CN117908002B - Non-line-of-sight distance estimation method based on IR-UWB radar - Google Patents
Non-line-of-sight distance estimation method based on IR-UWB radar Download PDFInfo
- Publication number
- CN117908002B CN117908002B CN202410316092.8A CN202410316092A CN117908002B CN 117908002 B CN117908002 B CN 117908002B CN 202410316092 A CN202410316092 A CN 202410316092A CN 117908002 B CN117908002 B CN 117908002B
- Authority
- CN
- China
- Prior art keywords
- distance
- distributed
- frequency shift
- value
- frequency
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 62
- 239000011159 matrix material Substances 0.000 claims abstract description 31
- 230000008859 change Effects 0.000 claims abstract description 18
- 238000000605 extraction Methods 0.000 claims abstract description 17
- 238000012545 processing Methods 0.000 claims abstract description 9
- 238000004364 calculation method Methods 0.000 claims description 7
- 238000001914 filtration Methods 0.000 claims description 4
- 238000010606 normalization Methods 0.000 claims description 3
- 230000008447 perception Effects 0.000 abstract description 26
- 238000001514 detection method Methods 0.000 description 6
- 230000008569 process Effects 0.000 description 6
- 239000000284 extract Substances 0.000 description 5
- 230000000694 effects Effects 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 238000012360 testing method Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000012935 Averaging Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 2
- 230000004888 barrier function Effects 0.000 description 2
- 230000006399 behavior Effects 0.000 description 2
- 230000003247 decreasing effect Effects 0.000 description 2
- 230000001747 exhibiting effect Effects 0.000 description 2
- 230000001965 increasing effect Effects 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 230000003068 static effect Effects 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000001629 suppression Effects 0.000 description 1
Landscapes
- Radar Systems Or Details Thereof (AREA)
Abstract
The invention belongs to the technical field of indoor track tracking, and relates to a non-line-of-sight distance estimation method based on an IR-UWB radar, which comprises the steps of firstly, utilizing the IR-UWB radar to obtain a radio frequency pulse ultra-wideband signal for sensing target movement, and extracting a distributed two-dimensional fast and slow time matrix signal characteristic from the radio frequency pulse ultra-wideband signal; secondly, performing fast Fourier transform on the distributed two-dimensional fast and slow time matrix signal characteristics to obtain distributed Doppler frequency shift characteristic components; then, combining the Doppler distribution unbalance characteristic and Doppler frequency shift change characteristic caused by the motion, carrying out differential operation and binarization processing on the distributed Doppler frequency shift characteristic component, and completing reconstruction; finally, a target distance extraction algorithm is designed for reconstruction Doppler frequency shift, so that the IR-UWB radar can acquire target distance information in a non-line-of-sight scene, and the problem of non-line-of-sight perception failure caused by obstacles in an indoor scene to the IR-UWB radar is solved.
Description
Technical Field
The invention relates to the technical field of indoor track tracking, in particular to a non-line-of-sight distance estimation method based on an IR-UWB radar.
Background
The high-precision position information of the personnel in the indoor environment has very important significance for realizing the application based on the position service, and the accurate pushing of the customized service and the application based on the perception of the position can be realized by acquiring the position information of the personnel.
The current technology for positioning indoor personnel generally divides a perception target into two types according to whether the perception target carries corresponding equipment, namely active positioning and passive positioning. Active positioning generally requires that a sensing body carries a dedicated sensor device and actively transmits signals to communicate with the surrounding environment, thereby enabling the acquisition of sensing body position information. Common active positioning technologies mainly include acquiring position information of a target by using a global positioning system (Global Positioning System, GPS) carried by a mobile phone, or estimating the position of the target by using signals reflected by a main body carrying an RFID tag and a UWB tag. Compared with the active positioning requiring that the target carries special equipment, the passive positioning does not require that the user carries any special equipment, and the non-contact type position estimation is realized through the change of signals caused by the movement of the user. The passive positioning technology can be divided into WiFi, millimeter wave radar, laser radar, pulse radar and different technical schemes based on vision and the like according to the signal category.
In the passive positioning technology, the pulse ultra-wideband IR-UWB is suitable for sensing target behaviors in a home scene due to the advantages of ultra-large bandwidth, lower power consumption, simplified deployment, high-precision distance checking precision and the like. In indoor scenes, continuous motion of a perceived target may cause a pulse signal to change at a point of location on a received signal frame. The most common method at present extracts the Time-of-flight (ToF) of a radio frequency pulse signal, the Angle-of-arrive (AoA) of arrival, the Time-difference-of-arrive (TDoA) of the signal, and the like from the change of the pulse position on the fast-slow Time matrix of the received signal, and combines the propagation speed of the radio signal, so as to accurately acquire the distance signal between the target and the radio frequency terminal, wherein the error is about decimeter level. However, due to uncertainty of arrangement of various furniture objects and the like in an indoor scene, an obstacle exists in a scene that blocks a radio frequency radar terminal from the object during movement of the object, so that a Non-light-of-sight (NLOS) scene is formed. In a non-line-of-sight scene, reflection information of human body motion is submerged by Direct Current (DC) components reflected by obstacles, so that the signal characteristics are invalid, distance information corresponding to a target cannot be extracted, and a distance sensing result of the pulse ultra-wideband radar is distorted.
The application publication number CN108828570A discloses a ranging method and a ranging device based on dynamic estimation of a path loss factor, which are mainly non-line-of-sight ranging methods for WiFi signals, and have a stable testing effect because WiFi is easy to receive the influence of the environment and has larger fluctuation interference ratio; the application publication number CN112731394A discloses bunching SAR clutter suppression and moving target refocusing based on an approximate observation matrix, which is mainly aimed at SAR radar and is difficult to popularize and use in daily life; the IR-UWB radar has the advantages of simple deployment and high detection precision, can be used for target behavior perception in household scenes, and aims at WiFi signals and SAR radars although some non-line-of-sight distance estimation exists in the prior art, and the existing target motion capture method cannot directly extract motion information of objects from the IR-UWB signals, so that a non-line-of-sight distance estimation method based on the IR-UWB radars is needed to realize the distance estimation in the non-line-of-sight scenes.
Disclosure of Invention
The invention aims to provide a non-line-of-sight distance estimation method based on an IR-UWB radar, which realizes the distance information extraction function of the IR-UWB radar in a non-line-of-sight scene by extracting distributed Doppler frequency shift from an IR-UWB radar received signal and reconstructing the distributed Doppler frequency shift and combining a self-designed target distance extraction algorithm; the method is used for solving the problem of non-line-of-sight perception failure caused by obstacles to the IR-UWB radar in an indoor scene.
Although the speed matrix data of the sensing target in the pulse ultra-wideband IR-UWR radar is submerged by the direct current component formed by the obstacle, for the receiving signal, the target movement speed can cause the receiving signal to generate a Doppler frequency shift component with a movement speed corresponding to weak.
The invention is realized by the following technical scheme, and the non-line-of-sight distance estimation method based on the IR-UWB radar specifically comprises the following steps:
Step 1: acquiring a radio frequency pulse ultra-wideband signal corresponding to the motion of a perceived target by using an IR-UWB radio frequency radar in an indoor environment, and extracting corresponding distributed two-dimensional fast and slow time matrix signal characteristics from the signal;
step 2: performing fast Fourier transform (Fast Fourier Transform, FFT) on the signal characteristics of the distributed two-dimensional fast and slow time matrix on the basis of the step 1 to obtain corresponding distributed Doppler frequency shift characteristic components;
Step 3: on the basis of the step 2, combining the Doppler distribution unbalance characteristic and Doppler frequency shift change characteristic caused by movement, carrying out differential operation and binarization processing on the distributed Doppler frequency shift characteristic component, and completing reconstruction of the distributed Doppler frequency shift characteristic component;
Step 4: on the basis of the step 3, a corresponding target distance extraction algorithm is designed for the reconstructed distributed Doppler frequency shift characteristic component, and the IR-UWB radio frequency radar is further used for obtaining the perceived target distance information.
Further, the specific steps of the step1 are as follows:
fixing the single-receiving IR-UWB radio frequency radar on an indoor wall, and reflecting the radio frequency pulse ultra-wideband signal into an indoor environment; a scene with alternate vision distance and non-vision distance appears in the motion process of a perception target; the perception of the target will cause a change in the received ultra wideband signal of the radio frequency pulse, the received signal of the IR-UWB radio frequency radar Expressed as:
Wherein the method comprises the steps of AndReal and imaginary information respectively of the received rf pulse ultra wideband signal,Is an imaginary unit;
From received signals The signal characteristics of the extracted distributed two-dimensional fast and slow time matrixExpressed as:
Wherein the method comprises the steps of Representing the value of the distance bin, t represents the time of the corresponding data frame,Representing the time tThe distance corresponds to the value of the radio frequency pulse ultra wideband signal.
Further, the step 2 specifically comprises:
In the process of indoor perception target positioning, the motion of a perception target can cause the propagation path of a radio frequency pulse ultra-wideband signal to change, so that the receiving and transmitting radio frequency pulse ultra-wideband signal generates Doppler frequency shift, and the signal amplitudes of different distances bin of the fast and slow time matrix signal characteristic data acquired by the IR-UWB radar show differences. The DC component is eliminated by the distributed two-dimensional fast and slow time matrix signal characteristics obtained in the step 1, and the fast Fourier transform FFT is carried out on the distributed two-dimensional fast and slow time matrix signal characteristic data sequences according to different distance bins, so as to obtain the distributed Doppler frequency shift characteristic components The method comprises the following steps:
Wherein the method comprises the steps of Representing the value of the radio frequency pulse ultra-wideband signal corresponding to the t second on the nth distance bin; Represents the starting position for performing the fourier transform, w represents the window length for performing the fast fourier transform, and T represents the transpose. In use, the IR-UWB radio frequency radar has a packet rate of 100 Hz per second and a window length w for performing a fast Fourier transform of 100.
Further, the step 3 specifically comprises:
Step 3-1: distributed doppler shift filtering. For indoor perception, usually the frequency shift component of the stationary object reflection is mainly concentrated in the low frequency range, i.e., -1hz to 1hz, and the noise in the environment is mainly present in the high frequency noise. Therefore, the motion information of a perceived target is reserved by removing low-frequency components between-2 Hz and high-frequency components between < -45Hz and >45Hz from the distributed Doppler characteristic components;
Step 3-2: and carrying out differential operation on the filtered distributed Doppler frequency shift characteristic components. In indoor environments, the distance between the perceived target and the device is typically in a single course of change, increasing or decreasing, and the doppler shift corresponding thereto is also non-negative, i.e., positive, thereby exhibiting an asymmetric characteristic. Based on the asymmetric characteristic of the frequency distribution, the Doppler component of the human motion is amplified by performing a difference operation on the values of the positive and negative frequency domains according to the positive and negative Doppler frequency ranges, and the specific operation is as follows:
Wherein the method comprises the steps of For a specific point of the frequency,The frequency value is represented as a positive number, and the corresponding range is [1 Hz-50 Hz ]; The frequency value is negative and the corresponding range is [ -50Hz to-1 Hz ]. Representing a distributed doppler shift feature component with a positive frequency value,A distributed Doppler shift feature component representing a negative frequency value; a value representing an nth distance bin;
Step 3-3: after the differential operation is carried out on the distributed Doppler frequency shift characteristic components in the step 3-2, the maximum frequency shift components on the corresponding distance bins of different time frames are extracted; the frequency offset maximum values of different distances bin at the current moment are extracted to construct a new distributed Doppler frequency shift characteristic component, and the specific calculation method is as follows:
Wherein the method comprises the steps of For the current moment of time,For a specific point of the frequency,For the differential value of the distributed Doppler shift characteristic componentTime-of-day distanceAt the maximum frequency difference value of the above,A distributed doppler shift feature component representing a positive value of the maximum frequency difference value,A distributed Doppler feature component representing a maximum frequency difference value of negative values;
Step 3-4: on the basis of the new distributed Doppler frequency shift characteristic component obtained in the step 3-3, binarization processing is carried out, firstly, random offset of the new distributed Doppler frequency shift characteristic component is eliminated, when the frequency offset value is larger than a preset threshold value, the random offset is reserved, and when the frequency offset value is smaller than the preset threshold value, the random offset is enabled to be equal to the value of the previous term, and the specific calculation method is as follows:
Wherein the method comprises the steps of Is thatIs used for the frequency offset value of (1),Is a predetermined threshold. Then toAccording to windowTo perform de-averaging to a corresponding data normalization:
Wherein the method comprises the steps of Is a windowMean of the inner sequence. Then combine the final thresholdFor a pair ofAnd (3) performing classification and division:
And finishing the two classifications of the corresponding distributed Doppler frequency shift characteristic components to obtain the reconstructed distributed Doppler frequency shift characteristic components. By reconstructing the signal, new signal characteristics are generated, and the motion information of the perceived target can be displayed simultaneously in the sight distance and non-sight distance scenes, so that the distance information between the target and the radar is extracted.
By reconstructing the original signal, the non-line-of-sight target motion information submerged in the original fast and slow time matrix can be extracted, and the target motion information in the non-line-of-sight state is further displayed; compared with the existing radar signal processing method, the method can extract the motion information of the target under the condition that the target is shielded, thereby realizing non-line-of-sight perception and solving the problem that the existing target motion capture method can not directly extract the motion information of the object from the IR-UWB signal.
Further, the step 4 specifically comprises:
after reconstructing the distributed Doppler frequency shift characteristic component, further extracting distance information of people at different moments on the basis of the characteristic to realize target distance detection in a non-line-of-sight scene; the corresponding specific distance extraction steps are as follows:
Step 4-1: differentiating the reconstructed distributed Doppler frequency shift characteristic components according to the dimension of the distance bin, and primarily acquiring maximum frequency shift variation values of different distances in the space by differentiating the Doppler frequency shifts corresponding to different distance bins at the same moment due to the fact that the motion of a perception target in the space;
Step 4-2: on the basis of obtaining the maximum frequency shift change values of different distances in the step 4-1, eliminating short-distance frequency shift noise of the differential reconstruction distributed Doppler frequency shift characteristic component, wherein human body motion can influence subsequent reflected signals, the space noise in a non-line-of-sight state is relatively short in distance, three conditions of-2, 0 and 2 can appear when the differential reconstruction distributed Doppler frequency shift characteristic component is carried out according to different distance bin, and the positive and negative alternate distances can be different; therefore, calculating the difference value of the distance points corresponding to the positive frequency shift value and the negative frequency shift value, and eliminating when the distance difference is smaller than the short-distance frequency shift noise distance;
step 4-3: setting the positive value of the differential reconstruction distributed Doppler frequency shift characteristic component to 0 on the basis of eliminating short-distance frequency shift noise in the step 4-2; the motion component of the distributed Doppler frequency shift target after binarization is basically negative, so that the interface also appears to be negative after differentiation;
Step 4-4: after the positive value is set to zero in the step 4-3, obtaining target distances of different time points according to a mode of the minimum path distance; firstly, acquiring a target distance point which is not 0 in a1 st time point, and generating a distance record matrix; then selecting a distance bin point which is not 0 in the reconstructed distributed Doppler frequency shift differential component in the dimension of the time frame, and calculating the distance difference value between the reconstructed distributed Doppler frequency shift differential component and each distance bin point which is not 0 at the previous moment; and finally, selecting the distance path with the nearest total path difference as a final IR-UWB radar non-line-of-sight distance estimation result.
The invention has the following beneficial effects: (1) The invention provides an IR-UWB radar non-line-of-sight distance estimation method, which adopts an IR-UWB radar to adopt a radio frequency pulse ultra-wideband signal reflected by a space moving target, extracts a distributed Doppler frequency shift characteristic component from an IR-UWB radar received signal and reconstructs the signal, solves the problem of distance information perception failure of the IR-UWB radar in a non-line-of-sight scene by combining a self-designed target distance extraction algorithm, realizes the target distance detection function of a single IR-UWB radar in the non-line-of-sight scene, and provides a novel perception method and a novel function for the field of IR-UWB radar radio frequency perception.
(2) The invention adopts the IR-UWB radar, can keep high robust perception effect in the perception range, has high detection precision, and can carry out fine-granularity activity perception. However, in the case of IR-UWB, since signal pulses are used, it is difficult to acquire motion information of an object due to the influence of a direct current component in an NLOS scene. According to the method, the Doppler effect of the signal is utilized, the motion information of the target in the non-line-of-sight scene is obtained through the frequency domain characteristics, and further the non-line-of-sight distance information is completed, and meanwhile, the method has higher precision.
Drawings
FIG. 1 is a flow chart of a non-line-of-sight distance estimation method based on an IR-UWB radar in an embodiment of the invention;
Figure 2 is a flow chart of a method of generating a reconstructed distributed doppler shift feature in an embodiment of the present invention;
FIG. 3 is a flow chart of a non-line-of-sight adaptive distance extraction algorithm in an embodiment of the invention;
FIG. 4 is a schematic view of a non-line-of-sight scene;
FIG. 5 is a two-dimensional fast and slow time matrix signal characteristic obtained for an IR-UWB device;
FIG. 6 is an extracted distributed Doppler shift feature component;
FIG. 7 is a feature distribution after reconstruction of a distributed Doppler shift feature component;
FIG. 8 is acquired person movement distance information for a non-line-of-sight scene;
FIG. 9 is a partial contrast plot of the actual trajectory of the perceived target versus the actual acquired trajectory;
FIG. 10 is a graph of test results of perceived target motion in a non-line-of-sight scene;
Fig. 11 is a comparison diagram of the non-line-of-sight estimation result of the method of the present invention and the MCA-CFAR method, in which fig. 11 (a) is an original signal acquired by using an IR-UWB device, fig. 11 (b) is a data feature obtained by processing the original signal in a non-line-of-sight scene by using the MCA-CFAR method, fig. 11 (c) is a target motion extracted from the data feature of fig. 11 (b), and fig. 11 (d) is a target motion extracted from the original signal in the non-line-of-sight scene by using the method of the present invention.
Detailed Description
The technical scheme of the invention is further described in detail below with reference to the attached drawings.
Fig. 1 is a flowchart of a non-line-of-sight distance estimation method based on an IR-UWB radar, and the invention aims to provide a non-line-of-sight distance estimation method based on an IR-UWB radar, which is used for realizing a distance information extraction function of the IR-UWB radar in a non-line-of-sight scene by extracting and reconstructing a distributed doppler shift characteristic component of an IR-UWB radar received signal and combining a self-designed target distance extraction algorithm.
Step 1: acquiring a radio frequency pulse ultra-wideband signal corresponding to the motion of a perceived target by using an IR-UWB radio frequency radar in an indoor environment, and extracting corresponding distributed two-dimensional fast and slow time matrix signal characteristics from the signal;
FIG. 4 is a schematic diagram of a non-line-of-sight scene in which an obstacle exists between a perceived target and an IR-UWB terminal such that no straight line path exists therebetween; during the use, the single-transceiver IR-UWB radio-frequency radar is fixed on the indoor wall and reflects the radio-frequency pulse ultra-wideband signal to the indoor environment. During the operation of the IR-UWB radar, the perception target freely and randomly moves in the perception area, and scenes with alternate vision distance and non-vision distance appear during the movement process. The perception of the target causes a change in the received ultra-wideband signal of the RF pulse, which receives the signal Expressed as:
Wherein the method comprises the steps of AndReal and imaginary information respectively of the received rf pulse ultra wideband signal,Is an imaginary unit;
For IR-UWB radar, the distance between the perceived target and the rf radar terminal may be determined by the pulse signal spacing between different frames, and the amplitude of the IR-UWB radar received signal is formed by a two-dimensional channel impulse response matrix (Channel Impulse Response, CIR), the dimensions of the corresponding matrix being the slow and fast dimensions, respectively. The fast time dimension data may be represented as corresponding distance information, which determines the distance resolution, also called distance bin, from the bandwidth; for the slow time dimension, each received signal frame is formed. FIG. 5 is a two-dimensional fast and slow time matrix signal characteristic obtained for an IR-UWB device, corresponding received signals may be represented as distributed two-dimensional fast and slow time matrix signal characteristics ：
Wherein the method comprises the steps ofRepresenting the value of the distance bin, is typical for commercial IR-UWB devicesIs 1-96.t represents the time of the corresponding data frame,Representing the time tAnd receiving signal values from the corresponding radio frequency pulse ultra wideband.
Step 2: based on the step 1, performing fast fourier transform on the signal characteristics of the distributed two-dimensional fast and slow time matrix to obtain corresponding distributed Doppler frequency shift characteristic components, and fig. 6 is an extracted distributed Doppler frequency shift characteristic component;
In the process of indoor perception target positioning, the motion of a perception target can cause the propagation path of a radio frequency pulse ultra-wideband signal to change, so that the receiving and transmitting radio frequency pulse ultra-wideband signal generates Doppler frequency shift, and the signal amplitudes of different distances bin of the fast and slow time matrix signal characteristic data acquired by the IR-UWB radar show differences. And (3) eliminating the direct current component by obtaining the signal characteristics of the distributed two-dimensional fast and slow time matrix in the step (1), namely subtracting the average value of the time sequence corresponding to each point on the slow time dimension. And performing fast Fourier transform FFT on the distributed two-dimensional fast and slow time matrix signal characteristic data sequences according to different distance bins to obtain distributed Doppler frequency shift characteristic components, namely:
Wherein the method comprises the steps of Representing an IR-UWB received signal corresponding to the t second at the nth distance bin; Representing the starting position for performing the fourier transform, w represents the window length for performing the fast fourier transform. In use, since the rate of the IR-UWB radar is 100 Hz per second, the window length w for performing the fast Fourier transform is also 100.
Step 3: on the basis of the step 2, combining the Doppler distribution unbalance characteristic and Doppler frequency shift change characteristic caused by movement, performing differential operation and binarization processing on the distributed Doppler frequency shift characteristic component, thereby completing the reconstruction of the distributed Doppler frequency shift characteristic component, and as shown in figure 7, reconstructing the characteristic distribution of the distributed Doppler frequency shift characteristic component; the two-dimensional fast and slow time matrix signal characteristics acquired by the IR-UWB device, i.e., the original signal, are shown in fig. 5, and it can be seen that when the target is blocked by the whiteboard, no corresponding target reflected signal is received in the radar signal. After the characteristics are reconstructed by the method, a relatively obvious track change curve is found along with the movement of the target, wherein the track change curve represents the distance relation between the perception target and the IR-UWB equipment. By performing capture analysis on the curve, target motion information in a non-line-of-sight state can be obtained.
Although the distributed doppler shift feature component obtained in step 2 contains motion information of the perceived target, since the dynamic frequency shift component is weaker than the static environment component, it needs to be subjected to feature reconstruction to show the perceived target dynamic signal information to the greatest extent. The calculation process of the reconstructed distributed Doppler frequency shift characteristic is shown in fig. 2, and the specific process is as follows:
Step 3-1: distributed doppler shift filtering. For indoor perception, usually the frequency shift component of the stationary object reflection is mainly concentrated in the low frequency range, i.e., -1hz to 1hz, and the noise in the environment is mainly present in the high frequency noise. Therefore, the motion information of the perceived target is reserved by removing low-frequency components between-2 Hz and high-frequency components between < -45Hz and >45Hz of the distributed Doppler characteristic by adopting a Butterworth band-pass filter, and the output signal of the Butterworth band-pass filter Can be expressed as:
wherein, A signal representative of the input is provided to the processor,Is the cut-off frequency of the filter,The order set in this scheme is 3 for the order of the filter.
Step 3-2: and (3) carrying out differential operation on the filtered distributed Doppler frequency shift characteristic components on the basis of the signal filtering in the step (3-1). In indoor environments, the distance between the perceived target and the device is typically in a single course of change, increasing or decreasing, and the doppler shift corresponding thereto is also non-negative, i.e., positive, thereby exhibiting an asymmetric characteristic. Based on the asymmetric characteristic of the frequency distribution, the Doppler component of the human motion is amplified by performing difference operation on the values of the positive and negative field ranges according to the positive and negative Doppler frequency ranges, and the specific operation is as follows:
Wherein the method comprises the steps of For a specific point of the frequency,The frequency value is represented as a positive number, and the corresponding range is [1 Hz-50 Hz ]; The frequency value is negative and the corresponding range is [ -50Hz to-1 Hz ]. Is the frequency pointCorresponding Doppler shift values; representing a distributed doppler shift feature component with a positive frequency value, A distributed Doppler shift feature component representing a negative frequency value; Representing the value of the nth distance bin.
Step 3-3: and 3-2, extracting maximum frequency shift components on the corresponding distance bin of different time frames after carrying out differential operation on the distributed Doppler frequency shift characteristic components. The new distributed Doppler frequency shift characteristic component is constructed by extracting the frequency offset maximum values of different distances bin at the current moment, and the specific calculation method is as follows:
Wherein the method comprises the steps of For the current moment of time,For the value of the n-th distance bin,For the differential value of the distributed Doppler shift characteristic componentTime-of-day distanceAt the maximum frequency difference value of the above,A distributed doppler shift feature component representing a positive value of the maximum frequency difference value,A distributed Doppler feature component representing a maximum frequency difference value of negative values;
Step 3-4: and (3) performing binarization processing on the basis of the new distributed Doppler frequency shift characteristic component obtained in the step (3-3). For radio frequency pulsed ultra wideband signals, the frequency offset in space is superimposed due to the randomness of the frequency shift of the static background noise, while for moving objects. The random offset of the new distributed Doppler frequency shift characteristic component is eliminated, and when the frequency offset value is larger than a preset threshold value, the random offset is reserved. When the frequency offset value is smaller than a preset threshold value, the frequency offset value is equal to the value of the previous term, and the specific calculation method is as follows:
Wherein the method comprises the steps of Is thatIs used in the field of the present invention,Is a predetermined threshold. Then toAccording to windowTo perform de-averaging to achieve corresponding data normalization:
Wherein the method comprises the steps of Is a windowMean of the inner sequence. Then combine the final thresholdFor a pair ofAnd (3) performing classification and division:
thus, two classifications corresponding to the distributed Doppler frequency shift are completed, and the reconstructed distributed Doppler frequency shift characteristic component is obtained.
Step 4: after reconstructing the distributed Doppler frequency shift characteristic component, further extracting distance information of people at different moments on the basis of the characteristic is needed to realize target distance detection in a non-line-of-sight scene. The corresponding specific distance extraction process is shown in fig. 3, and a corresponding target distance extraction algorithm (Distance Extraction Algorithm algorithm) is designed, wherein the algorithm code relates to the reconstructed distributed doppler shift feature component (rf_dfs) described in the step 3, and the short-distance frequency shift noise distanceIn the scheme, 1m can finish the extraction of the target distance by executing Distance Extraction Algorithm algorithm. The method comprises the following specific steps:
Step 4-1: and differentiating the reconstructed distributed Doppler frequency shift characteristic components according to the distance bin dimension. Since the motion of the sensing target in the space can cause the Doppler frequency shift corresponding to different distance bins at the same time to have the difference, the maximum frequency shift change value of different distances in the space can be obtained preliminarily by differentiating according to the distance bins.
Step 4-2: and (3) eliminating short-range frequency shift noise of the differential reconstruction distributed Doppler frequency shift characteristic component on the basis of acquiring different maximum frequency shift variation values of the distances in the step (4-1). Human motion affects subsequent reflected signals, and spatial noise in non-line-of-sight conditions is relatively short-range. Three cases-2, 0 and 2 occur when the reconstructed distributed doppler shift feature components are differentiated according to different distance bins, and the alternating positive and negative distances are different. Therefore, we perform cancellation when the distance difference is smaller than the short-distance frequency-shift noise distance (average width of human body 1 m) by calculating the distance bin position corresponding to the positive and negative values of the difference and calculating the corresponding distance bin difference.
Step 4-3: setting the positive value of the differential reconstruction distributed Doppler frequency shift characteristic component to 0 on the basis of eliminating short-distance frequency shift noise in the step 4-2; the motion component of the distributed doppler shift target after binarization is substantially negative, and therefore the interface also appears negative after differencing. Thus, by zeroing out the forward differential frequency-shifted component, interference can be irrelevant.
Step 4-4: after zeroing out the positive values in step 4-3, the target distances at different time points are obtained in a minimum path distance manner. First, a target distance point which is not 0 in the 1 st time point is acquired, and a distance record matrix is generated. And then selecting a distance bin point which is not 0 in the reconstructed distributed Doppler frequency shift differential component according to the dimension of the time frame, and calculating the distance difference value between the reconstructed distributed Doppler frequency shift differential component and each distance bin point which is not 0 at the previous moment. Finally, the closest distance path to the total path difference is selected as the final IR-UWB radar non-line-of-sight distance estimation result, as shown in FIG. 8.
Fig. 9 is a partial comparison diagram of a real track of a perceived target and an actually acquired track, ground Trurh (CV) represents a real motion track of the perceived target, IR-UWB 1 is a track of the perceived target acquired through testing by using an IR-UWB radar, and by comparing the real motion track with the track of the actually acquired perceived target, it can be seen from fig. 9 that the present invention realizes distance estimation under non-line-of-sight by reconstructing an original signal and combining a self-designed target distance extraction algorithm to obtain a substantial coincidence between the track of the actually perceived target and the real track.
FIG. 10 is a graph of test results of perceived target motion in a non-line-of-sight scene, the perceived target motion trajectory is rectangular, IR-UWB 1 acquires signals under the barrier of the perceived target, IR-UWB 2 acquires signals under the non-barrier of the perceived target, and the actual trajectory (Ground Trurh (CV) of the perceived target is obtained and the actual trajectory acquired by the method IR-UWB 1 of the invention, wherein the average error between the actual trajectory and the actual trajectory is only 24.88 cm through the record and the distance analysis of the perceived target positioning.
FIG. 11 is a graph comparing the result of non-line-of-sight estimation of the method of the present invention with that of the MCA-CFAR method, and the non-line-of-sight estimation is performed by using an IR-UWB device according to the above-mentioned non-line-of-sight scene, wherein FIG. 11 (a) is an original signal acquired by using the IR-UWB device, FIG. 11 (b) is a data feature obtained by processing the original signal in the non-line-of-sight scene by using the MCA-CFAR method, FIG. 11 (c) is a target motion extracted from the data feature of FIG. 11 (b), and FIG. 11 (d) is a target motion extracted from the original signal in the non-line-of-sight scene by using the method of the present invention, and it can be seen from FIGS. 11 (c) and 11 (d) that the motion track of the perceived target cannot be completely acquired by the MCA-CFAR method in the non-line-of-sight scene, but the method of the present invention can be precisely acquired. The result shows that the method can still perform high-precision distance detection under a non-line-of-sight scene.
The above description is merely of preferred embodiments of the present invention, and the scope of the present invention is not limited to the above embodiments, but all equivalent modifications or variations according to the present disclosure will be within the scope of the claims.
Claims (4)
1. An IR-UWB radar-based non-line-of-sight distance estimation method, comprising the steps of:
Step 1: acquiring a radio frequency pulse ultra-wideband signal corresponding to the motion of a perceived target by using an IR-UWB radio frequency radar in an indoor environment, and extracting corresponding distributed two-dimensional fast and slow time matrix signal characteristics from the signal;
Step 2: performing fast Fourier transform on the signal characteristics of the distributed two-dimensional fast and slow time matrix to obtain corresponding distributed Doppler frequency shift characteristic components;
Step 3: combining the Doppler distribution unbalance characteristic and Doppler frequency shift change characteristic caused by the motion, performing differential operation and binarization processing on the distributed Doppler frequency shift characteristic component, and completing reconstruction of the distributed Doppler frequency shift characteristic component;
Step 3-1: the distributed Doppler frequency shift filtering is used for removing low-frequency components between-2 Hz and high-frequency components between < -45Hz and >45Hz from distributed Doppler characteristic components, and motion information of a perceived target is reserved;
Step 3-2: and carrying out differential operation on the filtered distributed Doppler frequency shift characteristic components, wherein the specific operation is as follows:
Diff_DFS(ti,dn,h)＝|DFS(h+)-DFS(h-)|；
Wherein h is a specific frequency point, h + represents that the frequency value is a positive number, and the range is [1 Hz-50 Hz ]; h - denotes that the frequency value is negative, and the range is [ -50Hz to-1 Hz ]; DFS (h +) represents a distributed doppler shift feature component with a positive frequency value, and DFS (h -) represents a distributed doppler shift feature component with a negative frequency value; d n represents the value of the nth distance bin;
step 3-3: the frequency offset maximum values of different distances bin at the current moment are extracted to construct a new distributed Doppler frequency shift characteristic component, and the specific calculation method is as follows:
where t i is the current time of day, A maximum frequency difference value over a distance d n at time t i, which is a distributed doppler shift characteristic component difference value; /(I)A distributed doppler shift feature component representing a positive value of the maximum frequency difference value,A distributed Doppler feature component representing a maximum frequency difference value of negative values;
Step 3-4: on the basis of the new distributed Doppler frequency shift characteristic component obtained in the step 3-3, binarization processing is carried out, firstly, random offset of the new distributed Doppler frequency shift characteristic component is eliminated, when the frequency offset value is larger than a preset threshold value, the random offset is reserved, and when the frequency offset value is smaller than the preset threshold value, the random offset is enabled to be equal to the value of the previous term, and the specific calculation method is as follows:
where delta is the frequency offset value of RD_DFS '(t i,dn), delta is a predetermined threshold, and then the RD_DFS' (t i,dn) is de-averaged by window ω to achieve the corresponding data normalization:
Where μ is the mean of the sequence within window ω, then the RD_DFS' "(t i,dn) is divided into categories in combination with the final threshold θ:
Completing two classifications of the corresponding distributed Doppler frequency shift characteristic components to obtain a reconstructed distributed Doppler frequency shift characteristic component;
step 4: designing a corresponding target distance extraction algorithm for the reconstructed distributed Doppler frequency shift characteristic component, and obtaining perceived target distance information by using the IR-UWB radio frequency radar under a non-line-of-sight scene by using the corresponding target extraction algorithm;
step 4-1: differentiating the reconstructed distributed Doppler frequency shift characteristic components according to the dimension of the distance bin, and preliminarily obtaining the maximum frequency shift variation values of different distances in the space;
Step 4-2: on the basis of obtaining the maximum frequency shift change values of different distances in the step 4-1, calculating the difference value of the distance points corresponding to the positive frequency shift value and the negative frequency shift value, and eliminating when the distance difference is smaller than the short-distance frequency shift noise distance;
Step 4-3: after the step 4-2 is eliminated, setting the positive value of the differential reconstruction distributed Doppler frequency shift characteristic component to 0;
Step 4-4: after the positive value is set to zero in the step 4-3, obtaining target distances of different time points according to a mode of the minimum path distance; firstly, acquiring a target distance point which is not 0 in a1 st time point, and generating a distance record matrix; then selecting a distance bin point which is not 0 in the reconstructed distributed Doppler frequency shift differential component in the dimension of the time frame, and calculating the distance difference value between the reconstructed distributed Doppler frequency shift differential component and each distance bin point which is not 0 at the previous moment; and selecting the nearest distance path of the total path difference as a final IR-UWB radar non-line-of-sight distance estimation result.
2. The IR-UWB radar-based non-line-of-sight distance estimation method of claim 1, wherein step1 comprises the specific steps of:
Fixing the single-receiving IR-UWB radio frequency radar on an indoor wall, and reflecting the radio frequency pulse ultra-wideband signal into an indoor environment; the received signal r (t) of the IR-UWB radio-frequency radar is expressed as:
r(t)＝I(t)+Q(t)*j；
wherein I (t) and Q (t) are real part and imaginary part information of the received radio frequency pulse ultra wideband signal respectively, and j is an imaginary unit;
The distributed two-dimensional fast and slow time matrix signal characteristics R (t) extracted from the received signal R (t) are expressed as:
Where d represents the value of the distance bin, t represents the time of the corresponding data frame, and r t (d) represents the value of the radio frequency pulse ultra-wideband signal corresponding to the distance d at the time t.
3. The IR-UWB radar-based non-line-of-sight distance estimation method of claim 1, wherein step 2 is specifically:
the direct current component is eliminated by the distributed two-dimensional fast and slow time matrix signal characteristics obtained in the step 1, and the fast Fourier transform FFT is carried out on the distributed two-dimensional fast and slow time matrix signal characteristic data sequences according to different distances bin, so as to obtain a distributed Doppler frequency shift characteristic component DFS (t i,dn), namely:
Wherein the method comprises the steps of Representing the value of the radio frequency pulse ultra-wideband signal corresponding to the t second on the nth distance bin; i represents the starting position for performing the fourier transform, w represents the window length for performing the fast fourier transform, and T represents the transpose.
4. The IR-UWB radar-based non-line-of-sight distance estimation method of claim 3 wherein the IR-UWB radio frequency radar has a packet rate of 100Hz per second and the window length w for performing the fast fourier transform is 100.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410316092.8A CN117908002B (en) | 2024-03-20 | 2024-03-20 | Non-line-of-sight distance estimation method based on IR-UWB radar |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410316092.8A CN117908002B (en) | 2024-03-20 | 2024-03-20 | Non-line-of-sight distance estimation method based on IR-UWB radar |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117908002A CN117908002A (en) | 2024-04-19 |
| CN117908002B true CN117908002B (en) | 2024-05-24 |
Family
ID=90692739
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410316092.8A Active CN117908002B (en) | 2024-03-20 | 2024-03-20 | Non-line-of-sight distance estimation method based on IR-UWB radar |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117908002B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE102004045273A1 (en) * | 2004-09-16 | 2006-03-30 | Deutsches Zentrum für Luft- und Raumfahrt e.V. | Image generating method for synthetic aperture radar, involves producing transfer function and differential rank migration as partly numerical, and giving possible movement of object by measured flight paths using transmitter and sensor |
| CN103454624A (en) * | 2013-09-22 | 2013-12-18 | 河海大学 | Immediate data field moving target detection method based on dimensionality reduction sparse reconstructed space-time spectrum |
| CN110095765A (en) * | 2019-05-21 | 2019-08-06 | 西安电子科技大学 | High-resolution Doppler's reconstructing method of frequency-agile radar |
| CN114325695A (en) * | 2021-11-22 | 2022-04-12 | 中国人民解放军空军工程大学 | Radar target two-dimensional high-resolution imaging method based on rotating antenna |
| CN115552274A (en) * | 2020-05-14 | 2022-12-30 | 株式会社电装 | Position estimation system, position estimation method, and vehicle communication device |
| CN116961680A (en) * | 2022-04-26 | 2023-10-27 | 恩智浦有限公司 | UWB receiver and method of operation |
-
2024
- 2024-03-20 CN CN202410316092.8A patent/CN117908002B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE102004045273A1 (en) * | 2004-09-16 | 2006-03-30 | Deutsches Zentrum für Luft- und Raumfahrt e.V. | Image generating method for synthetic aperture radar, involves producing transfer function and differential rank migration as partly numerical, and giving possible movement of object by measured flight paths using transmitter and sensor |
| CN103454624A (en) * | 2013-09-22 | 2013-12-18 | 河海大学 | Immediate data field moving target detection method based on dimensionality reduction sparse reconstructed space-time spectrum |
| CN110095765A (en) * | 2019-05-21 | 2019-08-06 | 西安电子科技大学 | High-resolution Doppler's reconstructing method of frequency-agile radar |
| CN115552274A (en) * | 2020-05-14 | 2022-12-30 | 株式会社电装 | Position estimation system, position estimation method, and vehicle communication device |
| CN114325695A (en) * | 2021-11-22 | 2022-04-12 | 中国人民解放军空军工程大学 | Radar target two-dimensional high-resolution imaging method based on rotating antenna |
| CN116961680A (en) * | 2022-04-26 | 2023-10-27 | 恩智浦有限公司 | UWB receiver and method of operation |
Non-Patent Citations (1)
| Title |
|---|
| 宫峰勋 ; 马艳秋 ; 车业蒙 ; .场面信道特征重构的非视距多点定位性能研究.电波科学学报.(06),全文. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117908002A (en) | 2024-04-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102058411B (en) | UVB based multi-channel radar life detection instrument | |
| CN102243303B (en) | Method for locating static human body through wall based on respiratory character | |
| Pastina et al. | Parasitic exploitation of Wi-Fi signals for indoor radar surveillance | |
| Xu et al. | Range sidelobe suppression technique for coherent ultra wide-band random noise radar imaging | |
| CN102008291A (en) | Single-channel UWB-based radar type life detection instrument for multi-target detection | |
| Ding et al. | Noncontact multiple targets vital sign detection based on VMD algorithm | |
| CN103558597B (en) | Based on weak target detection method in the sea clutter of spectrum kurtosis | |
| Li et al. | Analysis of characteristics of two close stationary human targets detected by impulse radio UWB radar | |
| CN110095766B (en) | Maneuvering target coherent accumulation detection method based on non-uniform resampling technology | |
| CN102018515B (en) | Method for identifying and one-dimensionally positioning multiple human body targets by using single-channel UWB-based radar life detection instrument | |
| KR20040007534A (en) | System and method for detection and feature extraction in passive coherent location applications | |
| CN114019456A (en) | Micro-nano satellite networking radar system for space target detection | |
| CN110879388A (en) | Non-contact detection distinguishing method for human and animal based on IR-UWB (infrared-ultra wide band) biological radar signals | |
| Sobhani et al. | Multiple target detection and localization in UWB multistatic radars | |
| Mostafa et al. | Singular spectrum analysis-based algorithm for vitality monitoring using M-sequence UWB sensor | |
| CN117908002B (en) | Non-line-of-sight distance estimation method based on IR-UWB radar | |
| Deng et al. | Efficient living subject localization and weak vital-sign signal enhancement using impulse radio based UWB radar | |
| RU2296345C2 (en) | Mode of targets radar station clearance according to distance and a pulse radar station with compression of pulses and restoration of signals | |
| KR101916591B1 (en) | A bio-information determination apparatus and method using principal component analysis of radar signal | |
| Uney et al. | Maximum likelihood signal parameter estimation via track before detect | |
| Tran et al. | Toward passive RF tomography: Signal processing and experimental validation | |
| Hayashi et al. | Iterative data clustering algorithm of Doppler-associated RPM imaging for UWB human body imaging radar | |
| KAKOUCHE et al. | New measurement method for respiration and heartbeat estimation using ir-uwb radar | |
| Kakouche et al. | Respiration Rate Estimation Using IR-UWB Radar Based on the Second Maximum Eigenvalues | |
| KR20200032916A (en) | UWB Radar Receiving Device and Method for Processing Received Signal of UWB Radar |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |