Target and environment measurement situation generation method based on multipath signals

Abstract

The invention discloses a multi-path signal-based target and environment measurement situation generation method, which is applied to non-line-of-sight detection. The invention provides a target and environment measurement situation generation method based on multipath signal utilization by utilizing elliptical envelope approximate solution, aiming at the problems that the prior art depends on building layout information prior and the applicable scene is relatively simple and regular. The method comprises the following steps: firstly, stably and continuously associating multi-path arrival time delays of different periods based on a nearest neighbor algorithm; secondly, identifying and complementing arrival time delays corresponding to diffraction, primary reflection and secondary reflection through a strategy based on a least square method; and then, based on a circular cross positioning method, the multipath arrival time delay corresponding to diffraction and primary reflection is used for realizing shielded target positioning, and the arrival time delay corresponding to secondary reflection is used for elliptical envelope approximate solution to reconstruct shielded building layout. According to the method, on the basis of obtaining the position of the shielding target, estimation and reconstruction of the irregular and continuous shielding wall can be realized.

Description

Technical Field

[0001] The present invention relates to the field of radar situation awareness technology, and in particular to a method for generating target and environment measurement situation based on multipath signals. Background Art

[0002] As a technical means to break through the constraints of line-of-sight (LOS) detection, non-line-of-sight (NLOS) detection has important research and application value. Non-line-of-sight detection methods make full use of the multipath propagation characteristics of electromagnetic waves to detect targets, and many studies have proved the feasibility of using multipath to detect targets.

[0003] The Ilmanau University of Technology in Germany tracked the detected multipath arrival time delay (Time of Arrivals, ToA) based on the global nearest neighbor algorithm, and obtained the two-dimensional position of the non-line-of-sight target by intersecting the multipath trajectory corresponding to the diffraction path and the first-order reflection (R. Zetik, M. Eschrich, S. Jovanoska, et al., "Looking behind a corner using multipath-exploiting UWB radar," IEEE Trans. Aerosp. Electron. Syst., vol. 51, no. 3, pp. 1916–1926, Jul. 2015). Using only ToA information will undoubtedly ignore the phase characteristics carried by the signal. To this end, the University of Electronic Science and Technology of China used the minimum variance distortionless response (MVDR) algorithm to estimate the angle of arrival (Direction of Arrivals, DoA) of the multipath echo, and proposed a multipath identification strategy that combines ToA information and DoA information (J. Chen, S. Guo, H. Luo, et al., "Non-line-of-sight multi-target localization algorithm for driver-assistance radar system," IEEE Trans. Veh. Technol., vol. 72, no. 4, pp. 5332-5337, Dec. 2022). Furthermore, the literature (Z.Xu, S.Guo, J.Chen, et al., “Multi-domain featuresbased NLOS target localization method for MIMO UWB Radar,” IEEE Sensors J., vol.23, no.23, pp.29314-29322, Oct.2023) also made full use of Doppler information and proposed a multipath identification strategy based on the range-Doppler (RD) topology structure. Based on the multipath sorting results and combined with DoA information, target positioning was achieved.In order to better deal with the impact of rough walls and noise on ToA estimation, the University of Electronic Science and Technology of China used the back projection (BP) imaging algorithm to identify the ghost corresponding to the first multipath by matching, and then derived the target position by mirror symmetry (S.Li, S.Guo, J.Chen, et al., "Multipletargets lofcalization behind Lshaped corner via UWB radar," IEEE Trans. Veh. Technol., vol. 70, no. 4, pp. 3087–3100, Apr. 2021).

[0004] However, the above methods all assume that the two-dimensional detection scene is known, which is impractical in practice. Without loss of generality, the walls of the building layout are divided into line-of-sight reflection surfaces and non-line-of-sight reflection surfaces based on whether the direct radar signal is reachable. The estimation of the line-of-sight reflection surface is relatively simple and can be achieved based on the radar direct echo and other auxiliary means. The University of Electronic Science and Technology of China used Hough transform to extract the shape and position of the line-of-sight wall (Z. Zhu, S. Guo, J. Chen, et al., "Non-line-of-sight targets localization algorithm via joint estimation of DoD and DoA," IEEE Trans. Instrum. Meas., vol. 72, pp. 1-11, Oct. 2023). The National University of Defense Technology used lidar to obtain building layout information within line of sight (H.Du, C.Fan, C.Cao, et al., "A novel NLOS target localization method with a synthetic bistatic MMW radar," in Proc. IEEE 11th Sensor Array Multichannel Signal Process. Workshop (SAM), Hangzhou, China, Jun. 2020, pp. 1-5).

[0005] The estimation of non-line-of-sight reflective surfaces is more complicated and requires the assistance of multipath signals. Based on the premise that the non-line-of-sight corridor is straight, the University of Electronic Science and Technology of China proposed a joint estimation algorithm for non-line-of-sight channel width and non-line-of-sight target position based on the geometric distribution characteristics of multipath ghosts, the true position of the target and the building layout (P.Wu, J.Chen, S.Guo, et al., "NLOS positioning for building layout and target based on association and hypothesis method," IEEE Trans. Geosci. Remote Sens., vol. 61, pp. 1-13, Mar. 2023). Furthermore, the literature (J.Chen, Y.Zhang, S.Guo, et al., “Joint estimation of NLOS buildinglayout and targets via sparsity-driven approach,” IEEE Trans.Geosci.RemoteSens., vol.60, pp.1-13, Jun.2022) proposed a joint estimation of the tilt angle of the non-line-of-sight reflective surface and the target position based on the shape reconstruction group sparse algorithm and the particle swarm optimization algorithm.

[0006] In general, the existing non-line-of-sight reflective surface estimation algorithms only consider linear scenarios, that is, the default scenario is a straight line. However, in actual complex urban environments, building outlines are often complex and changeable. Therefore, studying a target and environment measurement situation generation method based on multipath signal utilization can achieve the estimation and reconstruction of unknown non-line-of-sight walls, which is of great research significance. Summary of the invention

[0007] In view of the problem that the existing technology relies on prior information of building layout and has relatively simple rules for applicable scenarios, the present invention proposes a method for generating target and environment measurement situation based on multipath signals to achieve estimation and reconstruction of unknown non-line-of-sight walls.

[0008] The technical solution adopted by the present invention is:

[0009] A method for generating a target and environment measurement situation based on multipath signals is applied to a scene with a continuously moving non-line-of-sight target. The scene includes: two buildings, a concealed target and a single-transmitter-single-receiver radar, wherein one of the buildings is an L-shaped wall, and the horizontal wall of the L-shaped wall is defined as a first wall, and the vertical wall is defined as a second wall; the other building is a linear wall, and its wall is defined as a third wall, wherein the second wall is opposite to the third wall, the single-transmitter-single-receiver radar is arranged on the same side of the first wall, and the second wall is a shielding wall of an unknown non-line-of-sight shape, and the third wall is a line-of-sight wall, and the concealed target (i.e., the non-line-of-sight target) is a moving single target, which is located in the non-line-of-sight area formed by the second wall and the third wall;

[0010] Perform the following target localization and building layout estimation in scenes with continuously moving non-line-of-sight targets:

[0011] Step S1, data collection: using a single-transmitting and single-receiving radar to collect K (K ≥ 2) cycles of echo data for a scene with a continuously moving non-line-of-sight target;

[0012] Step S2, multipath tracking: extract the arrival time delay (ToA) of the echo data collected in step S1, associate the multipath arrival time delays of K periods based on the nearest neighbor algorithm, and obtain the multipath arrival delays of several multipaths in each period;

[0013] Step S3, multipath sorting: Based on the least squares method, the multiple multipath arrival delays τ obtained in step S2 are sorted. n (k) Perform multipath sorting to identify or complete the arrival delays corresponding to diffraction, primary reflection, and secondary reflection;

[0014] Step S4, target positioning: performing target positioning on the arrival delay corresponding to the diffraction and the first reflection obtained in step S3 based on the circular crossing calculation to obtain the position of the obscured target;

[0015] Step S5, layout estimation: Based on the arrival delay corresponding to the secondary reflection obtained in step S3 and the shielding target position obtained in step S4, the elliptical envelope approximation solution method is used to estimate the position of the secondary reflection point to reconstruct the shape of the shielding wall.

[0016] The present invention firstly associates the multipath arrival delays of different periods robustly and continuously based on the nearest neighbor algorithm; secondly, the arrival delays corresponding to diffraction, primary reflection and secondary reflection are identified and completed through a strategy based on the least squares method; then, based on the circular crossing positioning method, the multipath arrival delays corresponding to diffraction and primary reflection are used to realize the obscured target positioning, while the arrival delays corresponding to the secondary reflection are used for elliptical envelope approximation solution to reconstruct the obscured building layout.

[0017] Furthermore, step S2 specifically includes the following steps:

[0018] Step S201, using a moving target indicator to eliminate the influence of static background echo and antenna coupling in the echo data;

[0019] Step S202: extract the arrival delay in the echo data based on the constant false alarm detector, and store the arrival delay of the echo data of each period in the vector p in ascending order. (k) In represents the nth arrival delay extracted in the kth cycle, where n=1,…N k , N k represents the total number of arrival delays extracted in the kth cycle;

[0020] Step S203, equating the multipath arrival delays of K periods to one-dimensional motions of multiple targets, and performing tracking processing on them based on a nearest neighbor algorithm to obtain measurements associated with the track;

[0021] During the tracking process, for the jth arrival delay observed in the kth cycle, Its new information with the i-th multipath track is:

[0022]

[0023] Where H = [1,0] T is the measurement matrix, is the state prediction of the i-th multipath track;

[0024] Step S204, performing multipath tracking filtering on the multipath signals based on the specified track length threshold, and then obtaining the delay matrix M based on the multipath arrival delays of the multipath signals at all times after the multipath tracking filtering. A :

[0025]

[0026] in, is the multipath arrival delay of the nth multipath in the kth cycle, N is the total number of arrival delays in each cycle after multipath tracking filtering, that is, the number of tracks after multipath tracking filtering, k is the cycle index, the multipath index n=1,…N, the multipath delay vector of the nth multipath

[0027] Further, in step S3, based on the specific linear relationship between the multipaths, a specific strategy is formulated using the least squares method to achieve the identification or completion of the ToA of the diffraction, primary reflection and secondary reflection paths, which specifically includes the following steps:

[0028] Step S301, check whether N / 3≥1, if not, the building layout estimation cannot be realized, and the multipath sorting is terminated directly; if yes, initialize the number of multipath combinations that meet the conditions χ S = 0, and set σ T After the value of is taken, step S302 is executed;

[0029] Step S302: From the delay matrix M A Three multipath delay vectors are selected from the non-repeated ones and arranged in ascending order as p n1 , p n2 , p n3 , and form them into a matrix A = [p n1 p n3 ] and vector b = p n2 ;

[0030] Step S303, calculate the least squares solution x=(A T A) -1 A T b. Judge ||x-α||<σ T Is it true? If so, execute step S304, otherwise return to step S302; where the auxiliary amount α = [1 / 2, 1 / 2] T ;

[0031] Step S304: The currently selected multipath delay vector p n1 and p n3 Treated as basis vectors and stored in matrix M B and the multipath combination number χ S Increment by 1;

[0032] Check whether χ S =2 or all If all the screening combinations are traversed, if yes, execute step S305; otherwise, return to step S302;

[0033] Step S305: if x S =2, then the matrix M B The column vectors in are arranged in ascending order to obtain the arrival delay vectors corresponding to the diffraction path, the first reflection path and the second reflection path;

[0034] If x S =1, then there are at least 4 multipaths corresponding to the two-way multipath Based on the arrival delay relationship between the base multipath and the combined multipath, the arrival delay vector corresponding to the secondary reflection path is derived;

[0035] If x S = 0, it is considered that there are at least 3 multipaths corresponding to Based on the arrival delay relationship between the base multipath and the combined multipath, the arrival delay vectors corresponding to the diffraction and secondary reflection paths are derived; where: It represents the two-way multipath captured by the receiving antenna of the single-transmitter single-receiver radar, and the propagation path index i,j∈{0,1,2}, where propagation path 0 refers to the electromagnetic wave diffracting at the corner of the L-shaped wall and then propagating to the hidden target; propagation path 1 refers to the electromagnetic wave reflecting once on the third wall and then propagating to the hidden target; propagation path 2 refers to the electromagnetic wave reflecting once on the third wall and then propagating to the hidden target by the second reflection on the second wall, and i=j is the two-way multipath. It is called the base multipath, and the two-way multipath of i≠j It is called combined multipath.

[0036] Furthermore, step S4 specifically includes:

[0037] Define R1 as the radar mirror point and C as the corner of the L-shaped wall. The hidden target Q is located on the circle with the corner C as the center and ||CQ|| as the radius and on the circle with R1 as the center and ||R1Q|| as the radius. Among them, the radar mirror point R1 is the mirror-symmetrical position of the single-transmitter and single-receiver radar about the third wall. The position of the hidden target Q is obtained based on the circle intersection algorithm.

[0038] Furthermore, in step S5, based on the coupling relationship between the target and the building layout contained in the multipath signal, the shape of the shielding wall (the second wall) is dynamically reconstructed by sensing the moving shielding target.

[0039] Furthermore, step S5 specifically includes:

[0040] Define S to represent the secondary reflection point on the shielding wall, which is located on an ellipse with the concealed target Q and the radar mirror point R1 as the focus and the secondary reflection path length (||R1S||+||SQ||) as the major axis; where the radar mirror point R1 is the mirror symmetric position of the single-transmitter and single-receiver radar about the third wall;

[0041] The contour shape of the shielding wall is reconstructed based on the ellipse envelope generated by the continuously moving shielding target.

[0042] Furthermore, the envelope of the ellipse family is discretized into the common tangent of two adjacent ellipses, and the two corresponding tangent points can be approximately equivalent to the position of the secondary reflection point S. The contour shape of the shielding wall is reconstructed using the ellipse envelope approximation solution method.

[0043] The technical solution provided by the present invention brings at least the following beneficial effects:

[0044] (1) The multipath sorting strategy based on the least squares method of the present invention can identify and even complete the arrival delay ToA corresponding to the diffraction, primary reflection and secondary reflection paths;

[0045] (2) Based on the elliptical envelope approximation solution, the reflection points on the non-line-of-sight wall are approximately estimated, and then the non-line-of-sight wall is reconstructed;

[0046] (3) The present invention can be applied to disaster relief, environmental perception and other fields. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0048] Figure 1 A schematic diagram of a process of a method for generating a target and environment measurement situation based on multipath signals proposed in an embodiment of the present invention;

[0049] Figure 2 This is a multipath propagation model diagram for the case of irregular shielding walls;

[0050] Figure 3 Schematic diagram for locating obscured targets;

[0051] Figure 4 Results of target positioning and building layout estimation in three different non-line-of-sight wall environments;

[0052] in, Figure 4 (4-a) is a straight wall scene; (4-b) is a curved concave wall; (4-c) is a sudden concave wall.

[0053] Figure 5 Estimate lateral deviations of walls for three different non-line-of-sight wall environments;

[0054] Figure 6 This is the scene diagram of the sudden concave experiment;

[0055] Figure 7 for Figure 6 Multipath tracking results in the experimental scenario shown;

[0056] Figure 8 for Figure 7 Results of target localization and building layout estimation in the experimental scene shown;

[0057] in, Figure 8 (8-a) is the experimental result; (8-b) is the numerical simulation result; (8-c) is the estimated lateral deviation. DETAILED DESCRIPTION

[0058] To make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be described in detail and completely in conjunction with the drawings in the implementation of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings can be arranged and designed using different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the present application for protection, but merely represents selected embodiments of the present invention.

[0059] See also Figure 1 The embodiment of the present invention proposes a method for generating target and environment measurement situation based on multipath signals, which comprises the following steps:

[0060] Step 1: Signal Analysis:

[0061] A scene with irregular non-line-of-sight reflective walls such as Figure 2 As shown in the figure, Wall-3 of Building 2 and Wall-1 of Building 1 are located in the line-of-sight area, and their positions and shapes can be obtained through the direct reflection echo of the wall. However, the continuously moving target and Wall-2 of Building 1 are located in the non-line-of-sight area, and their related information cannot be obtained using the traditional direct-view radar detection method.

[0062] A portable single-transmitter single-receiver radar R = [x R ,y R ] T Placed on the side of Wall-1, the target and the radar are not visible to each other. In this case, the radar detects the target through multipath propagation of electromagnetic waves rather than direct signal. Based on the ray tracing model, the one-way electromagnetic propagation path from the transmitting antenna to the target can be classified into three types, see Figure 2 , which are specifically:

[0063] 1) Path-0: The electromagnetic wave first diffracts at the corner and then propagates to the target, which is recorded as

[0064] 2) Path-1: The electromagnetic wave is reflected once on Wall-3 and propagates to the target, which is recorded as

[0065] 3) Path-2: After the electromagnetic wave is reflected once on Wall-3, it is reflected again on Wall-2 and propagates to the target, which is recorded as

[0066] After multiple reflections, the electromagnetic wave will be greatly attenuated, so the multipath higher than the secondary reflection is ignored. If the building surface can provide sufficiently strong mirror reflection in the considered band (1-3GHz), the mirror symmetric position of the radar is:

[0067] R1=[x R1 ,y R1 ] T =[2x w3 -x R ,y R ] T (1)

[0068] Among them, x w3 is the horizontal coordinate of Wall-3. Then the relationship between the arrival delays corresponding to the above three one-way multipaths is as follows:

[0069]

[0070] Among them, c is the speed of light, and R1 is the mirror-symmetric position of the radar about the wall 3. In the above formula, the value on the left side of the equal sign can be obtained by signal processing means, while the coordinates of the target Q and the secondary reflection point S are unknown and need to be solved. Since the problem considered is two-dimensional, the coordinates of Q can be solved by equations related to τ0 and τ1. Geometrically speaking, the target Q is located at the same time on a circle with the corner C as the center and ||CQ|| as the radius and a circle with R1 as the center and ||R1Q|| as the radius. The sum of ||R1S|| and ||SQ|| is a certain value τ2c, so the secondary reflection point S is located on an ellipse with Q and R1 as the focus and the secondary reflection path length (||R1S||+||SQ||) as the major axis. Therefore, the embodiment of the present invention can achieve the estimation of the shielding wall by making full use of the target and building layout information contained in the multipath.

[0071] The two-way multipath captured by the receiving antenna is a combination of the above three multipaths. The corresponding multipath arrival delay is τ ij , there is the following relationship:

[0072]

[0073] It is not difficult to see from the above formula that the received multipath exists in the space spanned by diffraction, primary reflection and secondary reflection paths, so it is better to use i = j is called the base multipath, and i≠j is a combined multipath. Correspondingly, the round-trip multipath arrival delay τ of i=j is ij is called the base multipath arrival delay, such as τ 00 , τ 11 , τ 22 , and the round-trip multipath arrival delay τ of i≠jij is called the combined multipath arrival delay, τ 01 , τ 02 , τ 12 etc. It is worth mentioning that since the propagation attenuation of the primary reflection path is greater than that of the diffraction and secondary reflection paths, the energy of the multipath including the primary reflection is greater, e.g. and

[0074] Step 2: Multipath tracking:

[0075] 2.1 Use the moving target indicator (MTI) technology to eliminate the influence of static background echo and antenna coupling in the received echo signal.

[0076] 2.2 Use the Constant False Alarm Rate (CFAR) detector to extract the ToA (Time of Arrival) in the echo. Consider that the radar captures K cycles of echoes in total. For the ToA of the kth cycle of echo data, store it in ascending order in the vector p (k) middle:

[0077]

[0078] in, represents the nth arrival delay extracted in the kth cycle, and N k represents the total number of arrival delays extracted in the kth cycle;

[0079] 2.3 The multipath ToA changes of all K periods are equivalent to the one-dimensional motion of multiple targets, which are tracked based on the nearest neighbor algorithm. As a result, the residual clutter will be further filtered out and the prediction of some missing multipaths can be achieved.

[0080] Specifically, the multi-radial quantity p extracted from the kth frame is (k) That is, the measurement vector, then the observed j-th multipath delay value and the i-th multipath track’s new information are:

[0081]

[0082] Where H = [1,0] T is the measurement matrix, is the state prediction of the i-th multipath track.

[0083] Define S(k) as the innovation covariance matrix, then the Mahalanobis distance can be calculated as:

[0084]

[0085] The basic idea of ​​the nearest neighbor algorithm is to calculate the Mahalanobis distance and select the measurement with the smallest Mahalanobis distance as the measurement associated with the track;

[0086] 2.4 Since the measurements generated by noise will not appear in every cycle, the corresponding tracking results often show short tracks, while the measurements corresponding to multipath will exist at most times, and the missing parts can be predicted and supplemented by the multipath tracking algorithm. Therefore, tracks with a length less than K will be discarded, and the number of remaining tracks is represented by N;

[0087] 2.5 Multipath Tracking The multipath signals at all times obtained after filtering are stored in the matrix:

[0088]

[0089] in, It represents the multipath arrival delay corresponding to the n-th multipath in the k-th cycle.

[0090] Step 3: Multipath sorting:

[0091] From formula (3), we can see that there is a definite linear relationship between the delay values ​​of the base multipath and the combined multipath:

[0092]

[0093] Among them, the auxiliary amount α=[1 / 2,1 / 2] T .

[0094] Define p n1 , p n2 , p n3 For three from M A The selected non-repeated column vectors, that is, the selected multi-radial quantities, are arranged in ascending order, so there are There are combinations of these, and C represents the combination. Then, the matrix A = [p n1 p n3 ] and vector b = p n2 .

[0095] Formula (8) can be converted into the form Ax = b. Theoretically, if x = α, then p n1 and p n3 is exactly the base multipath, and p n2 is the combined multipath. That is, the closer x and α are, the higher the possibility that the two vectors in the corresponding matrix A are the basis multipath. In the case of multiple cycles, the number of equations (the number of rows of matrix A) is greater than the number of unknowns (the number of columns of matrix A), and solving x is an overdetermined problem. The least squares method is used to solve x as follows:

[0096] x=(A T A) -1 A T b (9)

[0097] However, due to the existence of target expansion and measurement errors, the solved x is close to but not strictly equal to α. Therefore, the embodiment of the present invention believes that when ||x-α||<σ T In the multipath combination, p n1 and p n3 is the base multipath. Among them, σ T is an empirical threshold with a smaller value. By analyzing the simulation data and the measured data, the threshold σ in this embodiment is T The value of is set to 0.05, usually the threshold σ T The magnitude can be set to 10 -1 ~10 -2 Therefore, theoretically, we only need to obtain two sets of x that meet the conditions to obtain all the base multipaths.

[0098] However, in actual environments, it is often difficult for the radar echo to contain all six types of multipaths mentioned above. In this case, the missing multipaths can be supplemented by the following steps:

[0099] 3.1 Initialization parameter χ S =0, which means that ||x-α||<σ T If N / 3≥1, execute 3.2. Otherwise, the number of multipaths is small and the building layout estimation cannot be realized, so the multipath completion step ends directly.

[0100] 3.2 From M A Select three radial quantities without duplication and arrange them in order from small to large as p n1 , p n2 , p n3 , and form them into a matrix A = [p n1 p n3 ] and vector b = p n2 ;

[0101] 3.3 Calculate x=(A based on the least squares method T A) -1 A T b. Judge ||x-α||<σ T Is it true? If so, go to step 3.4; otherwise, go to step 3.2:

[0102] 3.4 Consider the selected p n1 and p n3 is the basis vector and is stored in the matrix M B In, χ S = x S +1, if χ S =2 or all If the traversal of the possible situations is completed, execute step 3.5, otherwise execute step 3.2;

[0103] 3.5 If x S =2, then the arrival delay vectors corresponding to the diffraction path, the first reflection path, and the second reflection path can be obtained by transforming the matrix M B The column vectors in are arranged in ascending order;

[0104] If x S =1, it is considered that there are at least 4 multipaths corresponding to According to the relationship between the base multipath and the combined multipath, the arrival delay vector corresponding to the secondary reflection path can be derived;

[0105] If x S = 0, it is considered that there are at least 3 multipaths corresponding to According to the relationship between the base multipath and the combined multipath, the arrival delay vectors corresponding to the diffraction and secondary reflection paths can be derived.

[0106] Step 4: Target positioning:

[0107] like Figure 3 As shown, the position of the target can be determined by two circles, that is, the target is located at the intersection of a circle with corner C as the center and ||CQ|| as the radius and a circle with R1 as the center and ||R1Q|| as the radius.

[0108] Let the length of line segment CR1 be L and the slope be l CR , the slope of line segment QE (target Q perpendicular to line segment CR1) is l QE , then:

[0109]

[0110] According to the Pythagorean theorem, we can get the following relationship:

[0111]

[0112] Therefore, the length of line segment CE can be calculated as:

[0113]

[0114] The coordinates of the vertical point E can be expressed as:

[0115]

[0116] According to the geometric relationship, we get:

[0117]

[0118] Then the length of line segment EF can be expressed as:

[0119]

[0120] Then the position of the target Q is:

[0121]

[0122] Step 5: Layout Estimation:

[0123] The principle of building layout estimation is: the reflection point S is located on an ellipse with Q and R1 as the focus and the secondary reflection path length (||R1S||+||SQ||) as the major axis. The building layout can be obtained by inverting the reflection point of the secondary reflection on Wall-2. However, the ellipse has an optical property, that is, the light emitted from one focus and irradiated on the ellipse will pass through another focus. Therefore, it is unrealistic to directly solve the position of the reflection point S.

[0124] Consider a moving target, and the envelope of the ellipse family formed by it can outline the outline of Wall-2. However, due to the uncertainty of the target's motion, the envelope of the ellipse family formed is difficult to solve. The envelope of the ellipse family can be discretized into the common tangent of adjacent ellipses. Therefore, the position of the reflection point can be approximately equivalent to the tangent point position. The specific calculation method is as follows:

[0125] Let a represent the semi-major axis of the ellipse, and b represent the semi-minor axis of the ellipse, which can be calculated from the arrival delays corresponding to the first reflection path and the second reflection path, respectively. Let (x0, y0) represent the center coordinates of the ellipse, and θ represent the major axis inclination (counterclockwise is positive), then the general equation coefficients of the ellipse are:

[0126]

[0127] in,

[0128] Then two adjacent ellipses can be expressed as

[0129]

[0130] The subscripts 1 and 2 of the coefficients of each ellipse equation are the distinguishers of two adjacent ellipses.

[0131] Consider two cases. First, when the slope of the common tangent exists, its equation is y=lx+m, then the above formula can be simplified to:

[0132]

[0133] in,

[0134] When the above formula has multiple roots at the same time, it means that the common tangent is tangent to both ellipses at the same time. Therefore, according to the multiple root discriminant, we get:

[0135]

[0136] Solving the above equation, we get the slope l and intercept m. Therefore, the horizontal coordinates of the two tangent points are:

[0137]

[0138] Substitute the horizontal coordinate into the equation y=lx+m to obtain its corresponding vertical coordinate.

[0139] In the second case, consider that the slope of the common tangent is infinite, that is, its equation is x = h. Then, the equations of two adjacent ellipses are:

[0140]

[0141] Similarly, the multiple root discriminant of the above formula is:

[0142]

[0143] Solving the above equation, we get h, which represents the horizontal coordinate of the two tangent points. Then its vertical coordinate is:

[0144]

[0145] The (x1, y1) and (x2, y2) obtained through the above steps are the tangent point coordinates, which are then equivalent to the position of the reflection point. After all the reflection points are estimated, the Hampel filter is used to remove the obvious outliers in the results, and the Kalman filter is used to smooth the results to obtain the target positioning and building layout estimation results.

[0146] The effects of the embodiments of the present invention are further illustrated by the following simulation and actual measurement verification:

[0147] Simulation results:

[0148] Without loss of generality, consider three L-shaped building corner scenarios, such as Figure 4 As shown. Among them, Figure 4In the scenario (4-a), Wall-2 is a conventional straight wall to test the applicability of the algorithm in conventional scenarios; while Wall-2 in (4-b) is a curved concave wall to demonstrate the universality of the proposed algorithm in any continuous wall situation. Wall-2 in (4-c) is concave 0.2m inward at the ordinate 5-7m to simulate the door frame / window in the corridor in the actual situation. In the non-line-of-sight channel, a cylinder with a radius of 0.1m is used to simulate the target, which moves along the curve. In order to detect non-line-of-sight targets and realize scene estimation, a single-transmit and single-receive radar is placed on the same side of Wall-1 with a coordinate of (1,1)m. The above scenario is simulated using gprMax electromagnetic simulation software, and the transmission waveform is a step frequency with a center frequency of 1.5GHz and a bandwidth of 1GHz.

[0149] Figure 4 The target positioning and building layout estimation results obtained by the method proposed in the embodiment of the present invention in the above three scenarios are shown. It can be seen from the figure that the estimated position of the non-line-of-sight target is consistent with the actual position. The estimated secondary reflection points are mostly distributed on Wall-2, which outlines the contour of Wall-2 well. For the concave and convex wall scene, such as Figure 4 As shown in (4-c) in the figure, at the concave mutation point of the wall, the estimated reflection point position deviation is more serious. This is because the multipath propagation is more complicated in the building mutation area, and due to the mutation of the multipath signal delay, there is an association error in the multipath tracking link.

[0150] Considering that the target is a cylinder with a radius of 0.1m, there is a certain extension, and electromagnetic propagation in the actual environment often does not strictly follow ray tracing, so the theoretical position of the secondary reflection point is often difficult to obtain. In order to measure the performance of building layout estimation, the lateral deviation between the estimated reflection point and the real wall is used to measure the performance of building layout estimation, such as Figure 5 It can be seen that the estimated deviations are all less than 0.25m, which is within the acceptable range.

[0151] Test results:

[0152] Experiments with concave structures such as Figure 6 As shown. Corner C is located at (2.0, 2.0) m, and the horizontal coordinate of Wall-3 is 4.38 m. There is a door on Wall-2 that is 0.14 m inward, 1.85 m away from the corner, and 1.15 m wide. The transmitting antenna position of the selected ultra-wideband radar is (1.05, 0.94) m, and the receiving antenna position is (1.3, 0.66) m. A non-line-of-sight target moves along a straight line from (2.85, 3.4) m to (2.85, 5.9) m.

[0153] The method proposed in the embodiment of the present invention is used to track the multipath in the multi-period range image, and the result is as follows: Figure 7 As shown by the black dotted line. The white box indicates the multipath fluctuation caused by the concave wall. As can be seen from the figure, the method proposed in the embodiment of the present invention can better associate the multipath arrival delays of multiple periods and to a certain extent complete the multipath missing in some periods.

[0154] The target positioning and building layout estimation results obtained by the method proposed in the embodiment of the present invention are as follows: Figure 8 As shown in (8-a), it is clear that the target positioning results are all around the real trajectory, and the inverted reflection points roughly outline the shape of the wall. Based on the experimental parameters, the numerical simulation inversion results are shown in (8-b). Comparing the experimental results with the numerical simulation results, it can be seen that the distribution positions of the estimated reflection points and the theoretical reflection points are basically consistent.

[0155] In order to measure the deviation between the estimated building layout and the actual wall, the deviation between the calculated second-order reflection point horizontal coordinate and the actual wall horizontal coordinate is shown in (8-c). The maximum deviation is about 0.16m, which is within the acceptable range. The maximum error is located in the concave part of the wall, which is consistent with the simulation results. This proves the accuracy of the wall inversion.

[0156] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

[0157] The above are only some embodiments of the present invention. For those skilled in the art, several modifications and improvements can be made without departing from the creative concept of the present invention, which all belong to the protection scope of the present invention.

Claims

1. A method for generating target and environment measurement situation based on multipath signals, applied to a scene with a continuously moving non-line-of-sight target, the scene including: Two buildings, a hidden target and a single-transmitter single-receiver radar, wherein one of the buildings is an L-shaped wall, the horizontal wall of the L-shaped wall is defined as the first wall, and the vertical wall is defined as the second wall; the other building is a linear wall, and its wall is defined as the third wall, wherein the second wall is opposite to the third wall, the single-transmitter single-receiver radar is arranged on the same side of the first wall, and the second wall is a shielding wall of unknown shape in non-line-of-sight, and the third wall is a line-of-sight wall, and the hidden target is a moving single target, which is located in the non-line-of-sight area formed by the second wall and the third wall; it is characterized in that the following target positioning and building layout estimation are performed in a scene with a continuously moving non-line-of-sight target: Step S1, using a single-transmitting and single-receiving radar to collect K cycles of echo data for a scene with a continuously moving non-line-of-sight target; Step S2, multipath tracking: extract the arrival delay of the echo data collected in step S1, associate the multipath arrival delays of K periods based on the nearest neighbor algorithm, and obtain the multipath arrival delays of several multipaths in each period; Step S3, multipath sorting: Based on the least squares method, the multiple multipath arrival delays obtained in step S2 are sorted. Perform multipath sorting to identify or complete the arrival delays corresponding to diffraction, primary reflection, and secondary reflection; Step S4, target positioning: performing target positioning on the arrival delay corresponding to the diffraction and the first reflection obtained in step S3 based on the circular crossing calculation to obtain the position of the obscured target; Step S5, layout estimation: Based on the arrival delay corresponding to the secondary reflection obtained in step S3 and the shielding target position obtained in step S4, the elliptical envelope approximation solution method is used to estimate the position of the secondary reflection point to reconstruct the shape of the shielding wall.

2. The method according to claim 1, characterized in that Step S2 specifically includes the following steps: Step S201, using a moving target indicator to eliminate the influence of static background echo and antenna coupling in the echo data; Step S202: extract the arrival delay in the echo data based on the constant false alarm detector, and store the arrival delay of the report data of each cycle in the vector p in ascending order. (k) In represents the nth arrival delay extracted in the kth cycle, where n=1,…N k , N k represents the total number of arrival delays extracted in the kth cycle; Step S203, equating the multipath arrival delays of K periods to one-dimensional motions of multiple targets, and performing tracking processing on them based on a nearest neighbor algorithm to obtain measurements associated with the track; During the tracking process, for the jth arrival delay observed in the kth cycle, Its new information with the i-th multipath track is: Where H = [1,0] T is the measurement matrix, is the state prediction of the i-th multipath track; Step S204, performing multipath tracking filtering on the multipath signals based on the specified track length threshold, and then obtaining the delay matrix M based on the multipath arrival delays of the multipath signals at all times after the multipath tracking filtering. A : in, is the multipath arrival delay of the nth multipath in the kth cycle, N is the total number of arrival delays in each cycle after multipath tracking filtering, k is the cycle index, the multipath index n=1,…N, and the multipath delay vector of the nth multipath 3. The method according to claim 2, characterized in that The tracking process based on the nearest neighbor algorithm in step S203 is specifically as follows: Define S(k) as the innovation covariance matrix, according to the innovation v of the i-th multipath track ij (k) Calculate the Mahalanobis distance: The measurement with the smallest Mahalanobis distance is selected as the measurement associated with the track.

4. The method according to claim 2, characterized in that Step S3 specifically includes the following steps: Step S301, check whether N / 3≥1, if not, the building layout estimation cannot be realized, and the multipath sorting is terminated directly; if yes, initialize the number of multipath combinations that meet the conditions χ S = 0, and set σ T After the value of is taken, step S302 is executed; Step S302: From the delay matrix M A Three multipath delay vectors are selected from the non-repeated ones and arranged in ascending order as p n1 , p n2 , p n3 , and form them into a matrix A = [p n1 p n3 ] and vector b = p n2 ; Step S303, calculate the least squares solution x=(A T A) -1 A T b. Judge ||x-α||<σ T Is it true? If so, execute step S304, otherwise return to step S302; where the auxiliary amount α = [1 / 2, 1 / 2] T ; Step S304: The currently selected multipath delay vector p n1 and p n3 Treated as basis vectors and stored in matrix M B and the multipath combination number χ S Increment by 1; Check whether χ S =2 or all If all the screening combinations are traversed, if yes, execute step S305; otherwise, return to step S302; Step S305: if x S =2, then the matrix M B The column vectors in are arranged in ascending order to obtain the arrival delay vectors corresponding to the diffraction path, the first reflection path and the second reflection path; If x S =1, then there are at least 4 multipaths corresponding to the two-way multipath Based on the arrival delay relationship between the base multipath and the combined multipath, the arrival delay vector corresponding to the secondary reflection path is derived; If x S = 0, it is considered that there are at least 3 multipaths corresponding to Based on the arrival delay relationship between the base multipath and the combined multipath, the arrival delay vectors corresponding to the diffraction and secondary reflection paths are derived; where: It represents the two-way multipath captured by the receiving antenna of the single-transmitter single-receiver radar, and the propagation path index i,j∈{0,1,2}, where propagation path 0 refers to the electromagnetic wave diffracting at the corner of the L-shaped wall and then propagating to the hidden target; propagation path 1 refers to the electromagnetic wave reflecting once on the third wall and then propagating to the hidden target; propagation path 2 refers to the electromagnetic wave reflecting once on the third wall and then propagating to the hidden target by the second reflection on the second wall, and i=j is the two-way multipath. It is called the base multipath, and the two-way multipath of i≠j It is called combined multipath.

5. The method according to claim 4, characterized in that Threshold σ T The magnitude is set to 10 -1 ~10 -2 .

6. The method according to claim 1, characterized in that Step S4 specifically includes: Define R1 as the radar mirror point and C as the corner of the L-shaped wall. Then the hidden target Q is located on the circle with the corner C as the center and ||CQ|| as the radius and on the circle with R1 as the center and ‖R1Q|| as the radius. Among them, the radar mirror point R1 is the mirror symmetric position of the single-transmitter and single-receiver radar about the third wall. The position of the hidden target Q is obtained based on the circle intersection algorithm.

7. The method according to claim 6, characterized in that In step S4, the position of the hidden target Q is specifically: Define the length of line segment CR1 as L and the slope as l CR , the radii of the two circles are r1=||CQ||, r2=||R1Q||; Define E as the vertical point of the hidden target Q on line segment CR1, then the slope of line segment QE is Based on the length of line segment CE||CE||, the position coordinates of corner C (x C ,y C ), radar mirror point R1 position coordinates (x R1 ,y R1 ) Calculate the position coordinates of the vertical point E: in, According to the coordinates of the vertical point E, the length of the line segment QE ||QE|| and the slope l QE Calculate the position coordinates of the hidden target Q:

8. The method according to claim 1, characterized in that In step S5, based on the coupling relationship between the target and the building layout contained in the multipath signal, the shape of the shielding wall is dynamically reconstructed by sensing the moving shielding target.

9. The method according to claim 1, characterized in that Step S5 specifically includes: Define S to represent the secondary reflection point on the shielding wall, which is located on an ellipse with the concealed target Q and the radar mirror point R1 as the focus and the secondary reflection path length (||R1S||+||SQ||) as the major axis; where the radar mirror point R1 is the mirror symmetric position of the single-transmitter and single-receiver radar about the third wall; The contour shape of the shielding wall is reconstructed based on the ellipse envelope generated by the continuously moving shielding target.

10. The method according to claim 9, characterized in that In step S5, the envelope of the ellipse family is discretized into the common tangents of two adjacent ellipses, and the two corresponding tangent points can be approximately equivalent to the position of the secondary reflection point S, and the contour shape of the shielding wall is reconstructed using the ellipse envelope approximation solution method.

Images