CN111045008B - Vehicle millimeter wave radar target identification method based on widening calculation - Google Patents
Vehicle millimeter wave radar target identification method based on widening calculation Download PDFInfo
- Publication number
- CN111045008B CN111045008B CN202010042315.8A CN202010042315A CN111045008B CN 111045008 B CN111045008 B CN 111045008B CN 202010042315 A CN202010042315 A CN 202010042315A CN 111045008 B CN111045008 B CN 111045008B
- Authority
- CN
- China
- Prior art keywords
- doppler
- target
- distance
- data
- dimensional matrix
- 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
- 238000004364 calculation method Methods 0.000 title claims abstract description 52
- 238000000034 method Methods 0.000 title claims abstract description 52
- 239000011159 matrix material Substances 0.000 claims abstract description 98
- 238000012549 training Methods 0.000 claims abstract description 31
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 20
- 238000001514 detection method Methods 0.000 claims abstract description 16
- 238000012360 testing method Methods 0.000 claims description 16
- 238000002790 cross-validation Methods 0.000 claims description 7
- 238000012935 Averaging Methods 0.000 claims description 2
- 238000000605 extraction Methods 0.000 abstract description 2
- 239000000523 sample Substances 0.000 description 26
- 230000008569 process Effects 0.000 description 10
- 238000004590 computer program Methods 0.000 description 7
- 230000006870 function Effects 0.000 description 5
- 239000013598 vector Substances 0.000 description 5
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 238000012706 support-vector machine Methods 0.000 description 3
- 238000004891 communication Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 102100034112 Alkyldihydroxyacetonephosphate synthase, peroxisomal Human genes 0.000 description 1
- 101000799143 Homo sapiens Alkyldihydroxyacetonephosphate synthase, peroxisomal Proteins 0.000 description 1
- 238000000848 angular dependent Auger electron spectroscopy Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000035699 permeability Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
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
- G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
- G01S13/88—Radar or analogous systems specially adapted for specific applications
- G01S13/93—Radar or analogous systems specially adapted for specific applications for anti-collision purposes
- G01S13/931—Radar or analogous systems specially adapted for specific applications for anti-collision purposes of land vehicles
Abstract
The application belongs to the technical field of radar target type recognition, and provides a vehicle-mounted millimeter wave radar target recognition method based on widening calculation, which comprises the following steps: acquiring a received radar echo signal, and extracting space distribution information and Doppler information of the radar echo signal to obtain a distance-Doppler two-dimensional matrix; performing target detection on the data of the distance-Doppler two-dimensional matrix to obtain characteristic data of target point tracks; clustering the characteristic data according to a preset clustering algorithm to obtain clustering data of the target; extracting target features from the clustered data according to the distance-Doppler two-dimensional matrix, and constructing feature sample data according to the target features; training a target recognition model according to the characteristic sample data, and constructing the target recognition model. The method and the device solve the problem that target feature extraction is difficult due to point cloud data sparseness of the millimeter wave radar.
Description
Technical Field
The invention relates to the technical field of radar target type recognition, in particular to a vehicle millimeter wave radar target recognition method based on widening calculation.
Background
Millimeter wave radar has been widely used in the automotive field since the development of the 20 th century, and is well known as an important component of intelligent traffic in driving assistance and autopilot applications. BSD, i.e., blind spot detection, has remained the highest permeability and steadily increased from 2017. The basic function of millimeter wave radar, which is an important component in (ADAS) advanced driving assistance systems, has shown very strong market vitality, not only is technically mature and stable, but also has high user acceptance.
In the BSD scheme system, it is important that the radar can effectively identify the type of the target in the blind area, so as to feed back the target to the control system, and then make a correct driving operation. Pedestrian and vehicle targets are the most typical road targets and therefore correct identification of pedestrian and vehicle targets is essential to BSD systems.
In the prior art, point cloud data of millimeter wave radars are clustered to obtain target point cloud data of different categories, then a feature vector sample set is established, the feature sample set is divided into a training set and a testing set by a ten-fold cross validation method, and a kernel support vector machine is trained by the training set. However, since the point cloud of the millimeter wave radar is generally sparse, the detection target is likely to be an isolated point in actual use, and in this case, the algorithm is disabled.
Disclosure of Invention
In view of the above, the embodiment of the invention provides a vehicle millimeter wave radar target identification method based on widening calculation, so as to solve the problem of difficult target feature extraction caused by sparse point cloud data of a millimeter wave radar.
The first aspect of the embodiment of the invention provides a vehicle millimeter wave radar target identification method based on widening calculation, which comprises the following steps:
acquiring a received radar echo signal, and extracting space distribution information and Doppler information of the radar echo signal to obtain a distance-Doppler two-dimensional matrix;
performing target detection on the data of the distance-Doppler two-dimensional matrix to obtain characteristic data of target point tracks;
clustering the characteristic data according to a preset clustering algorithm to obtain clustering data of the target;
extracting target features from the clustered data according to the distance-Doppler two-dimensional matrix, and constructing feature sample data according to the target features; the target features comprise Doppler broadening of the maximum amplitude point trace in the clustered data and distance broadening of the maximum amplitude point trace in the clustered data;
training a target recognition model according to the characteristic sample data, and constructing the target recognition model.
In one implementation example, the target feature further includes a target average speed of the cluster data, a maximum value of different trace speed differences in the cluster data, and a maximum value of different trace distance differences in the cluster data.
In one implementation example, when extracting doppler spread of a trace of maximum amplitude in the cluster data or distance spread of a trace of maximum amplitude in the cluster data from the cluster data, the extracting a target feature from the cluster data according to the distance-doppler two-dimensional matrix, constructing feature sample data according to the target feature includes:
determining a point trace with the largest amplitude in the cluster data as a reference point trace;
searching points in the range of Doppler and range-widening calculation corresponding to the datum mark in the distance-Doppler two-dimensional matrix; the coordinates of the Doppler dimension in the distance-Doppler two-dimensional matrix are Doppler units, and the coordinates of the distance dimension are distance units;
and calculating Doppler broadening or distance broadening of the datum mark according to the searched points.
In one implementation example, the calculating the doppler spread or the range spread of the fiducial trace from the found points includes:
For a plurality of searched points with the same Doppler units or range units, if the amplitudes of the plurality of searched points are larger than the amplitude of the datum mark, carrying out amplitude limiting calculation on the amplitudes of the plurality of searched points according to the amplitude of the datum mark, calculating the average value of the plurality of searched points after amplitude limitation, and accumulating the calculated average value into Doppler broadening or range broadening;
for a plurality of searched points with the same Doppler units or range units, if the amplitudes of the plurality of searched points are smaller than the amplitudes of the datum mark, calculating the average value of the plurality of searched points, and accumulating the calculated average value into Doppler broadening or range broadening;
and traversing all points in the range of the spread calculation corresponding to the reference point trace in the distance-Doppler two-dimensional matrix to obtain the Doppler spread or the distance spread of the reference point trace.
In one implementation example, when doppler spread of a point trace with the largest amplitude in the cluster data is extracted from the cluster data, a spread calculation range corresponding to the point trace in the distance-doppler two-dimensional matrix includes points in a doppler dimension and a distance dimension centered on the point trace; and defining a range of the Doppler dimensions in the spread calculation range according to the characteristics of the target.
In one implementation example, when extracting the range spread of the maximum-amplitude point trace in the cluster data from the cluster data, the range of spread calculation corresponding to the reference point trace in the range-doppler two-dimensional matrix includes points in the doppler dimension and the range dimension centered on the reference point trace; and defining a range of the distance dimension in the spread calculation range based on the characteristics of the target.
In one implementation example, the training the object recognition model according to the feature sample data, and constructing the object recognition model includes:
dividing the characteristic sample data into training data and test data in a cross-validation mode;
training the target recognition model according to the training data to construct the target recognition model so as to recognize the target;
and inputting the test data into the target recognition model to test the target recognition model.
In one implementation example, after training the object recognition model according to the feature sample data and constructing the object recognition model, the method further includes:
extracting the space distribution information and Doppler information of each received radar echo signal frame to obtain a distance-Doppler two-dimensional matrix;
Performing target detection on the data of the distance-Doppler two-dimensional matrix to obtain characteristic data of target point tracks;
clustering the characteristic data according to a preset clustering algorithm to obtain clustering data of the target;
and extracting target features from the clustered data according to the distance-Doppler two-dimensional matrix, and inputting the target features into the target recognition model to obtain a target recognition result.
In one implementation example, after extracting the target feature from the cluster data according to the distance-doppler two-dimensional matrix, inputting the target feature into the target recognition model to obtain a target recognition result, the method further includes:
and if the target recognition results corresponding to the two frames of radar echo signals are the same in the three continuous frames of radar echo signals, the target detected by the radar is the target recognition result.
In one implementation example, the characteristic data includes distance, speed, angle, and amplitude.
According to the vehicle-mounted millimeter wave radar target identification method based on widening calculation, the space distribution information and the Doppler information of the radar echo signals are extracted by acquiring the received radar echo signals, so that a distance-Doppler two-dimensional matrix is obtained; performing target detection on the data of the distance-Doppler two-dimensional matrix to obtain characteristic data of target point tracks; clustering the characteristic data according to a preset clustering algorithm to obtain clustering data of the target; extracting target features from the clustered data according to the distance-Doppler two-dimensional matrix, and constructing feature sample data according to the target features; the target features comprise Doppler broadening of the maximum amplitude point trace in the clustered data and distance broadening of the maximum amplitude point trace in the clustered data; training a target recognition model according to the characteristic sample data, and constructing the target recognition model. Under the condition that only a single point trace exists in the clustered data, the distance and Doppler broadening characteristics of targets in the clustered data can be accurately extracted according to the distance-Doppler two-dimensional matrix. The problem that target characteristics are difficult to extract due to the fact that point cloud data of the millimeter wave radar are sparse is solved, and effective identification and distinguishing of targets are achieved.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings that are needed in the embodiments or the description of the prior art will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and that other drawings can be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a schematic flow chart of a vehicle millimeter wave radar target recognition method based on widening calculation according to an embodiment of the present invention;
figure 2 is a schematic diagram of a structure providing a range-doppler two-dimensional matrix in accordance with an embodiment of the present invention;
FIG. 3 is a flowchart illustrating a method for calculating Doppler spread of maximum amplitude traces in clustered data according to an embodiment of the present invention;
FIG. 4 is a schematic flow chart of calculating the distance broadening of the maximum amplitude trace in the clustered data according to the first embodiment of the present invention;
FIG. 5 is a table diagram of test results of a target recognition model according to an embodiment of the present invention;
fig. 6 is a schematic flow chart of a vehicle millimeter wave radar target recognition method based on widening calculation according to a second embodiment of the present invention.
Detailed Description
In order that those skilled in the art will better understand the present invention, a technical solution of an embodiment of the present invention will be clearly described below with reference to the accompanying drawings in the embodiment of the present invention, and it is apparent that the described embodiment is a part of the embodiment of the present invention, but not all the embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the present invention without making any inventive effort, shall fall within the scope of the present invention.
The term "comprising" in the description of the invention and the claims and in the above figures and any variants thereof is intended to cover a non-exclusive inclusion. For example, a process, method, or system, article, or apparatus that comprises a list of steps or elements is not limited to only those listed but may optionally include additional steps or elements not listed or inherent to such process, method, article, or apparatus. Furthermore, the terms "first," "second," and "third," etc. are used for distinguishing between different objects and not for describing a particular sequential order.
Example 1
Fig. 1 is a schematic flow chart of a vehicle millimeter wave radar target recognition method based on widening calculation according to an embodiment of the present invention. In the embodiment of the application, the training device of the target recognition model is taken as an execution main body for explanation, and the method specifically comprises the following steps:
s110, acquiring a received radar echo signal, and extracting spatial distribution information and Doppler information of the radar echo signal to obtain a distance-Doppler two-dimensional matrix;
in the prior art, the training process of the target recognition model of the radar is to cluster point cloud data of the millimeter wave radar to obtain different types of target point cloud data, then a feature vector sample set is established, the feature sample set is divided into a training set and a testing set by a ten-fold cross validation method, and a kernel support vector machine is trained by the training set. However, since the point cloud of the millimeter wave radar is generally sparse, the detection target is likely to be an isolated point in actual use, and in this case, the algorithm is disabled. In order to solve the problem, the embodiment extracts the spatial distribution information and the Doppler information of the sampled radar echo signals to obtain a distance-Doppler two-dimensional matrix, and extracts the target features from the clustering data of the target point trace according to the distance-Doppler two-dimensional matrix, thereby solving the problem that the target features are difficult to extract due to the sparse point cloud data of the millimeter wave radar. Alternatively, the radar may be a vehicle millimeter wave radar.
Specifically, after a target, such as a pedestrian, a vehicle, or the like, is detected by transmitting a chirp signal through a millimeter wave radar, a radar echo signal reflected by the target is received. Alternatively, the received radar return signals may be sorted into a radar data cube (radarCube). And extracting the space distribution information and Doppler information of the received radar echo signals to obtain a Range-Doppler Matrix (RDM). As shown in fig. 2, a schematic structure of the calculated range-doppler two-dimensional matrix is shown, and a range cell (range bin) and a doppler cell (dopplerBin) in the range-doppler two-dimensional matrix are used as coordinate positions to form a range dimension and a doppler dimension. In fig. 2, each square in the range-doppler two-dimensional matrix corresponds to a point in the range-doppler two-dimensional matrix, and each point in the range-doppler two-dimensional matrix has a different [ range bin, doppler bin ] coordinate position and magnitude.
S120, performing target detection on the data of the distance-Doppler two-dimensional matrix to obtain characteristic data of target point tracks;
after the space distribution information and Doppler information of the received radar echo signals are extracted to obtain a distance-Doppler two-dimensional matrix, target detection is carried out on the distance-Doppler two-dimensional matrix (RDM), and characteristic data of corresponding points of targets (such as pedestrians, vehicles and the like) in the distance-Doppler two-dimensional matrix (RDM) are extracted to obtain characteristic data of the target points. In one implementation example, the acquired feature data of the target point trace includes a distance R, a velocity V, an angle θ, an amplitude a, and the like.
S130, clustering the characteristic data according to a preset clustering algorithm to obtain clustering data of the target;
after target detection is carried out on the data of the distance-Doppler two-dimensional matrix to obtain characteristic data of target points, the characteristic data can be clustered through a preset clustering algorithm, and characteristic data belonging to the same target in the characteristic data of the target points are clustered together to obtain clustered data of each target. Optionally, the clustering algorithm and the clustering parameters in the clustering algorithm are preset and adjusted according to the actual parameter configuration of the radar.
S140, extracting target features from the clustered data according to the distance-Doppler two-dimensional matrix, and constructing feature sample data according to the target features; the target features comprise Doppler broadening of the maximum amplitude point trace in the clustered data and distance broadening of the maximum amplitude point trace in the clustered data;
after clustering the feature data according to a preset clustering algorithm to obtain the clustering data of the target, the feature vector can be established by extracting the feature of the target from the clustering data of the target according to the distance-Doppler two-dimensional matrix (RDM) calculated in the step 110. And acquiring a large number of radar echo signals to obtain a large number of target features, and constructing feature sample data according to the acquired large number of target features. The method comprises the steps of extracting target characteristics from clustering data of a target, wherein the target characteristics comprise Doppler broadening of a maximum-amplitude point trace in the clustering data and distance broadening of the maximum-amplitude point trace in the clustering data.
In one implementation example, extracting the target feature from the cluster data of the target further includes a target average speed of the cluster data, a maximum value of different trace speed differences in the cluster data, and a maximum value of different trace distance differences in the cluster data. Specifically, the target feature: the target average speed of the cluster data, the maximum value of the difference values of the speeds of different points in the cluster data and the maximum value of the difference values of the distances of different points in the cluster data can be directly extracted from the cluster data of the target.
In one implementation example, when extracting the doppler spread of the maximum amplitude trace in the cluster data from the cluster data, the extracting the target feature from the cluster data according to the distance-doppler two-dimensional matrix, and constructing feature sample data according to the target feature includes: determining a point trace with the largest amplitude in the cluster data as a reference point trace; searching points in the range of Doppler and range-widening calculation corresponding to the datum mark in the distance-Doppler two-dimensional matrix; calculating Doppler spread of the datum mark according to the searched points; the coordinates of the Doppler dimension in the distance-Doppler two-dimensional matrix are Doppler units, and the coordinates of the distance dimension are distance units.
Specifically, since the target feature to be extracted is doppler broadening of the trace with the largest amplitude in the clustered data, the trace with the largest amplitude needs to be found out from the clustered data. The trace of the point in the cluster data with the largest amplitude may be determined to be the trace of the reference point. In order to avoid the problem that the target characteristics cannot be extracted from the clustered data due to the fact that only a single point trace exists in the clustered data, doppler broadening of the point trace with the largest amplitude in the clustered data is calculated according to the distance-Doppler two-dimensional matrix and the clustered data, wherein the distance-Doppler two-dimensional matrix and the clustered data are obtained by extracting the space distribution information and the Doppler information of the received radar echo signals. And determining the point trace with the largest amplitude in the cluster data as a reference point trace, and determining the point in the widening calculation range corresponding to the reference point trace in the distance-Doppler two-dimensional matrix according to the coordinate positions of the reference point trace in the distance and Doppler dimensions in the distance-Doppler two-dimensional matrix. Thereby calculating the Doppler spread of the reference trace from the found points.
In one implementation example, the specific process of calculating the doppler spread of the fiducial trace from the found points may be: for a plurality of searched points with the same Doppler units in each group, if the amplitudes of the plurality of searched points are larger than the amplitude of the datum mark, carrying out limit amplitude calculation on the amplitudes of the plurality of searched points according to the amplitude of the datum mark, calculating the average value of the plurality of searched points after amplitude limitation, and accumulating the calculated average value into Doppler broadening; for a plurality of searched points with the same Doppler units in each group, if the amplitudes of the plurality of searched points are smaller than the amplitudes of the datum mark, calculating the average value of the plurality of searched points, and accumulating the calculated average value into Doppler broadening; and traversing all points in the range of spread calculation corresponding to the reference point trace in the distance-Doppler two-dimensional matrix to obtain the Doppler spread of the reference point trace.
Because the points with the amplitude larger than the reference point trace in the widening calculation range corresponding to the reference point trace in the distance-Doppler two-dimensional matrix may be the points corresponding to the adjacent targets, in order to reduce the calculation interference of the points of the adjacent targets on the Doppler widening of the reference point trace, if the amplitudes of a plurality of searched points which are the same in each group of Doppler units are all larger than the amplitude of the reference point trace, the amplitude limiting calculation can be performed on a plurality of points in the group, the average value of a plurality of searched points in the group after the amplitude limitation is calculated, and the calculated average value is accumulated into the Doppler widening. If the amplitudes of the same plurality of searched points in each group of Doppler units are smaller than the amplitudes of the datum mark, calculating the average value of the plurality of searched points in the group, and accumulating the calculated average value into Doppler broadening; and traversing all points in the range of the spread calculation corresponding to the reference point trace in the distance-Doppler two-dimensional matrix to obtain the Doppler spread of the reference point trace.
In one implementation example, when doppler spread of a point trace with the largest amplitude in the cluster data is extracted from the cluster data, a spread calculation range corresponding to the point trace in the distance-doppler two-dimensional matrix includes points in a doppler dimension and a distance dimension centered on the point trace. Due to the extra motion of the arms and legs of the pedestrian while walking, the speed of the pedestrian target will have extra spread, while the distance direction will not have obvious spread; in contrast, since the motor vehicle has a certain length, the motor vehicle has an additional widening in the distance direction without a significant widening in the speed. The range of the Doppler dimensions in the spread calculation range may be defined according to characteristics of the target.
Specifically, in order to make the doppler spread of the maximum amplitude point trace in the cluster data calculated according to the distance-doppler two-dimensional matrix and the cluster data more accurate and stable, the spread calculation range corresponding to the reference point trace in the distance-doppler two-dimensional matrix may be a point in the doppler dimension and the distance dimension centered on the reference point trace, that is, a point in a shadow portion in the range 21 as shown in fig. 2; and defining a range of the Doppler dimensions in the spread calculation range according to the characteristics of the target.
In detail, as shown in fig. 2, the black square position in the preset graph is the trace of the point with the largest amplitude in the cluster data, namely the trace of the reference point; acquiring a coordinate position [ dopplerBin, rangeBin ] of the reference point trace in a distance-Doppler two-dimensional matrix, and searching a point positioned in a widening calculation range corresponding to the reference point trace in the distance-Doppler two-dimensional matrix; the specific process of calculating the Doppler spread of the datum mark according to the searched points can be as follows: defining and setting a reference unit number N according to the characteristics of the target, wherein N is an odd number; FIG. 3 is a schematic flow chart showing the calculation of Doppler spread of maximum amplitude traces in clustered data;
specifically, RDM [ dopplerBin, rangeBin ] is amplitude information of a point whose coordinate position is [ dopplerBin, rangeBin ] in the range-doppler two-dimensional matrix.
And step 18, determining the Doppler spread of the trace with the largest amplitude in the cluster data as D_e.
In one implementation example, when extracting the distance broadening of the maximum-amplitude trace in the cluster data from the cluster data, extracting target features from the cluster data according to the distance-doppler two-dimensional matrix, and constructing feature sample data according to the target features includes: determining a point trace with the largest amplitude in the cluster data as a reference point trace; searching points in the range of Doppler and range-widening calculation corresponding to the datum mark in the distance-Doppler two-dimensional matrix; and calculating the distance broadening of the datum mark according to the searched points.
Specifically, since the target feature to be extracted is the distance broadening of the maximum amplitude trace in the clustered data, the trace with the maximum amplitude needs to be found out from the clustered data. The trace of the point in the cluster data with the largest amplitude may be determined to be the trace of the reference point. In order to avoid the problem that the target characteristics cannot be extracted from the clustered data due to the fact that only a single point trace exists in the clustered data, the distance broadening of the point trace with the largest amplitude in the clustered data is calculated according to the distance-Doppler two-dimensional matrix obtained by extracting the space distribution information and Doppler information of the received radar echo signals and the clustered data. And determining the point trace with the largest amplitude in the cluster data as a reference point trace, and determining the point in the widening calculation range corresponding to the reference point trace in the distance-Doppler two-dimensional matrix according to the coordinate positions of the reference point trace in the distance and Doppler dimensions in the distance-Doppler two-dimensional matrix. Thereby calculating the distance broadening of the reference point trace according to the searched points.
In one implementation example, the specific process of calculating the distance broadening of the reference trace according to the found point may be: for a plurality of searched points with the same distance units in each group, if the amplitudes of the plurality of searched points are larger than the amplitude of the datum mark, carrying out limit amplitude calculation on the amplitudes of the plurality of searched points according to the amplitude of the datum mark, calculating the average value of the plurality of searched points after amplitude limitation, and accumulating the calculated average value into the distance broadening; for a plurality of searched points with the same distance units in each group, if the amplitudes of the plurality of searched points are smaller than the amplitudes of the datum mark, calculating the average value of the plurality of searched points, and accumulating the calculated average value into the distance broadening; and traversing all points in the range of the spread calculation corresponding to the reference point trace in the distance-Doppler two-dimensional matrix to obtain the distance spread of the reference point trace.
Because the points with the amplitude larger than the datum mark in the widening calculation range corresponding to the datum mark in the distance-Doppler two-dimensional matrix may be the points corresponding to the adjacent targets, in order to reduce the calculation interference of the points of the adjacent targets on the distance widening of the datum mark, if the amplitudes of a plurality of searched points with the same distance units in each group are all larger than the amplitude of the datum mark, the amplitude limiting calculation can be performed on a plurality of points in the group, the average value of a plurality of searched points in the group after the amplitude limitation is calculated, and the calculated average value is accumulated into the distance widening. If the amplitudes of the plurality of searched points of each group of distance units are smaller than the amplitudes of the datum mark, calculating the average value of the plurality of searched points in the group, and accumulating the calculated average value into the distance broadening; and traversing all points in the range of the spread calculation corresponding to the reference point trace in the distance-Doppler two-dimensional matrix to obtain the distance spread of the reference point trace.
Specifically, in order to make the distance broadening of the maximum-amplitude point trace in the cluster data calculated according to the distance-doppler two-dimensional matrix and the cluster data more accurate and stable, the broadening calculation range corresponding to the reference point trace in the distance-doppler two-dimensional matrix may be a point in the doppler dimension and the distance dimension centered on the reference point trace, that is, a point in a shadow portion in the range 22 as shown in fig. 2; and defining a range of distance dimensions in a spread calculation range based on characteristics of the target.
In detail, as shown in fig. 2, the black square position in the preset graph is the trace of the point with the largest amplitude in the cluster data, namely the trace of the reference point; acquiring a coordinate position [ dopplerBin, rangeBin ] of the reference point trace in a distance-Doppler two-dimensional matrix, and searching a point positioned in a widening calculation range corresponding to the reference point trace in the distance-Doppler two-dimensional matrix; the specific process of calculating the distance broadening of the datum mark according to the searched points can be as follows: defining and setting a reference unit number N according to the characteristics of the target, wherein N is an odd number; FIG. 4 is a schematic flow chart of calculating the distance broadening of the maximum amplitude trace in the clustered data;
specifically, RDM [ dopplerBin, rangeBin ] is amplitude information of a point whose coordinate position is [ dopplerBin, rangeBin ] in the range-doppler two-dimensional matrix.
and 28, determining the distance broadening of the trace with the largest amplitude in the cluster data as R_e.
And S150, training a target recognition model according to the characteristic sample data, and constructing the target recognition model.
After the feature sample data is constructed according to the extracted target features, the feature sample data can be used as training data to train a target recognition model, so that the construction of the target recognition model is completed, and the target recognition model can be a vector machine (SVM) optionally.
In one implementation example, training the object recognition model according to the feature sample data, and the specific process of constructing the object recognition model may be: dividing the characteristic sample data into training data and test data in a cross-validation mode; training the target recognition model according to the training data to construct the target recognition model so as to recognize the target; and inputting the test data into the target recognition model to test the target recognition model.
Specifically, after the feature sample data is constructed according to the extracted target features, the feature sample data can be further divided into training data and test data by adopting a cross-validation mode. Training the target recognition model by using training data to construct the target recognition model so as to recognize the target. Alternatively, the targets include pedestrians, vehicles, and the like. The test data may also be input into the trained object recognition model to test the constructed object recognition model. In detail, in this embodiment, a NXP S32R274 radar chip may be used, and a TEF8102 radio frequency chip may be mounted. The relevant radar configuration parameters are as follows: the 2 x 4MIMO-TDM mode is adopted for the distance resolution 0.3979m, the distance unit number 128, the speed resolution 0.1152m/s, the speed unit number 256 and the maximum non-blurring speed 29.5 m/s. 1000 groups of data are collected by the pedestrian target and the vehicle target, 5 features are extracted from the collected data respectively through the method provided by the invention, and a feature set is constructed. The support vector machine kernel function adopts a Gaussian kernel function, and adopts a simple cross-validation mode, namely a 70% training set and a 30% testing set. The final results are shown in fig. 5, and the recognition accuracy of pedestrians and vehicles is up to 91.67% and 93.33%.
According to the vehicle-mounted millimeter wave radar target identification method based on widening calculation, the space distribution information and the Doppler information of the radar echo signals are extracted by acquiring the received radar echo signals, so that a distance-Doppler two-dimensional matrix is obtained; performing target detection on the data of the distance-Doppler two-dimensional matrix to obtain characteristic data of target point tracks; clustering the characteristic data according to a preset clustering algorithm to obtain clustering data of the target; extracting target features from the clustered data according to the distance-Doppler two-dimensional matrix, and constructing feature sample data according to the target features; the target features comprise Doppler broadening of the maximum amplitude point trace in the clustered data and distance broadening of the maximum amplitude point trace in the clustered data; training a target recognition model according to the characteristic sample data, and constructing the target recognition model. Under the condition that only a single point trace exists in the clustered data, the distance and Doppler broadening characteristics of targets in the clustered data can be accurately extracted according to the distance-Doppler two-dimensional matrix. The problem that target characteristics are difficult to extract due to the fact that point cloud data of the millimeter wave radar are sparse is solved, and effective identification and distinguishing of targets are achieved.
Example two
Fig. 6 is a schematic flow chart of a vehicle millimeter wave radar target recognition method based on widening calculation according to a second embodiment of the present invention. On the basis of the first embodiment, the present embodiment further provides a method for identifying a target detected by the radar echo signal by using the target identification model, so as to identify the detected target of the radar. The method specifically comprises the following steps:
s210, extracting space distribution information and Doppler information of each received frame of radar echo signal to obtain a distance-Doppler two-dimensional matrix;
since the radar transmits a probe signal at a preset frequency, a plurality of frames of radar echo signals can be received. For each frame of radar return signals received, each frame of radar return signals may be consolidated into a radar data cube (radarCube). And extracting the space distribution information and Doppler information of the received radar echo signals to obtain a Range-Doppler Matrix (RDM).
S220, performing target detection on the data of the distance-Doppler two-dimensional matrix to obtain characteristic data of target point tracks;
after the space distribution information and Doppler information of a received radar echo signal frame are extracted to obtain a distance-Doppler two-dimensional matrix, target detection is carried out on the distance-Doppler two-dimensional matrix (RDM), and characteristic data of corresponding points of targets (such as pedestrians, vehicles and the like) in the distance-Doppler two-dimensional matrix (RDM) are extracted to obtain characteristic data of the target points. In one implementation example, the acquired feature data of the target point trace includes a distance R, a velocity V, an angle θ, an amplitude a, and the like.
S230, clustering the characteristic data according to a preset clustering algorithm to obtain clustering data of the target;
after target detection is carried out on the data of the distance-Doppler two-dimensional matrix to obtain characteristic data of target points, the characteristic data can be clustered through a preset clustering algorithm, and characteristic data belonging to the same target in the characteristic data of the target points are clustered together to obtain clustered data of each target. Optionally, the clustering algorithm and the clustering parameters in the clustering algorithm are preset and adjusted according to the actual parameter configuration of the radar.
S240, extracting target features from the clustered data according to the distance-Doppler two-dimensional matrix, and inputting the target features into the target recognition model to obtain a target recognition result;
after clustering the feature data according to a preset clustering algorithm to obtain the clustering data of the target, extracting target features from the clustering data of the target according to the calculated distance-Doppler two-dimensional matrix (RDM) to establish feature vectors. The extracting the target features from the clustering data of the targets comprises the following steps: the method comprises the steps of target average speed of cluster data, maximum values of different trace speed differences in the cluster data and maximum values of different trace distance differences in the cluster data, doppler broadening of maximum-amplitude traces in the cluster data and distance broadening of maximum-amplitude traces in the cluster data. Specifically, the target feature: the target average speed of the cluster data, the maximum value of the difference values of the speeds of different points in the cluster data and the maximum value of the difference values of the distances of different points in the cluster data can be directly extracted from the cluster data of the target.
And inputting the extracted 5 target features into the target recognition model trained in the second embodiment to obtain a target recognition result corresponding to the radar echo signal of the frame. Alternatively, the target recognition result may be a pedestrian or a vehicle, or the like.
S250, if two frames of target recognition results corresponding to the radar echo signals exist in three continuous frames of radar echo signals, the targets detected by the radar are the target recognition results.
And (3) caching target recognition results corresponding to the continuous three-frame radar echo signals, and if the target recognition results corresponding to the two-frame radar echo signals are the same in the continuous three-frame radar echo signals, determining the target detected by the current radar as the target recognition result according to the comprehensive Kalman tracking condition. The target type is comprehensively judged by comprehensively tracking the Kalman condition and referring to the history category judgment condition and adopting methods such as 2/3 rule, so that the probability of erroneously judging the target type due to accidental conditions is reduced, and the stability of the system is improved.
It will be apparent to those skilled in the art that, for convenience and brevity of description, only the above-described division of the functional units and modules is illustrated, and in practical application, the above-described functional distribution may be performed by different functional units and modules according to needs, i.e. the internal structure of the apparatus is divided into different functional units or modules to perform all or part of the above-described functions. The functional units and modules in the embodiment may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit, where the integrated units may be implemented in a form of hardware or a form of a software functional unit. In addition, specific names of the functional units and modules are only for convenience of distinguishing from each other, and are not used for limiting the protection scope of the present application. The specific working process of the units and modules in the above system may refer to the corresponding process in the foregoing method embodiment, which is not described herein again.
In the foregoing embodiments, the descriptions of the embodiments are emphasized, and in part, not described or illustrated in any particular embodiment, reference is made to the related descriptions of other embodiments.
Those of ordinary skill in the art will appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.
In the embodiments provided in the present invention, it should be understood that the disclosed apparatus/terminal device and method may be implemented in other manners. For example, the apparatus/terminal device embodiments described above are merely illustrative, e.g., the division of the modules or units is merely a logical function division, and there may be additional divisions in actual implementation, e.g., multiple units or components may be combined or integrated into another system, or some features may be omitted or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection via interfaces, devices or units, which may be in electrical, mechanical or other forms.
The units described as separate units may or may not be physically separate, and units shown as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment. In addition, each functional unit in the embodiments of the present invention may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in software functional units.
The integrated modules/units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer readable storage medium. Based on such understanding, the present invention may implement all or part of the flow of the method of the above embodiment, or may be implemented by a computer program to instruct related hardware, where the computer program may be stored in a computer readable storage medium, and when the computer program is executed by a processor, the computer program may implement the steps of each of the method embodiments described above. Wherein the computer program comprises computer program code which may be in source code form, object code form, executable file or some intermediate form etc. The computer readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a U disk, a removable hard disk, a magnetic disk, an optical disk, a computer Memory, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), an electrical carrier signal, a telecommunications signal, a software distribution medium, and so forth. It should be noted that the computer readable medium contains content that can be appropriately scaled according to the requirements of jurisdictions in which such content is subject to legislation and patent practice, such as in certain jurisdictions in which such content is subject to legislation and patent practice, the computer readable medium does not include electrical carrier signals and telecommunication signals.
The above embodiments are only for illustrating the technical solution of the present invention, and not for limiting the same; although the invention 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 technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention, and are intended to be included in the scope of the present invention.
Claims (9)
1. The vehicle millimeter wave radar target identification method based on widening calculation is characterized by comprising the following steps of:
acquiring a received radar echo signal, and extracting space distribution information and Doppler information of the radar echo signal to obtain a distance-Doppler two-dimensional matrix;
performing target detection on the data of the distance-Doppler two-dimensional matrix to obtain characteristic data of target point tracks;
clustering the characteristic data according to a preset clustering algorithm to obtain clustering data of the target;
extracting target features from the clustered data according to the distance-Doppler two-dimensional matrix, and constructing feature sample data according to the target features; the target features comprise Doppler broadening of the maximum amplitude point trace in the clustered data and distance broadening of the maximum amplitude point trace in the clustered data;
Training a target recognition model according to the characteristic sample data, and constructing the target recognition model;
when Doppler broadening of the maximum-amplitude point trace in the clustered data is extracted from the clustered data, extracting target features from the clustered data according to the distance-Doppler two-dimensional matrix, and constructing feature sample data according to the target features, wherein the method comprises the following steps: determining a point trace with the largest amplitude in the cluster data as a reference point trace; searching points in the range of Doppler and range-widening calculation corresponding to the datum mark in the distance-Doppler two-dimensional matrix; calculating Doppler spread of the datum mark according to the searched points; the coordinates of the Doppler dimension in the distance-Doppler two-dimensional matrix are Doppler units, and the coordinates of the distance dimension are distance units;
calculating Doppler spread of the maximum amplitude trace in the clustered data comprises;
step 11, acquiring a coordinate position [ dopplerBin, rangeBin ] of a reference point trace in a distance-Doppler two-dimensional matrix;
step 12, limiting and setting a reference unit number N according to the characteristics of the target, wherein N is an odd number; initializing a loop variable i= - (N-1)/2; doppler spread d_e=0;
Step 13, calculating a temporary variable rdmt=rdm [ dopplerBin, rangeBin ]; tem1=rdm [ dopplerbin+i, rangeBin-1], tem2=rdm [ dopplerbin+i, rangeBin ], tem3=rdm [ dopplerbin+i, rangebin+1], RDM [ dopplerBin, rangeBin ] is amplitude information of a point whose coordinate position is [ dopplerBin, rangeBin ] in the range-doppler two-dimensional matrix;
step 14, if tem1 is greater than rdmT, tem1=tem1 rdmT/(tem1+rdmt); if tem2 is greater than rdmT, tem2=tem2 rdmT/(tem2+rdmt); if tem3 is greater than rdmT, tem3=tem3 rdmT/(tem3+rdmt);
step 15, if tem1 is less than rdmT, tem2 is less than rdmT, and tem3 is less than rdmT, then step 16 is performed;
step 16, calculating Doppler spread D_e=D_e+mean ([ tem1/rdmT, tem2/rdmT, tem3/rdmT ]), mean representing averaging;
step 17, judging whether i is less than (N-1)/2, if yes, i=i+1, and returning to execute step 13, otherwise, executing step 18;
step 18, determining Doppler spread of the trace with the largest amplitude in the cluster data as D_e;
when extracting the distance broadening of the maximum-amplitude point trace in the clustered data from the clustered data, extracting target features from the clustered data according to the distance-Doppler two-dimensional matrix, and constructing feature sample data according to the target features, wherein the method comprises the following steps: determining a point trace with the largest amplitude in the cluster data as a reference point trace; searching points in the range of Doppler and range-widening calculation corresponding to the datum mark in the distance-Doppler two-dimensional matrix; and calculating the distance broadening of the datum mark according to the searched points.
2. The spread computation based vehicle millimeter wave radar target recognition method of claim 1, wherein the target features further comprise a target average speed of the clustered data, a maximum value of different trace speed differences in the clustered data, and a maximum value of different trace distance differences in the clustered data.
3. The spread computation based vehicle millimeter wave radar target recognition method according to claim 1, wherein the calculating the doppler spread or the distance spread of the reference trace from the found points includes:
for a plurality of searched points with the same Doppler units or distance units in each group, if the amplitudes of the plurality of searched points are larger than the amplitude of the datum mark, carrying out limit amplitude calculation on the amplitudes of the plurality of searched points according to the amplitude of the datum mark, calculating the average value of the plurality of searched points after amplitude limitation, and accumulating the calculated average value into Doppler broadening or distance broadening;
for a plurality of searched points with the same Doppler units or distance units in each group, if the amplitudes of the plurality of searched points are smaller than the amplitudes of the datum mark, calculating the average value of the plurality of searched points, and accumulating the calculated average value into Doppler broadening or distance broadening;
And traversing all points in the range of the spread calculation corresponding to the reference point trace in the distance-Doppler two-dimensional matrix to obtain the Doppler spread or the distance spread of the reference point trace.
4. The spread computation-based vehicle-mounted millimeter wave radar target recognition method according to claim 3, wherein when doppler spread of a point trace with a maximum amplitude in the cluster data is extracted from the cluster data, a spread computation range corresponding to the reference point trace in the range-doppler two-dimensional matrix includes points in a doppler dimension and a range dimension centered on the reference point trace; and defining a range of the Doppler dimensions in the spread calculation range according to the characteristics of the target.
5. The spread computation-based vehicle-mounted millimeter wave radar target recognition method according to claim 4, wherein when a distance spread of a maximum-amplitude point trace in the cluster data is extracted from the cluster data, a spread computation range corresponding to the reference point trace in the distance-doppler two-dimensional matrix includes points in a doppler dimension and a distance dimension centered on the reference point trace; and defining a range of the distance dimension in the spread calculation range based on the characteristics of the target.
6. The vehicle millimeter wave radar target recognition method based on widening calculation according to claim 5, wherein the training the target recognition model according to the characteristic sample data, constructing the target recognition model, comprises:
dividing the characteristic sample data into training data and test data in a cross-validation mode;
training the target recognition model according to the training data to construct the target recognition model so as to recognize the target;
and inputting the test data into the target recognition model to test the target recognition model.
7. The spread computation based vehicle millimeter wave radar target recognition method according to claim 6, further comprising, after training a target recognition model based on the feature sample data, after constructing the target recognition model:
extracting the space distribution information and Doppler information of each received radar echo signal frame to obtain a distance-Doppler two-dimensional matrix;
performing target detection on the data of the distance-Doppler two-dimensional matrix to obtain characteristic data of target point tracks;
Clustering the characteristic data according to a preset clustering algorithm to obtain clustering data of the target;
and extracting target features from the clustered data according to the distance-Doppler two-dimensional matrix, and inputting the target features into the target recognition model to obtain a target recognition result.
8. The spread computation based vehicle-mounted millimeter wave radar target recognition method according to claim 7, wherein after extracting target features from the clustered data according to the distance-doppler two-dimensional matrix, inputting the target features into the target recognition model to obtain a target recognition result, further comprising:
and if the target recognition results corresponding to the two frames of radar echo signals are the same in the three continuous frames of radar echo signals, the target detected by the radar is the target recognition result.
9. The spread computation based vehicle millimeter wave radar target identification method according to any one of claims 1 to 6, wherein the characteristic data includes a distance, a speed, an angle, and an amplitude.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010042315.8A CN111045008B (en) | 2020-01-15 | 2020-01-15 | Vehicle millimeter wave radar target identification method based on widening calculation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010042315.8A CN111045008B (en) | 2020-01-15 | 2020-01-15 | Vehicle millimeter wave radar target identification method based on widening calculation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111045008A CN111045008A (en) | 2020-04-21 |
| CN111045008B true CN111045008B (en) | 2023-06-09 |
Family
ID=70244631
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010042315.8A Active CN111045008B (en) | 2020-01-15 | 2020-01-15 | Vehicle millimeter wave radar target identification method based on widening calculation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111045008B (en) |
Families Citing this family (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111427032B (en) * | 2020-04-24 | 2022-02-01 | 森思泰克河北科技有限公司 | Room wall contour recognition method based on millimeter wave radar and terminal equipment |
| CN111983595B (en) * | 2020-07-14 | 2023-11-10 | 北京航空航天大学杭州创新研究院 | Indoor positioning method and device |
| CN112162271A (en) * | 2020-08-18 | 2021-01-01 | 河北省交通规划设计院 | Vehicle type recognition method of microwave radar under multiple scenes |
| CN111929655B (en) * | 2020-09-08 | 2023-07-14 | 中国电子科技集团公司第三十八研究所 | Method and system for tracking road target of millimeter wave radar of automobile |
| CN112906519B (en) * | 2021-02-04 | 2023-09-26 | 北京邮电大学 | Vehicle type identification method and device |
| CN113269242B (en) * | 2021-05-18 | 2022-03-08 | 电子科技大学 | Target detection trace aggregation method based on peak clustering |
| CN114814778B (en) * | 2022-06-29 | 2022-09-20 | 长沙莫之比智能科技有限公司 | Carrier speed calculation method based on millimeter wave radar |
| CN115128571B (en) * | 2022-09-02 | 2022-12-20 | 长沙莫之比智能科技有限公司 | Multi-person and non-motor vehicle identification method based on millimeter wave radar |
| CN115236627B (en) * | 2022-09-21 | 2022-12-16 | 深圳安智杰科技有限公司 | Millimeter wave radar data clustering method based on multi-frame Doppler velocity dimension expansion |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110488278A (en) * | 2019-08-20 | 2019-11-22 | 深圳锐越微技术有限公司 | Doppler radar signal kind identification method |
| CN110658506A (en) * | 2019-09-10 | 2020-01-07 | 武汉大学 | Micro Doppler clutter filtering method based on angle clustering and Doppler analysis |
Family Cites Families (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7567203B2 (en) * | 2005-04-11 | 2009-07-28 | Raytheon Canada Limited | Classification system for radar and sonar applications |
| US7965148B2 (en) * | 2009-08-03 | 2011-06-21 | Northrop Grumman Guidance And Electronics Company, Inc. | Atomic frequency clock systems and methods |
| CN103117823B (en) * | 2013-01-28 | 2015-02-11 | 中国电子科技集团公司第二十二研究所 | Short wave channel model building method |
| CN104020451B (en) * | 2014-06-03 | 2017-02-15 | 西安电子科技大学 | Outer transmitter-based radar target track processing method based on clustering |
| US10317521B2 (en) * | 2016-06-16 | 2019-06-11 | The United States Of America As Represented By The Administrator Of The National Aeronautics And Space Administration | Frequency diversity pulse pair determination for mitigation of radar range-doppler ambiguity |
| CN107121677B (en) * | 2017-06-02 | 2019-10-11 | 太原理工大学 | Avoidance radar method and device based on ultra wide band cognition CPPM signal |
| CN109917347B (en) * | 2019-04-10 | 2021-09-24 | 电子科技大学 | Radar pedestrian detection method based on time-frequency domain sparse reconstruction |
| CN110188628A (en) * | 2019-05-13 | 2019-08-30 | 北京遥感设备研究所 | A kind of RD image ship seakeeping method based on deep learning model |
-
2020
- 2020-01-15 CN CN202010042315.8A patent/CN111045008B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110488278A (en) * | 2019-08-20 | 2019-11-22 | 深圳锐越微技术有限公司 | Doppler radar signal kind identification method |
| CN110658506A (en) * | 2019-09-10 | 2020-01-07 | 武汉大学 | Micro Doppler clutter filtering method based on angle clustering and Doppler analysis |
Non-Patent Citations (1)
| Title |
|---|
| 基于特征辅助的距离高分辨雷达航迹关联算法;刘海波;刘洋;王额;;北京理工大学学报(第09期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111045008A (en) | 2020-04-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111045008B (en) | Vehicle millimeter wave radar target identification method based on widening calculation | |
| CN111316128A (en) | Continuous obstacle detection method, device, system and storage medium | |
| Li et al. | An adaptive 3D grid-based clustering algorithm for automotive high resolution radar sensor | |
| CN113009441B (en) | Method and device for identifying multipath target of radar moving reflecting surface | |
| CN113219462B (en) | Target identification method and device based on time-frequency diagram and terminal equipment | |
| CN113076922A (en) | Object detection method and device | |
| CN114609609A (en) | Speed estimation method for extracting static point cloud by FMCW laser radar random sampling | |
| CN108693517B (en) | Vehicle positioning method and device and radar | |
| US20230139751A1 (en) | Clustering in automotive imaging | |
| CN114137505A (en) | Target detection method and device based on wireless radar | |
| CN114296095A (en) | Method, device, vehicle and medium for extracting effective target of automatic driving vehicle | |
| CN106772357B (en) | AI-PHD filter multi-object tracking method under signal-to-noise ratio unknown condition | |
| Kim et al. | Deep-learning based multi-object detection and tracking using range-angle map in automotive radar systems | |
| CN112689773B (en) | Radar signal processing method and radar signal processing device | |
| JP3407045B2 (en) | Millimeter-wave sensor signal processing method | |
| CN113625266A (en) | Method, device, storage medium and equipment for detecting low-speed target by using radar | |
| CN113534096A (en) | LiDAR signal geometric feature extraction method and system based on spline function | |
| CN105242272A (en) | Vehicle-mounted millimeter wave crashproof radar constant false alarm detection method based on autoregression time sequence model | |
| Wang et al. | Classification and tracking of moving objects from 77 GHz automotive radar sensors | |
| CN113687348B (en) | Pedestrian recognition method and device based on tracking micro Doppler graph | |
| CN110782475A (en) | Multipath component processing method, terminal and computer storage medium | |
| CN116879863B (en) | Multi-target measuring method and system for continuous wave 4D millimeter wave radar | |
| CN113050056B (en) | Radar working environment identification method and device and terminal equipment | |
| CN116430349B (en) | Space group target detection method and system based on physical priori knowledge | |
| CN116224280B (en) | Radar target detection method, radar target detection device, radar equipment and storage medium |
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 | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20230712 Address after: 518000 107, building 37, chentian Industrial Zone, chentian community, Xixiang street, Bao'an District, Shenzhen, Guangdong Province Patentee after: Shenzhen Huaxun ark Intelligent Information Technology Co.,Ltd. Address before: 518102 East, 2nd floor, building 37, chentian Industrial Zone, Baotian 1st Road, Xixiang street, Bao'an District, Shenzhen City, Guangdong Province Patentee before: SHENZHEN HUAXUN FANGZHOU MICROELECTRONIC SCIENCE & TECHNOLOGY Co.,Ltd. Patentee before: CHINA COMMUNICATION TECHNOLOGY Co.,Ltd. |
|
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20231201 Address after: Building 7-6, Industrial Control Cloud Creation Port, No. 58 Hongtu Avenue, Honggutan District, Nanchang City, Jiangxi Province, 330000 Patentee after: Jiangxi Huaxun Fangzhou Intelligent Technology Co.,Ltd. Address before: 518000 107, building 37, chentian Industrial Zone, chentian community, Xixiang street, Bao'an District, Shenzhen, Guangdong Province Patentee before: Shenzhen Huaxun ark Intelligent Information Technology Co.,Ltd. |