting process in the remaining point cloud to extract the second and third walls in turn; and obtain the geometric model of the three walls: ,in , , This method achieves automatic outlier removal through the maximum interior point criterion, thereby suppressing the effects of multipath and clutter.

[0052] Step S6: Geometric Relationship Determination and Intersection Point Calculation. Calculate the angle between the direction vectors of each line, and only calculate the intersection point for line pairs that satisfy the approximately perpendicular relationship.

[0053] To avoid geometric degradation caused by parallel walls, the corner coordinates are obtained as follows: and .

[0054] Step S7: Room Size Reconstruction. Calculate the room length and width parameters based on the Euclidean distance between the intersection points to achieve geometric reconstruction of the room boundaries, i.e.:

[0055] Through the above steps, robust room boundary extraction and spatial size estimation under complex multipath and projection interference conditions are achieved.

[0056] In one specific embodiment, to verify the effectiveness of the algorithm, this application verified the algorithm in both simulation and real-world scenarios: like Figure 4 As shown, this application simulates point cloud results for walls, multipath propagation, clutter, and noise. In the simulation, the length and width of the room are 6m and 8m, respectively. Based on this point cloud result, the algorithm of this application is used to detect walls in the simulated point cloud result. As described in steps S6 and S7 above, the wall is fitted three times, and the intersection points are calculated based on the vertical relationship to determine the geometric information of the room. The results are as follows. Figure 5 As shown in Table 1, the red, green, and blue lines represent the walls, and the purple intersections represent the intersections of the walls. The room boundary reconstruction results are shown in Table 1. The algorithm estimated the room length to be 5.98m and the width to be 8.01m in the simulation scenario, with estimation errors of 0.02m and 0.01m for the length and width, respectively. Figure 6 The image shows the actual experimental scenario; the room's length and width are 5.2m and 3m, respectively. The radar board used in this experiment was a 2243 cascaded board, operating in TDMA-MIMO mode with 12 transmit and 16 receive signals. First, the radar was calibrated, and then the radar echo data was analyzed. After performing range estimation, MUSIC angle super-resolution estimation, and CFAR target detection on the radar echo, the following results were obtained: Figure 7 The target point cloud information shown is consistent with the simulation process described above. The wall surface is fitted three times, and the intersection points are determined based on the vertical relationship to ascertain the room's geometric information. The results are as follows: Figure 8 As shown in Table 1, the red, green, and blue lines represent the walls, and the purple intersections represent the wall intersections. The room boundary reconstruction results are shown in Table 1. In the actual test scenario, the algorithm estimated the room length to be 5.21m and the width to be 3.04m, with estimation errors of 0.01m and 0.04m, respectively.

[0057] Table 1. Algorithm estimation error table for simulation and actual scenarios.

[0058] The proposed method was comprehensively validated through simulation and actual millimeter-wave radar acquisition experiments. Experimental results show that, under complex indoor environmental conditions such as multipath reflection, furniture clutter interference, and pitch-dimensional projection aliasing, this method, through temporal fusion and sequential robust geometric fitting strategies, can stably and accurately extract room wall boundaries, achieving effective separation of real walls from spurious echoes. This method demonstrates significant advantages in wall detection integrity, outlier resistance, and size estimation accuracy, with low room length and width estimation errors and a significantly improved boundary recognition success rate, exhibiting good robustness and engineering applicability.

[0059] Experiments have fully demonstrated that the robust geometric reconstruction method for room boundaries of millimeter-wave radar with multipath suppression proposed in this application is feasible and can effectively solve the problem of inaccurate room size estimation in complex indoor environments. It has high practical value and significance for promotion.

[0060] The robust geometric reconstruction method for room boundaries using millimeter-wave radar, designed for multipath suppression, achieves two key improvements. First, by calibrating the antenna and constructing a spatial point cloud geometric model, it introduces robust linear fitting and sequential model separation strategies to stably extract the true wall structure, effectively suppressing multipath echoes and outlier interference. Geometric constraints are then used to reconstruct the room boundary and accurately estimate its dimensions. Second, through temporal fusion and sequential robust geometric fitting strategies, it can stably and accurately extract the room wall boundaries, effectively separating the true wall from spurious echoes. This method demonstrates significant advantages in wall detection integrity, outlier resistance, and dimension estimation accuracy. It exhibits low room length and width estimation errors, significantly improved boundary recognition success rate, and good robustness and engineering applicability.

[0061] It should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," and "counterclockwise" in the above description indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing the embodiments of this disclosure and simplifying the description, and are not intended to indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the embodiments of this disclosure.

[0062] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of embodiments of this disclosure, "a plurality of" means two or more, unless otherwise explicitly specified.

[0063] In the embodiments of this disclosure, unless otherwise expressly specified and limited, the terms "installation," "connection," "linking," "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this disclosure according to the specific circumstances.

[0064] In embodiments of this disclosure, unless otherwise expressly specified and limited, "above" or "below" the second feature can include direct contact between the first and second features, or contact between the first and second features through another feature between them. Furthermore, "above," "over," and "on top" of the second feature includes the first feature being directly above or diagonally above the second feature, or simply indicates that the first feature is at a higher horizontal level than the second feature. "Below," "below," and "under" the second feature includes the first feature being directly below or diagonally below the second feature, or simply indicates that the first feature is at a lower horizontal level than the second feature.

[0065] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this disclosure. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. In addition, those skilled in the art can combine and integrate the different embodiments or examples described in this specification.

[0066] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the appended claims.

Claims

1. A robust geometric reconstruction method for room boundaries in millimeter-wave radar with multipath suppression, characterized in that, include: Step S1: Acquisition of raw radar data The millimeter-wave radar transmits linear frequency modulated continuous wave signals into the indoor space and receives echo signals. The echo signals are then mixed, filtered, and converted from analog to digital to obtain complex baseband data, forming multiple frames of raw radar data. Step S2: Distance-Angle Information Extraction Perform range-dimensional FFT on each frame of raw radar data to obtain the complex echo of each range cell; The range image is obtained based on the complex echoes of each range cell; Based on the range profile, spatial spectrum estimation is performed according to the array receiving vector on the same range cell to obtain the target direction, thereby generating a range-angle map; Constant false alarm rate detection is performed on the distance-angle map to obtain a set of effective target points in multiple frames; Step S3: Temporal fusion and multipath suppression The time domain is accumulated for the set of effective target points in multiple frames, the occurrence frequency of each target point is counted, and points that occur stably are selected according to a preset stability threshold to form a high-confidence candidate point cloud for the wall. Step S4: Spatial Point Cloud Modeling Convert the polar coordinates of the candidate point cloud on the wall to Cartesian coordinates to construct a two-dimensional spatial point cloud set; Step S5: Sequential Robust Geometric Fitting A robust line fitting algorithm is used to fit the two-dimensional point cloud set. Line models of multiple walls are extracted sequentially. After each wall line model is extracted, the interior points corresponding to the line are removed. The remaining point cloud is then fitted again until the required number of wall line models are obtained. Step S6: Determine geometric relationships and solve intersection points Calculate the angle between the direction vectors of the straight line models on each wall surface, and find the intersection point of the straight line pairs that satisfy the approximately perpendicular relationship to obtain multiple wall corner coordinates; Step S7: Room Dimension Reconstruction The length and width of the room are calculated based on the Euclidean distance between all corner coordinates, thus achieving geometric reconstruction of the room boundary.

2. The robust geometric reconstruction method for millimeter-wave radar room boundaries for multipath suppression as described in claim 1, characterized in that, The method also includes: A corner reflector is placed at a preset position, and raw data is acquired using TDMA-MIMO mode. After performing range dimension FFT processing on the raw data, frequency compensation and phase amplitude compensation are performed to complete the radar channel calibration.

3. The robust geometric reconstruction method for millimeter-wave radar room boundaries for multipath suppression according to claim 1, characterized in that, The step of generating a range-angle map by performing spatial spectrum estimation based on the range image and array receive vectors on the same range cell to obtain the target direction includes: Based on the range image, the array reception vector is constructed using all array data elements on the same range cell: In the formula, For the first Each antenna received a value. , This represents the total number of antennas. Construct the covariance matrix based on all array receive vectors: In the formula, L is the accumulated number of Chirp. Represents conjugate transpose. For the first Each antenna received a value; Inverse eigenvalue decomposition of the covariance matrix yields the noise subspace matrix. ; The MUSIC spectrum is calculated based on the noise subspace matrix to generate a distance-angle map; where the MUSIC spectrum is: In the formula, For azimuth angle variables, The spectral peak position corresponds to the direction in which the target is located, which is the array steering vector.

4. The robust geometric reconstruction method for millimeter-wave radar room boundaries for multipath suppression according to claim 3, characterized in that, In step S3, the number of times each target point appears in the multi-frame constant false alarm rate detection results is counted: in, Representing the The number of times each point appears. For the set of valid target points, For the first One effective target point; when If the condition is met, then the point is considered a stable point and retained; where, As the stability threshold, Total number of frames; Iterate through all valid target points in the set of valid target points to obtain a stable point cloud set. .

5. The robust geometric reconstruction method for millimeter-wave radar room boundaries for multipath suppression according to claim 4, characterized in that, In step S5, the linear model is as follows: In the formula, For two-dimensional point cloud coordinates, a As the first coefficient, b As the second coefficient, c It is the third coefficient; The distance from the point to the line is: when Then the point is determined to be an interior point; where; This is the distance tolerance.

6. The robust geometric reconstruction method for millimeter-wave radar room boundaries for multipath suppression according to claim 5, characterized in that, Robust line fitting algorithms include: The random sampling consensus algorithm is used to search for the line model with the most interior points in the point cloud as the first wall. Remove the interior points corresponding to the line; Repeat the fitting process in the remaining point cloud to extract the second and third walls in sequence.

7. The robust geometric reconstruction method for millimeter-wave radar room boundaries for multipath suppression according to claim 6, characterized in that, In step S6, the intersection point of the straight lines whose angle between the direction vectors is within a preset range is obtained and used as the corner coordinates.

Images