CN116879863B - Multi-target measuring method and system for continuous wave 4D millimeter wave radar - Google Patents
Multi-target measuring method and system for continuous wave 4D millimeter wave radar Download PDFInfo
- Publication number
- CN116879863B CN116879863B CN202311159083.4A CN202311159083A CN116879863B CN 116879863 B CN116879863 B CN 116879863B CN 202311159083 A CN202311159083 A CN 202311159083A CN 116879863 B CN116879863 B CN 116879863B
- Authority
- CN
- China
- Prior art keywords
- target
- echo signal
- point
- probability density
- cluster
- 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 48
- 238000001514 detection method Methods 0.000 claims abstract description 42
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 29
- 238000001914 filtration Methods 0.000 claims abstract description 21
- 230000006870 function Effects 0.000 claims description 95
- 239000011159 matrix material Substances 0.000 claims description 35
- 238000005259 measurement Methods 0.000 claims description 19
- 239000002245 particle Substances 0.000 claims description 19
- 230000003044 adaptive effect Effects 0.000 claims description 6
- 238000004590 computer program Methods 0.000 claims description 5
- 238000000691 measurement method Methods 0.000 claims description 4
- 238000005516 engineering process Methods 0.000 abstract description 5
- 238000005070 sampling Methods 0.000 description 4
- 238000012952 Resampling Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000002776 aggregation Effects 0.000 description 1
- 238000004220 aggregation Methods 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000007621 cluster analysis Methods 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000002592 echocardiography Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000007620 mathematical function Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 239000000758 substrate Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
- G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
- G01S7/41—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The application provides a continuous wave 4D millimeter wave radar multi-target measuring method and a system, which relate to radar detection technology, in particular to a continuous wave 4D millimeter wave radar multi-target measuring technology, wherein the method comprises the following steps: determining a first probability density function when the original echo signal meets a first assumption condition and a second probability density function when the original echo signal meets a second assumption condition based on the original echo signal corresponding to the millimeter wave radar, and filtering interference signals in the original echo signal according to a preset detection threshold function by combining the first probability density function and the second probability density function to obtain a target echo signal; obtaining a clustering center of the point cloud set based on the point cloud set corresponding to the target echo signal, and forming a clustering cluster according to the clustering center and points in the point cloud set; and acquiring a point cluster centroid according to the cluster, the measuring point data and the track which are acquired in advance, and tracking a target object corresponding to the original echo signal through a multi-target tracking algorithm based on the point cluster centroid.
Description
Technical Field
The application relates to a radar detection technology, in particular to a continuous wave 4D millimeter wave radar multi-target measurement technology.
Background
Millimeter wave radar is a sensor technology widely applied to multi-target detection, has high resolution, can penetrate the bad weather conditions and obstacles, and is widely applied to the fields of military, civilian use and automatic driving.
Compared with a general radar, a linear frequency modulation continuous wave (LFM-CW) radar has a larger time-band product, so that the radar has higher speed resolution and distance resolution, and meanwhile, the radar has the advantages of simple hardware design, small volume, strong reliability and the like, is particularly suitable for the fields of radar imaging, target feature research and the like, but the accuracy of a measured target is reduced due to the problem of distance speed coupling when the LFM-CW radar detects a moving target, and distance speed decoupling can be realized by adopting a single-period symmetrical triangular wave signal, and the accuracy of pairing of upper and lower sweep targets determines the measuring accuracy of the radar to the speed and the distance in a multi-target moving scene, so that the radar is a main difficulty existing at present.
In the related art, CN105549012a discloses a multi-target detection device of a vehicle-mounted millimeter wave radar system, which is used for solving the problems of inaccurate target detection and missed target detection existing in the multi-target detection of the vehicle-mounted millimeter wave radar, and the technical points are as follows: comprising the following steps: a combined waveform transmitting unit which transmits a combined waveform of the periodic FMCW and CW; the relative velocity matrix calculating unit is used for obtaining a Doppler frequency matrix of the CW waveform echo signal by using a frequency clustering algorithm and calculating a relative velocity matrix; and the resolving unit also uses a frequency aggregation algorithm in the FMCW up and down frequency sweep to obtain frequency values of the up and down frequency sweep, and resolves a speed matrix and a distance matrix of the target.
CN114609626a discloses a multi-target detection method suitable for the field of vehicle millimeter wave radar target detection, which aims at solving the problems of multi-target azimuth measurement limitation and multi-target angle parameter matching under the condition of insufficient number of radar receiving array elements, and the main technical scheme comprises: acquiring millimeter wave radar A/D sampling data; rearranging the multi-array element A/D sampling data into three-dimensional matrix data; performing two-dimensional FFT processing on the A/D sampling data in the distance dimension and the speed dimension, and storing a distance dimension FFT result; obtaining target two-dimensional matrix data by using GO-CFAR; performing DBSCAN density clustering on the detected two-dimensional motion parameters of the target points, and taking the maximum peak point in each class set after clustering as a detection target corresponding to the set to acquire speed distance information of a plurality of targets; extracting a multi-array element beat signal corresponding to each target from a distance dimension FFT result; and (3) performing MUSIC algorithm estimation on the beat signal corresponding to each target, estimating azimuth angle corresponding to the target, and outputting the azimuth angle and the target speed distance information in a matching way.
In summary, in the related art, the target may be detected by changing the radar emission waveform, and the problems of prediction and accurate distinction of the motion state of the target, false alarm and the like are not achieved, but in the practical application process, the accuracy of the target distinction and the estimated position of the target are very important, and the occurrence of the false alarm is unacceptable.
The information disclosed in the background section of the application is only for enhancement of understanding of the general background of the application and should not be taken as an acknowledgement or any form of suggestion that this information forms the prior art already known to a person skilled in the art.
Disclosure of Invention
The embodiment of the application provides a continuous wave 4D millimeter wave radar multi-target measuring method and system, which can track and predict targets in a complex scene with multiple targets.
In a first aspect of the embodiment of the present application, a continuous wave 4D millimeter wave radar multi-target measurement method is provided, including:
determining a first probability density function when the original echo signal meets a first assumption condition and a second probability density function when the original echo signal meets a second assumption condition based on an original echo signal corresponding to the millimeter wave radar, and filtering interference signals in the original echo signal according to a preset detection threshold function by combining the first probability density function and the second probability density function to obtain a target echo signal;
based on a point cloud set corresponding to the target echo signal, obtaining a cluster center of the point cloud set by maximizing intra-cluster similarity of the same type of clusters in the point cloud set and minimizing similarity of different types of clusters, and forming a cluster according to the cluster center and the points in the point cloud set;
and according to the cluster and the measuring point data and the track which are acquired in advance, acquiring a point cluster centroid, and tracking a target object corresponding to the original echo signal through a multi-target tracking algorithm based on the point cluster centroid, wherein the multi-target tracking algorithm is formed based on a Kalman filtering algorithm and particle filtering.
According to some embodiments of the present application, the determining, based on an original echo signal corresponding to a millimeter wave radar, a first probability density function when the original echo signal meets a first assumption condition, and a second probability density function when the original echo signal meets a second assumption condition, combining the first probability density function and the second probability density function, filtering an interference signal in the original echo signal according to a preset detection threshold function, and obtaining a target echo signal includes:
the first assumption condition is that the original echo signal only contains an interference signal, and the second assumption condition is that the original echo signal simultaneously contains a target echo signal and an interference signal;
and enabling the first probability density function to be in a preset threshold range through the detection threshold function, enabling the second probability density function to be maximum, filtering signals meeting a first assumption condition from the original echo signals, and enabling the rest signals to be the target echo signals.
According to some embodiments of the application, the original echo signal comprises:
；
wherein the method comprises the steps ofRepresenting the original echo signal(s),as a first assumption of the condition of the first hypothesis,as a second assumption of the condition of the assumption,in order to provide a signal for the interference signal,is the target echo signal;
the first probability density function and the second probability density function are:
;
;
wherein the method comprises the steps ofAs a function of the first probability density,as a function of the second probability density,is the standard deviation of the two-dimensional image,is the mean value, the interference signalObeying the mean value was 0, and the variance wasX is the original echo signal;
and enabling the first probability density function to be in a preset threshold range through the detection threshold function, and enabling the second probability density function to be maximum, wherein the following objective function is required to be minimum:
；
wherein the method comprises the steps ofRepresenting the lagrangian multiplier and,representation ofAndand (5) weighted sum.
According to some embodiments of the present application, based on a point cloud set corresponding to the target echo signal, obtaining a cluster center of the point cloud set by maximizing intra-cluster similarity of the same class of clusters in the point cloud set and minimizing similarity of different classes of clusters, and forming a cluster according to the cluster center and points in the point cloud set includes: determining an adjacent matrix corresponding to any two vertexes through the initial connectivity of the two vertexes in the point cloud collection;
determining K-order adjacency connectivity and K-order self connectivity corresponding to any two vertexes based on the adjacency matrix, and obtaining a clustering center through a comparison formula;
and distributing each vertex in the point cloud set to a corresponding clustering center to form a clustering cluster.
According to some embodiments of the application, the initial connectivity of any two vertices in the point cloud set may be expressed as:
；
wherein,is the vertexAnd a vertexIs used for the communication degree of the (a),andfor two different vertices of the model,is the variance;
the clustering center is obtained through a comparison formula, and the comparison formula specifically comprises:
；
wherein the method comprises the steps ofIs the vertexAndthe degree of K-th order of adjacency connectivity,is the vertexFor each K, a clustering center can be obtained by searching points meeting the formula;
and distributing each vertex in the point cloud collection to a corresponding clustering center, wherein the method specifically comprises the following steps:
；
wherein the method comprises the steps ofFor other vertices in the point cloud collection,to evaluate the function.
According to some embodiments of the present application, the obtaining a point cluster centroid according to the cluster and the pre-obtained measurement point data and track, and tracking a target object corresponding to the original echo signal through a multi-target tracking algorithm based on the point cluster centroid includes:
calculating the distance between each measuring point and an initial target, obtaining the association degree between each measuring point and the initial target, judging whether each measuring point is in a preset self-adaptive threshold, marking the measuring points in the self-adaptive threshold as effective points, and deleting the rest measuring points;
determining a cost function of an effective point track and an initial target, mutually pairing all measuring points with all tracks by minimizing the cost function, regarding all measuring points associated with the same track as a target point cluster, and acquiring a point cluster centroid of the target point cluster;
and determining a weight value of each track through a multi-target tracking algorithm based on the point cluster centroid, and determining an estimated state of a target object corresponding to the original echo signal based on the weight value and the point cluster centroid.
According to some embodiments of the application, the preset adaptive threshold comprises:
；
wherein target is atThe predicted position of the moment is the center of the wave gate,is aimed atThe group residual covariance matrix of the time instant,for the predicted state of the track at time n,representing the conversion of the predicted state in the cartesian coordinate system to a predicted value in the polar coordinate system,representation ofIs used to determine the transposed matrix of (a),is on trackA covariance matrix of the time instant predictors,a covariance matrix representing the measurement noise is presented,is a group discrete matrix;
will beSubstituting the self-adaptive threshold G of the current time target into the following formula:
；
wherein,for the target volume whenWhen the target group residual covariance matrix is constant, the adaptive threshold G is the target group residual covariance matrixA non-fixed value of influence;
the distance between each measuring point and the initial target is calculated, and the association degree between each measuring point and the initial target is obtained, specifically:
；
wherein the method comprises the steps ofRepresenting measurementsThe distance between the two adjacent substrates is determined,representing the center of the wave gate, l representing the measuring point,covariance of the space distance between the center of the wave gate and the measuring point is represented;
the cost function for determining the effective point trace and the initial target comprises the following steps:
；
wherein D is the cost function of the effective trace and the initial target.
A second aspect of an embodiment of the present disclosure provides a continuous wave 4D millimeter wave radar-based multi-target measurement system, including:
a first unit: the method comprises the steps of determining a first probability density function when the original echo signal meets a first assumption condition and a second probability density function when the original echo signal meets a second assumption condition based on an original echo signal corresponding to a millimeter wave radar, combining the first probability density function and the second probability density function, and filtering interference signals in the original echo signal according to a preset detection threshold function to obtain a target echo signal;
a second unit: the method comprises the steps of obtaining a clustering center of a point cloud set by maximizing intra-cluster similarity of the same type of clusters in the point cloud set and minimizing similarity of different types of clusters based on the point cloud set corresponding to the target echo signal, and forming a cluster according to the clustering center and the points in the point cloud set;
a third unit: the method is used for acquiring the point cluster centroid according to the cluster and the measuring point data and the track which are acquired in advance, and tracking the target object corresponding to the original echo signal through a multi-target tracking algorithm based on the point cluster centroid, wherein the multi-target tracking algorithm is formed based on a Kalman filtering algorithm and particle filtering.
In a third aspect of the embodiments of the present disclosure,
there is provided an electronic device including:
a processor;
a memory for storing processor-executable instructions;
wherein the processor is configured to invoke the instructions stored in the memory to perform the method described previously.
In a fourth aspect of an embodiment of the present application,
there is provided a computer readable storage medium having stored thereon computer program instructions which, when executed by a processor, implement the method as described above.
The beneficial effects of the embodiments of the present application may refer to the effects corresponding to technical features in the specific embodiments, and are not described herein.
Drawings
In order to more clearly illustrate the embodiments of the application or the solutions of the prior art, the drawings which are necessary for the description of the embodiments or the prior art will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the application, and that other embodiments may be obtained from these drawings without inventive effort to a person skilled in the art,
fig. 1 exemplarily shows a flow diagram of a continuous wave 4D millimeter wave radar multi-target measurement method according to an embodiment of the present disclosure;
fig. 2 exemplarily shows a schematic structural diagram of a continuous wave 4D millimeter wave radar multi-target measurement system according to an embodiment of the present disclosure.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present application more apparent, the technical solutions of the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application, and it is apparent that the described embodiments are only some embodiments of the present application, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.
The technical scheme of the application is described in detail below by specific examples. The following embodiments may be combined with each other, and some embodiments may not be repeated for the same or similar concepts or processes.
Fig. 1 is a schematic flow chart of a continuous wave 4D millimeter wave radar multi-target measurement method according to an embodiment of the disclosure, as shown in fig. 1, the method includes:
s1, determining a first probability density function when the original echo signal meets a first assumption condition and a second probability density function when the original echo signal meets a second assumption condition based on an original echo signal corresponding to a millimeter wave radar, combining the first probability density function and the second probability density function, and filtering interference signals in the original echo signal according to a preset detection threshold function to obtain a target echo signal;
the method comprises the steps of acquiring original echo signals through radar or sensor measurement data, analyzing the original echo signal data, estimating a probability density function (first probability density function) meeting a first assumption condition and a probability density function (second probability density function) meeting a second assumption condition, designing a detection threshold function, keeping the value of the first probability density function within a preset threshold value, enabling the value of the second probability density function to be maximum, applying the detection threshold function to the original echo signal data, filtering signals under the first assumption condition, and taking the rest echo signals as target echo signals.
The "detection threshold function" is typically used to determine whether a received signal exceeds a particular threshold, and in the presence of noise, to determine whether a signal is strong enough to be properly detected and identified; combining a preset threshold value according to the received signal strength, and if the signal strength exceeds the threshold value, considering the signal as effective; if the signal strength is below the threshold value, the signal may be considered noise.
The first assumption condition is that the original echo signal only contains an interference signal, and the second assumption condition is that the original echo signal simultaneously contains a target echo signal and an interference signal;
for example, since radar target detection is always accompanied by interference of environmental clutter such as thermal noise inside the radar and ground objects, rain and snow, these interference signals are also inevitably contained in the target echo signal. When the radar detects a non-target signal as including a real target, it is called generating a false alarm, and when the radar determines an echo including a real target as a non-target signal, it is called generating a false alarm. The occurrence of false alarms is difficult to avoid, but the target detection probability of the radar can be maximized under the condition that the false alarm probability is kept at a reasonable level.
In an alternative embodiment of the present application,
and enabling the first probability density function to be in a preset threshold range through the detection threshold function, enabling the second probability density function to be maximum, filtering signals meeting a first assumption condition from the original echo signals, and enabling the rest signals to be the target echo signals.
The original echo signal received by the radar is marked as X, the condition that the original echo signal only contains the interference signal is used as a first assumption condition and is marked asTaking the condition that the original echo signal simultaneously contains the target echo signal and the interference signal as a second assumption condition, marking asThe original echo signal is expressed as:
；
wherein the method comprises the steps ofRepresenting the original echo signal(s),as a first assumption of the condition of the first hypothesis,as a second assumption of the condition of the assumption,in order to provide a signal for the interference signal,is the target echo signal;
the first probability density function and the second probability density function are:
;
;
wherein the method comprises the steps ofAs a function of the first probability density,as a function of the second probability density,is the standard deviation of the two-dimensional image,is the mean value, the interference signalObeying the mean value was 0, and the variance wasIs a gaussian distribution of (c);
in radar target detection we typically useAndas two important performance indicators for CFAR detection, and hopeSmaller and betterThe larger the better. In order to increase the probability of target detection as much as possible, the detection threshold needs to be reduced as much as possible whenWhen the detection threshold reaches the maximum, the report missing rate reaches the minimum, but the reduction of the detection threshold inevitably leads toIncrease, so according to the preset detection threshold, willIs limited within a certain threshold value to enableAt maximum, even though the objective function of the following formula is at minimum:
；
wherein the method comprises the steps ofRepresenting the lagrangian multiplier and,representation ofAndand (5) weighted sum.
The method effectively filters interference signals in the original echo, extracts the echo signals of the target, is beneficial to improving the precision of target detection and tracking, reduces the false alarm rate and improves the performance of the system.
S2, based on a point cloud set corresponding to the target echo signal, obtaining a clustering center of the point cloud set by maximizing intra-cluster similarity of the same type of clusters in the point cloud set and minimizing similarity of different types of clusters, and forming a clustering cluster according to the clustering center and points in the point cloud set;
illustratively, a point is randomly selected as the initial cluster center, and then the similarity of each point to this center is calculated. Points having a similarity to the center greater than a certain threshold are assigned to the same class of clusters. For each point allocated to the cluster, calculating the similarity between the point and the center of the cluster, and taking the point with the highest similarity between the point and the center as a new center. This step is repeated until the new cluster center is no longer changed or changes little. The similarity is determined by calculating the distance between different clusters, and if the distance between two clusters is smaller than a certain threshold value, the clusters are combined into the same cluster.
The processed radar echo data still exists in the form of a point cloud set, but the number of target point clouds detected by the radar can be influenced due to the difference of different targets in terms of volume, shape, reflection characteristics, distance between the radar and the target point cloud. Meanwhile, if the distance between a plurality of targets is small, the point cloud may appear as a large cluster, thereby affecting the detection effect of the radar.
In an alternative embodiment of the present application,
determining an adjacent matrix corresponding to any two vertexes through the initial connectivity of the two vertexes in the point cloud collection;
determining K-order adjacency connectivity and K-order self connectivity corresponding to any two vertexes based on the adjacency matrix, and obtaining a clustering center through a comparison formula;
and distributing each vertex in the point cloud set to a corresponding clustering center to form a clustering cluster.
Wherein, the initial connectivity of any two vertices in the point cloud set can be expressed as:
；
wherein,is the vertexAnd a vertexIs used for the communication degree of the (a),andfor two different vertices of the model,is the variance;
the clustering center is obtained through a comparison formula, and the comparison formula specifically comprises:
；
wherein the method comprises the steps ofIs the vertexAndthe degree of K-th order of adjacency connectivity,is the vertexFor each K, a clustering center can be obtained by searching points meeting the formula;
and distributing each vertex in the point cloud collection to a corresponding clustering center, wherein the method specifically comprises the following steps:
；
wherein the method comprises the steps ofFor other vertices in the point cloud collection, i.e. vertices to be assigned,for evaluating the function, the method is used for measuring the correlation between the vertexes to be allocated and the clustering center,is a mathematical function that represents taking a parameter that maximizes the value of the function, here for selecting the cluster center that is most relevant to the vertex.
The purpose of this expression is to select the most appropriate cluster center based on the correlation between each vertex in the point cloud and the respective cluster center, thereby assigning each vertex to the corresponding cluster center.
After each vertex is distributed to the corresponding clustering center, a clustering cluster in the point cloud collection can be obtained, and whether different echoes are formed by the reflection of the same detection target can be identified in radar measurement.
In the radar detection field, parameters of a detected target need to be considered in many aspects, the traditional radar usually only recognizes the target according to the intensity and position information of a reflected signal, but is easy to be interfered by noise and clutter in a complex environment, so that the target recognition accuracy is reduced. Subsequent target recognition and analysis may be made simpler.
S3, acquiring a point cluster centroid according to the cluster and the measurement point data and the track which are acquired in advance, and tracking a target object corresponding to the original echo signal through a multi-target tracking algorithm based on the point cluster centroid, wherein the multi-target tracking algorithm is formed based on a Kalman filtering algorithm and particle filtering.
In an alternative embodiment of the present application,
calculating the distance between each measuring point and an initial target, obtaining the association degree between each measuring point and the initial target, judging whether each measuring point is in a preset self-adaptive threshold, marking the measuring points in the self-adaptive threshold as effective points, and deleting the rest measuring points;
determining a cost function of an effective point track and an initial target, mutually pairing all measuring points with all tracks by minimizing the cost function, regarding all measuring points associated with the same track as a target point cluster, and acquiring a point cluster centroid of the target point cluster;
and determining a weight value of each track through a multi-target tracking algorithm based on the point cluster centroid, and determining an estimated state of a target object corresponding to the original echo signal based on the weight value and the point cluster centroid.
Wherein "resampling" is a technique commonly used in data analysis and machine learning by varying the sampling frequency of the data set or reorganizing various tasks such as data processing sample imbalance, estimating model performance, estimating uncertainty of parameters, etc.;
"point cluster centroid" refers to the center point or average point of each cluster (also referred to as a cluster) in a cluster analysis or data clustering task. The cluster centroid is typically calculated from the average of the coordinates of all the points in the cluster.
The preset adaptive threshold is expressed as:
；
wherein target is atThe predicted position of the moment is the center of the wave gate,is aimed atThe group residual covariance matrix of the time instant,for the predicted state of the track at time n,representing the conversion of the predicted state in the cartesian coordinate system to a predicted value in the polar coordinate system,representation ofIs used to determine the transposed matrix of (a),is on trackA covariance matrix of the time instant predictors,a covariance matrix representing the measurement noise is presented,is a group discrete matrix;
will beSubstituting the self-adaptive threshold G of the current time target into the following formula:
；
wherein,for the target volume whenWhen the target group residual covariance matrix is constant, the adaptive threshold G is the target group residual covariance matrixA non-fixed value of influence;
the distance between each measuring point and the initial target is calculated, and the association degree between each measuring point and the initial target is obtained, specifically:
；
wherein the method comprises the steps ofThe measurement distance is indicated as being the distance to be measured,representing the center of the wave gate, l representing the measuring point,covariance of the space distance between the center of the wave gate and the measuring point is represented;
the cost function for determining the effective point trace and the initial target comprises the following steps:
；
wherein D is the cost function of the effective trace and the initial target.
Determining a weight value of each track through a multi-target tracking algorithm based on the point cluster centroid, and determining an estimated state of a target object corresponding to the original echo signal based on the weight value and the point cluster centroid specifically comprises the following steps:
randomly generating a set of particles, each particle representing a possible target state; using the predicted state of each particle as prediction data, and performing state prediction and covariance prediction on the particle;
for each predicted particle, calculating its weight from the observed data from the target detector or sensor, the weight representing the degree of fit of the particle to the observed data, a measurement model is used in the present application to calculate the weight taking into account the difference between the target position and the sensor measurement;
for each particle, correcting the state estimation and covariance estimation according to the weight and the corresponding observed data thereof by using a measurement updating step of a multi-target tracking algorithm;
resampling the particles according to the weight of the particles, increasing the selected probability of the particles with high weight, and reducing the selected probability of the particles with low weight, so that the particles with high weight are increased, and the particles with low weight are reduced;
and counting the resampled particle set to obtain a more accurate target estimation state.
The traditional radar detection mode is generally simpler and more direct, but has certain limitation in aspects of multi-target tracking, target recognition and the like, is not supported by a multi-target tracking algorithm, can possibly cause confusion and tracking loss of targets, is generally difficult to distinguish different types of targets by the traditional radar, and can continuously update and correct the estimated state of the targets at different time nodes according to observation data from a radar system after the multi-target tracking algorithm is used, so that the accuracy and stability of detection are maintained, and meanwhile, the method can be suitable for more and more complex scenes.
A second aspect of an embodiment of the present disclosure provides a continuous wave 4D millimeter wave radar-based multi-target measurement system, including:
a first unit: the method comprises the steps of determining a first probability density function when the original echo signal meets a first assumption condition and a second probability density function when the original echo signal meets a second assumption condition based on an original echo signal corresponding to a millimeter wave radar, combining the first probability density function and the second probability density function, and filtering interference signals in the original echo signal according to a preset detection threshold function to obtain a target echo signal;
a second unit: the method comprises the steps of obtaining a clustering center of a point cloud set by maximizing intra-cluster similarity of the same type of clusters in the point cloud set and minimizing similarity of different types of clusters based on the point cloud set corresponding to the target echo signal, and forming a cluster according to the clustering center and the points in the point cloud set;
a third unit: the method is used for acquiring the point cluster centroid according to the cluster and the measuring point data and the track which are acquired in advance, and tracking the target object corresponding to the original echo signal through a multi-target tracking algorithm based on the point cluster centroid, wherein the multi-target tracking algorithm is formed based on a Kalman filtering algorithm and particle filtering.
In a third aspect of the embodiments of the present disclosure,
there is provided an electronic device including:
a processor;
a memory for storing processor-executable instructions;
wherein the processor is configured to invoke the instructions stored in the memory to perform the method described previously.
In a fourth aspect of embodiments of the present disclosure,
there is provided a computer readable storage medium having stored thereon computer program instructions which, when executed by a processor, implement the method as described above.
The present application may be a method, apparatus, system, and/or computer program product. The computer program product may include a computer readable storage medium having computer readable program instructions embodied thereon for performing various aspects of the present application.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present application, and not for limiting the same; although the application has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some or all of the technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit of the application.
Claims (7)
1. The continuous wave 4D millimeter wave radar multi-target measurement method is characterized by comprising the following steps of:
determining a first probability density function when the original echo signal meets a first assumption condition and a second probability density function when the original echo signal meets a second assumption condition based on an original echo signal corresponding to the millimeter wave radar, and filtering interference signals in the original echo signal according to a preset detection threshold function by combining the first probability density function and the second probability density function to obtain a target echo signal;
the first assumption condition is that the original echo signal only contains an interference signal, and the second assumption condition is that the original echo signal simultaneously contains a target echo signal and an interference signal;
the first probability density function is in a preset threshold range through the detection threshold function, the second probability density function is maximized, signals meeting a first assumption condition are filtered out of the original echo signals, and the rest signals are the target echo signals;
based on a point cloud set corresponding to the target echo signal, obtaining a cluster center of the point cloud set by maximizing intra-cluster similarity of the same type of clusters in the point cloud set and minimizing similarity of different types of clusters, and forming a cluster according to the cluster center and the points in the point cloud set;
determining an adjacent matrix corresponding to any two vertexes through the initial connectivity of the two vertexes in the point cloud collection;
determining K-order adjacency connectivity and K-order self connectivity corresponding to any two vertexes based on the adjacency matrix, and obtaining a clustering center through a comparison formula;
distributing each vertex in the point cloud set to a corresponding clustering center to form a clustering cluster;
acquiring a point cluster centroid according to the cluster and measurement point data and tracks acquired in advance, and tracking a target object corresponding to the original echo signal through a multi-target tracking algorithm based on the point cluster centroid, wherein the multi-target tracking algorithm is formed based on a Kalman filtering algorithm and particle filtering;
calculating the distance between each measuring point and an initial target, obtaining the association degree between each measuring point and the initial target, judging whether each measuring point is in a preset self-adaptive threshold, marking the measuring points in the self-adaptive threshold as effective points, and deleting the rest measuring points;
determining a cost function of an effective point track and an initial target, mutually pairing all measuring points with all tracks by minimizing the cost function, regarding all measuring points associated with the same track as a target point cluster, and acquiring a point cluster centroid of the target point cluster;
and determining a weight value of each track through a multi-target tracking algorithm based on the point cluster centroid, and determining an estimated state of a target object corresponding to the original echo signal based on the weight value and the point cluster centroid.
2. The method of claim 1, wherein the original echo signal comprises:
wherein X represents the original echo signal, H 0 For the first assumption, H 1 N (t) is the interference signal, s (t) is the target echo signal;
the first probability density function and the second probability density function are:
wherein P is f As a first probability density function, P m Sigma as a second probability density function n The standard deviation, u is the mean value, the interference signal n (t) obeys Gaussian distribution with the mean value of 0 and the variance of 2 n;
and enabling the first probability density function to be in a preset threshold range through the detection threshold function, and enabling the second probability density function to be maximum, wherein the following objective function is required to be minimum:
P e ＝P m +λP f ＝P(H 0 |H 1 )+λP(H 1 |H 0 )；
wherein λ represents the Lagrangian multiplier, P e Representing P m And P f And (5) weighted sum.
3. The method of claim 1, wherein the initial connectivity of any two vertices in the point cloud collection is represented as:
S ij ＝exp(-||V i -V j ||/σ 2 )；
wherein S is ij Is the vertex V i And vertex V j Degree of connectivity, V i And V j For two different vertices, sigma 2 Is the variance;
the clustering center is obtained through a comparison formula, and the comparison formula specifically comprises:
con (k) (V i ,V i )>con (k) (V i ,V j ),j＝1....n(j≠n)；
wherein con (k) (V i ,V j ) Is the vertex V i And V j K-th order adjacency connectivity, con (k) (V i ,V i ) Is the vertex V i For each K, a clustering center can be obtained by searching points meeting the formula;
and distributing each vertex in the point cloud collection to a corresponding clustering center, wherein the method specifically comprises the following steps:
wherein the method comprises the steps ofRcon for other vertices in the point cloud collection (k) To evaluate the function.
4. The method of claim 1, wherein the preset adaptive threshold comprises:
C G ＝H(X apr (n)P apr (n)H T (X apr (n)))+R G +C D ；
wherein, the predicted position of the target at the time n is taken as the center of the wave gate, C G For the group residual covariance matrix of the target at time n-1, X apr (n) is the predicted state of the track at time n, H (X) apr (n) represents converting the predicted state in the Cartesian coordinate system into the predicted value in the polar coordinate system, H T (X apr (n)) represents H (X) apr Transposed matrix of (n), P apr (n) is covariance matrix of predicted value of track at n time, R G Covariance matrix representing measurement noise, C D Is a group discrete matrix;
c is C G Substituting the self-adaptive threshold G of the current time target into the following formula:
wherein V is the target volume, and when V is constant, the adaptive threshold G is the target group residual covariance matrix C G A non-fixed value of influence;
the distance between each measuring point and the initial target is calculated, and the association degree between each measuring point and the initial target is obtained, specifically:
Z K ＝(C-l) T ST -1 (C-l)；
wherein Z is K The measuring distance is represented by C, the center of the wave gate is represented by l, the measuring point is represented by l, and the covariance of the space distance between the center of the wave gate and the measuring point is represented by ST;
the cost function for determining the effective point trace and the initial target comprises the following steps:
D＝ln|C G |+(Z K ) 2 ；
wherein D is the cost function of the effective trace and the initial target.
5. A continuous wave 4D millimeter wave radar-based multi-target measurement system, comprising:
a first unit: the method comprises the steps of determining a first probability density function when the original echo signal meets a first assumption condition and a second probability density function when the original echo signal meets a second assumption condition based on an original echo signal corresponding to a millimeter wave radar, combining the first probability density function and the second probability density function, and filtering interference signals in the original echo signal according to a preset detection threshold function to obtain a target echo signal;
the first assumption condition is that the original echo signal only contains an interference signal, and the second assumption condition is that the original echo signal simultaneously contains a target echo signal and an interference signal;
the first probability density function is in a preset threshold range through the detection threshold function, the second probability density function is maximized, signals meeting a first assumption condition are filtered out of the original echo signals, and the rest signals are the target echo signals;
a second unit: the method comprises the steps of obtaining a clustering center of a point cloud set by maximizing intra-cluster similarity of the same type of clusters in the point cloud set and minimizing similarity of different types of clusters based on the point cloud set corresponding to the target echo signal, and forming a cluster according to the clustering center and the points in the point cloud set;
determining an adjacent matrix corresponding to any two vertexes through the initial connectivity of the two vertexes in the point cloud collection;
determining K-order adjacency connectivity and K-order self connectivity corresponding to any two vertexes based on the adjacency matrix, and obtaining a clustering center through a comparison formula;
distributing each vertex in the point cloud set to a corresponding clustering center to form a clustering cluster;
a third unit: the method comprises the steps of acquiring a point cluster centroid according to the cluster and measurement point data and tracks acquired in advance, and tracking a target object corresponding to the original echo signal through a multi-target tracking algorithm based on the point cluster centroid;
calculating the distance between each measuring point and an initial target, obtaining the association degree between each measuring point and the initial target, judging whether each measuring point is in a preset self-adaptive threshold, marking the measuring points in the self-adaptive threshold as effective points, and deleting the rest measuring points;
determining a cost function of an effective point track and an initial target, mutually pairing all measuring points with all tracks by minimizing the cost function, regarding all measuring points associated with the same track as a target point cluster, and acquiring a point cluster centroid of the target point cluster;
and determining a weight value of each track through a multi-target tracking algorithm based on the point cluster centroid, and determining an estimated state of a target object corresponding to the original echo signal based on the weight value and the point cluster centroid.
6. An electronic device, comprising:
a processor;
a memory for storing processor-executable instructions;
wherein the processor is configured to invoke the instructions stored in the memory to perform the method of any of claims 1 to 4.
7. A computer readable storage medium having stored thereon computer program instructions, which when executed by a processor, implement the method of any of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311159083.4A CN116879863B (en) | 2023-09-09 | 2023-09-09 | Multi-target measuring method and system for continuous wave 4D millimeter wave radar |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311159083.4A CN116879863B (en) | 2023-09-09 | 2023-09-09 | Multi-target measuring method and system for continuous wave 4D millimeter wave radar |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116879863A CN116879863A (en) | 2023-10-13 |
| CN116879863B true CN116879863B (en) | 2023-12-05 |
Family
ID=88270311
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311159083.4A Active CN116879863B (en) | 2023-09-09 | 2023-09-09 | Multi-target measuring method and system for continuous wave 4D millimeter wave radar |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116879863B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118568525B (en) * | 2024-07-29 | 2024-11-12 | 中国人民解放军海军航空大学 | An adaptive clustering method for radar echoes |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102023294A (en) * | 2009-09-09 | 2011-04-20 | 中国人民解放军海军航空工程学院 | Detection method for radar multi-target Hough transform target-by-target elimination |
| CN104077498A (en) * | 2014-07-22 | 2014-10-01 | 西安电子科技大学 | Multi-target tracking method by adopting external illuminating radar and combining target angles |
| CN111338338A (en) * | 2020-02-20 | 2020-06-26 | 山东科技大学 | Robot speed self-adaptive control method based on road surface characteristic cluster analysis |
| CN114137525A (en) * | 2021-10-29 | 2022-03-04 | 际络科技(上海)有限公司 | Multi-target detection method and system based on vehicle millimeter-wave radar |
| CN218037334U (en) * | 2022-08-16 | 2022-12-13 | 德心智能科技(常州)有限公司 | Millimeter wave radar with infrared sensor |
| CN115494472A (en) * | 2022-11-16 | 2022-12-20 | 中南民族大学 | A positioning method based on enhanced radar wave signal, millimeter wave radar, and device |
| CN115542308A (en) * | 2022-12-05 | 2022-12-30 | 德心智能科技(常州)有限公司 | Indoor personnel detection method, device, equipment and medium based on millimeter wave radar |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| IL250253B (en) * | 2017-01-24 | 2021-10-31 | Arbe Robotics Ltd | Method for separating targets and clutter from noise in radar signals |
| US10712437B2 (en) * | 2017-07-07 | 2020-07-14 | Veoneer Us, Inc. | Radar systems and methods utilizing composite waveforms for customization of resolution requirements |
| US20200116851A1 (en) * | 2018-10-10 | 2020-04-16 | Government Of The United States, As Represented By The Secretary Of The Air Force | Robust Constant False Alarm Rate (CFAR) Detector for Interference-Plus-Noise Covariance Matrix Mismatch |
-
2023
- 2023-09-09 CN CN202311159083.4A patent/CN116879863B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102023294A (en) * | 2009-09-09 | 2011-04-20 | 中国人民解放军海军航空工程学院 | Detection method for radar multi-target Hough transform target-by-target elimination |
| CN104077498A (en) * | 2014-07-22 | 2014-10-01 | 西安电子科技大学 | Multi-target tracking method by adopting external illuminating radar and combining target angles |
| CN111338338A (en) * | 2020-02-20 | 2020-06-26 | 山东科技大学 | Robot speed self-adaptive control method based on road surface characteristic cluster analysis |
| CN114137525A (en) * | 2021-10-29 | 2022-03-04 | 际络科技(上海)有限公司 | Multi-target detection method and system based on vehicle millimeter-wave radar |
| CN218037334U (en) * | 2022-08-16 | 2022-12-13 | 德心智能科技(常州)有限公司 | Millimeter wave radar with infrared sensor |
| CN115494472A (en) * | 2022-11-16 | 2022-12-20 | 中南民族大学 | A positioning method based on enhanced radar wave signal, millimeter wave radar, and device |
| CN115542308A (en) * | 2022-12-05 | 2022-12-30 | 德心智能科技(常州)有限公司 | Indoor personnel detection method, device, equipment and medium based on millimeter wave radar |
Non-Patent Citations (5)
| Title |
|---|
| Multiple Targets Tracking of Vehicle Anti-collision Radar Based on Probability Hypothesis Density;Yu Z;《Modern Radar》;全文 * |
| 刘美 ; 徐小玲 ; 黄瑞龙 ; 方湃盛 ; .无线传感器网络多目标跟踪技术研究进展.广东石油化工学院学报.2023,全文. * |
| 基于概率假设密度滤波和数据关联的脉冲多普勒雷达多目标跟踪算法;谭顺成;《电子与信息学报》;全文 * |
| 基于聚类和随机参考分类器的对海雷达目标识别算法;范学满;胡生亮;贺静波;;模式识别与人工智能(11);全文 * |
| 有源压制干扰下的相控阵雷达多目标跟踪时间资源优化配置算法;陶庭宝;张弓;张劲东;;数据采集与处理(05);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116879863A (en) | 2023-10-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108303692B (en) | Multi-target tracking method for solving Doppler ambiguity | |
| CN111899568B (en) | Bridge anti-collision early warning system, method and device and storage medium | |
| US8314732B2 (en) | Adaptive radar | |
| US8094060B2 (en) | Method of detecting a target | |
| US8718323B2 (en) | Batch detection association for enhanced target descrimination in dense detection environments | |
| CN112881993B (en) | Method for automatically identifying false flight path caused by radar distribution clutter | |
| CN106918807A (en) | A kind of Targets Dots condensing method of radar return data | |
| CN107576959B (en) | A Pre-detection Tracking Method for High Repetition-Frequency Radar Targets Based on Area Map Deblurring | |
| CN115061113B (en) | Target detection model training method and device for radar and storage medium | |
| JP2019138862A (en) | Radar signal processor | |
| CN113960587B (en) | Millimeter wave radar multi-target tracking method based on category information feedback | |
| CN107561509A (en) | A kind of airborne millimetre-wave radar electric power line detecting method | |
| CN116879863B (en) | Multi-target measuring method and system for continuous wave 4D millimeter wave radar | |
| CN112731307A (en) | RATM-CFAR detector based on distance-angle joint estimation and detection method | |
| JP2009236720A (en) | Moving target detector | |
| Kim et al. | Deep-learning based multi-object detection and tracking using range-angle map in automotive radar systems | |
| CN106772357B (en) | AI-PHD filter multi-object tracking method under signal-to-noise ratio unknown condition | |
| Ott et al. | Uncertainty-based meta-reinforcement learning for robust radar tracking | |
| KR102361816B1 (en) | Method for detecting target and readable medium | |
| KR102211844B1 (en) | Method and apparatus for estimating behind wall Multi-target in an IR-UWB Radar system | |
| Kohlleppel | Ground target tracking with signal adaptive measurement error covariance matrix | |
| Mecocci et al. | Radar image processing for ship-traffic control | |
| CN116224280B (en) | Radar target detection method, radar target detection device, radar equipment and storage medium | |
| Ren et al. | Discrimination of automotive radar distributed targets | |
| Chatterjee et al. | Effects of measurement uncertainties on multitarget tracking |
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 |