Search

CN-121995370-A - Ultra-wideband through-wall radar wall parameter measurement method

CN121995370ACN 121995370 ACN121995370 ACN 121995370ACN-121995370-A

Abstract

The invention discloses a method for measuring ultra-wideband through-wall radar wall parameters, which is characterized by comprising the following steps of S1, data acquisition and input of a hardware part; S2, clutter and noise suppression, S3, data imaging, S4, parameter identification and output. The method optimizes the characteristic enhancement technology and the high-precision recognition technology of the aiming through-wall radar human body activity recognition, obtains parameters more quickly, calculates more accurately, and has better excellent performance.

Inventors

  • ZHANG JING

Assignees

  • 苏州红警科技有限公司

Dates

Publication Date
20260508
Application Date
20231201

Claims (8)

  1. 1. The ultra-wideband through-the-wall radar wall parameter measurement method is characterized by comprising the following steps of: S1, data acquisition and input of a hardware part; S2, clutter and noise suppression; s3, data imaging; S4, parameter identification output.
  2. 2. The method for measuring parameters of an ultra-wideband through-wall radar wall according to claim 1, wherein the step S1 comprises the following steps: S11, determining a coordinate system, namely under the spherical coordinate system, using (R, alpha, beta) to represent the target azimuth, wherein alpha is the azimuth angle of the measuring point from the radar, beta is the pitch angle of the measuring point from the radar, and R is the distance of the measuring point from the radar; S12, calculating the distance between the measuring point and the radar, wherein R is the distance between the radar and the target, C is the speed of light, t r is the time delay of the transmitted pulse and the received pulse, and the distance between the measuring point and the radar is expressed as S13, measuring the target speed by using Doppler effect, namely, the frequency difference between the echo signal frequency of the radar and the main wave transmitting signal frequency exists when the target moves radially, wherein the frequency difference is defined as f d ＝f r -f t ; Where f d is the Doppler frequency, f r is the received echo signal frequency, f t is the transmitted signal frequency, and for the RF signal S t (t): S t (t)＝A×cos(2πf c t+φ); the frequency is expressed as a phase derivative, and there are: Radar typically uses a chirp signal, denoted as: Its frequency varies with time by f t , expressed as: f(t)=f 0 +μt。
  3. 3. The method for measuring parameters of an ultra-wideband through-wall radar wall body according to claim 1, wherein the angle range of alpha is 0-360 degrees, and the angle range of beta is 0-180 degrees.
  4. 4. The method for measuring parameters of an ultra-wideband through-the-wall radar wall according to claim 1, wherein S4 performs parameter identification by using a machine learning method when the parameter identification is output, and the machine learning algorithm comprises SVM, CNN, viT.
  5. 5. The method for measuring parameters of an ultra-wideband through-wall radar wall according to claim 1, wherein S2 comprises the steps of receiving a signal frequency domain, converting the signal frequency domain into a time domain, mixing the local oscillation, low-pass filtering, splicing along slow time, and activating target display during clutter and noise suppression.
  6. 6. The method for measuring the wall parameters of the ultra-wideband through-the-wall radar according to claim 5, wherein the method for measuring the wall parameters of the ultra-wideband through-the-wall radar is characterized by comprising the following steps of firstly carrying out SVD decomposition on an input image by utilizing SVD, dividing the input image into three subspaces which respectively represent a wall subspace, a human motion characteristic subspace and a noise subspace, and secondly sequentially inputting the image into a network of coordinate attention and LISTA, wherein the coordinate attention embeds position information into a channel, and the coordinate attention decomposes the channel attention into two one-dimensional characteristic coding processes, and respectively aggregates characteristics along two spatial directions.
  7. 7. The method for measuring parameters of an ultra-wideband through-wall radar wall according to claim 1, wherein S3 comprises the following steps: The input LISTA network receives as input a corrupted or compressed signal; the sparse representation initialization comprises the steps that a network firstly carries out linear transformation on an input signal once and maps the input signal to a sparse representation space; The iterative process comprises the steps that LISTA the network gradually improves the quality of sparse representation through multiple iterations, wherein each iteration comprises two steps of sparsification and updating; the sparse representation is output for signal reconstruction or other subsequent tasks.
  8. 8. The method for measuring parameters of an ultra-wideband through-wall radar wall body according to claim 6, wherein S3 is characterized in that when data are imaged, images of three subspaces are fused together in a summation mode to determine the weight applied by the image, and an adopted network is constructed by LeNet.

Description

Ultra-wideband through-wall radar wall parameter measurement method Technical Field The invention relates to the technical field of radars, in particular to an ultra-wideband through-wall radar wall parameter measurement method. Background A through-wall radar is a device that detects objects behind walls or other shelters by electromagnetic wave signals. The ultra-wideband radar signal is utilized to have strong penetrability to barriers such as nonmetal medium walls, and radar echoes are generated after the ultra-wideband radar signal penetrates through nonmetal barriers such as the walls and meets personnel targets. The weak motions such as human body motion, respiratory heartbeat, body shaking, limb swinging and the like can form a Doppler effect on the echo, and the received echo is processed and analyzed based on the Doppler effect, so that information such as the position, the number and the gesture of personnel targets behind the wall body is obtained and displayed on a control terminal client software interface. The prior China patent with the publication number of CN109696672B discloses a high-resolution through-wall radar imaging method based on spatial structure relativity, which comprises the following steps of acquiring echo signals of an N-th array element receiving observation scene in an N-dimensional uniform linear array, vectorizing the N array element and M frequency point receiving echo signals, carrying out electromagnetic wave multipath propagation caused by internal wall reflection, acquiring echo signals of the N-dimensional array receiving multipath propagation, carrying out downsampling on measurement signals to acquire sparsely reconstructed observation vectors, adding a set sparse structure simulating the vectors to direct-arrival waves and K-1 multipath reflection coefficient vectors, and carrying out iterative solution on each parameter by utilizing a parameter estimation formula to obtain a high-resolution through-wall radar imaging result. The method fully utilizes the group sparse structure of multipath propagation and the space distribution structure of target continuity in the imaging process of the through-wall radar, and obtains the high-resolution radar image. The above patent has some advantages, but also has some disadvantages of complex parameter measurement and low accuracy. Disclosure of Invention Aiming at the problems in the background art, the invention aims to provide an ultra-wideband through-wall radar wall parameter measurement method for solving the problems in the background art. The technical aim of the invention is realized by the following technical scheme: a ultra-wideband through-wall radar wall parameter measurement method comprises the following steps: S1, data acquisition and input of a hardware part; S2, clutter and noise suppression; s3, data imaging; S4, parameter identification output. Preferably, the step S1 includes the following steps: S11, determining a coordinate system, namely under the spherical coordinate system, using (R, alpha, beta) to represent the target azimuth, wherein alpha is the azimuth angle of the measuring point from the radar, beta is the pitch angle of the measuring point from the radar, and R is the distance of the measuring point from the radar; S12, calculating the distance between the measuring point and the radar, wherein R is the distance between the radar and the target, C is the speed of light, t r is the time delay of the transmitted pulse and the received pulse, and the distance between the measuring point and the radar is expressed as S13, measuring the target speed by using Doppler effect, namely, the frequency difference between the echo signal frequency of the radar and the main wave transmitting signal frequency exists when the target moves radially, wherein the frequency difference is defined as f d＝fr-ft; Where f d is the Doppler frequency, f r is the received echo signal frequency, f t is the transmitted signal frequency, and for the RF signal S t (t): St(t)＝A×cos(2πfct+φ); the quite rate is expressed as the waited derivative, and then there are: Radar typically uses a chirp signal, denoted as: Its frequency varies with time by f t, expressed as: f(t)=f0+μt。 Preferably, the angle of alpha is in the range of 0-360 degrees, and the angle of beta is in the range of 0-180 degrees. Preferably, in the step S4, a machine learning method is adopted to perform parameter identification during parameter identification output, and the machine learning algorithm includes SVM, CNN, viT. Preferably, the S2 process includes the steps of receiving signal frequency domain to time domain, local oscillation mixing, low-pass filtering, splicing along slow time and activating target display during clutter and noise suppression. The method comprises the steps of firstly dividing an input image into three subspaces by SVD, wherein the three subspaces respectively represent a wall subspace, a human motion characteristic subspac