JP7173735B2 - Radar device and signal processing method - Google Patents

Description

The present invention relates to a radar device and signal processing method for detecting a target.

Pedestrian recognition technology using an in-vehicle camera has been studied. However, pedestrian recognition technology using in-vehicle cameras has the following drawbacks: (1) When monocular cameras are used, the distance to pedestrians cannot be measured. (2) It is weak against scenes such as fog, snow, and backlight, and has low environmental robustness.

Therefore, in recent years, as a substitute technology for pedestrian recognition technology using in-vehicle cameras, research has been conducted on pedestrian recognition technology using a radar device that has high accuracy in measuring distances to targets and high environmental robustness.

“Pedestrian recognition in automotive radar sensors,” Radar Symposium (IRS), 2013 14th International

When a pedestrian's Doppler frequency is detected by a radar device, it is known that Doppler fluctuations due to swinging of arms and legs occur around the central Doppler component, which is detected mainly by reflected waves from the body. This is called the micro-Doppler effect.

For example, the pedestrian recognition technology using a radar device disclosed in Non-Patent Document 1 is a pedestrian recognition technology that utilizes the above-mentioned micro-Doppler effect, and bundles detection peak information that occurs instantaneously in a short distance. Pedestrians and moving vehicles are classified based on feature quantities such as distance spread and speed spread in the group information.

However, in the pedestrian recognition technology using the radar device disclosed in Non-Patent Document 1, mainly when the target exists in the distance, only one detection peak information in the group information can be obtained, and the distance There is a problem that feature quantities such as spread and velocity spread cannot be calculated.

SUMMARY OF THE INVENTION In view of the above problems, an object of the present invention is to provide a radar apparatus and a signal processing method capable of identifying the type of a distant target.

A radar apparatus according to the present invention includes: a derivation unit that derives target data related to the target based on a received signal obtained by receiving a reflected wave of a transmitted wave reflected by a target; a clustering unit that clusters the target data into clusters based on a predetermined measurement value; and a first feature value based on time-series information obtained by time-series acquisition of the target data belonging to the same cluster. and an identification unit that identifies the type of the target based on the first feature amount (first configuration).

In the radar device having the first configuration, the extraction unit extracts a second feature amount that is an instantaneous feature amount of the target data belonging to the same cluster, and the identification unit extracts the first and the second feature amount to identify the type of the target (second configuration).

In the radar device having the first or second configuration, the extraction unit extracts a second feature amount that is an instantaneous feature amount of the target data belonging to the same cluster, and the extraction unit extracts the A configuration (a third configuration) may be employed in which the identification unit identifies the type of the target based only on the second feature amount during a period in which the first feature amount cannot be extracted.

In the radar device having the first or second configuration, the identification section may not perform identification during a period when the extraction section cannot extract the first feature amount (fourth configuration).

In the radar device having any one of the first to fourth configurations, the extraction unit extracts a plurality of types of feature amounts as the first feature amount, and the identification unit identifies the target by supervised machine learning. A configuration (fifth configuration) for identifying the type may be used.

In the radar device of the fifth configuration, the extraction unit extracts at least two types of feature quantities, ie, the power of the received signal and the distance to the target, as the first feature quantity (the sixth configuration).

In the radar device having any one of the first to sixth configurations, a tracking processing unit that tracks the target is provided, and used in the tracking processing unit according to the type of the target identified by the identification unit A configuration (seventh configuration) in which the value of the parameter is changed may be used.

A signal processing method according to the present invention includes a deriving step of deriving target data related to the target based on a received signal obtained by receiving a reflected wave of a transmitted wave reflected by a target; a clustering step of clustering the target data into clusters based on a predetermined measurement value of; A configuration (eighth configuration) comprising an extraction step of extracting an amount, and an identification step of identifying the type of the target based on the first feature amount.

According to the radar device and the signal processing method of the present invention, the first feature amount can be detected even for a distant target, so the type of the distant target can be identified.

A diagram showing a configuration example of a radar device A diagram showing frequency modulation applied to a transmission signal Flowchart showing an operation example of the signal processing device A diagram showing the result of Fourier transform processing in the range bin direction A diagram showing the result of Fourier transform processing in the direction of velocity bins Schematic diagram showing outline of tracking process using αβ filter Flowchart showing an example of identification processing Schematic diagram showing a configuration example of a window Schematic diagram showing other configuration examples of windows Schematic diagram showing the outline of the sliding window method Flowchart showing another example of identification processing Schematic diagram showing an overview of supervised machine learning Diagram showing distance characteristics of received power

Exemplary embodiments of the invention are described in detail below with reference to the drawings.

<1. Configuration of Radar Device>
FIG. 1 is a diagram showing the configuration of a radar device 1 according to this embodiment. The radar device 1 is mounted on a vehicle such as an automobile. Hereinafter, the vehicle on which the radar device 1 is mounted is referred to as "own vehicle". In addition, the direction from the driver's seat to the steering wheel, which is the straight traveling direction of the own vehicle, is referred to as "forward". In addition, the direction from the steering wheel to the driver's seat, which is the straight traveling direction of the own vehicle, is called "rear". Also, a direction that is perpendicular to the straight traveling direction and the vertical line of the own vehicle and that extends from the right side to the left side of the driver who is facing forward is referred to as the "left direction." Also, the direction perpendicular to the straight traveling direction and the vertical line of the own vehicle and from the left side to the right side of the driver who is facing forward is referred to as the "right direction."

When the radar device 1 is mounted at the front end of the own vehicle, the radar device 1 acquires target object data related to a target existing in front of the own vehicle using transmission waves.

The radar device 1 measures the distance [m] until the reflected wave reflected from the target is received by the receiving antenna of the radar device 1, the relative speed of the target with respect to the vehicle [km/h], the front and rear direction of the vehicle The distance from the radar device 1 to the target (hereinafter referred to as “vertical position”) [m], the distance from the radar device 1 to the target in the lateral direction of the vehicle (hereinafter referred to as “lateral position”) [m ] to derive target data with parameters such as The vertical position is represented by a positive value in front of the own vehicle and a negative value in the rear of the own vehicle, with the position where the radar device 1 of the own vehicle is mounted as the origin O, for example. The lateral position is represented by a positive value on the right side of the own vehicle and a negative value on the left side of the own vehicle, with the position where the radar device 1 of the own vehicle is mounted as the origin O, for example.

The radar device 1 detects a target by an FCM (Fast Chirp Modulation) method. Since the FCM method does not require the up-peak and down-peak pairing processing required in the FMCW method, the problem of false recognition of targets due to incorrect pairing does not occur, and more accurate target detection can be expected. .

Here, a method for calculating the distance and relative velocity in the FCM method will be briefly described. In the FCM method, one chirp is one waveform of a transmission wave whose frequency changes like a sawtooth wave. A beat signal is obtained by taking the difference between the received signal and the transmitted wave, and the distance and relative velocity to the target are obtained by performing a two-dimensional FFT on this beat signal. Specifically, the frequency of the beat signal is proportional to the distance because the longer the target is, the greater the time delay of the received signal with respect to the transmitted wave. Therefore, by subjecting the beat signal to FFT processing, a peak appears at the position of the frequency corresponding to the distance to the target. FFT extracts the reception level and phase information for each frequency point (hereinafter sometimes referred to as range bin) set at a predetermined frequency interval. A peak appears. Therefore, the distance to the target can be obtained by detecting the peak frequency. Since the FFT processing for obtaining this distance is performed for each beat signal, it is repeated by the number of beat signals, that is, the number of chirps.

Next, regarding the calculation of the relative velocity, in the FCM method, when the relative velocity of the target is not 0 km/h, the Doppler frequency is detected using the phase change that appears between the beat signals according to the Doppler frequency. and calculate the relative velocity. That is, when there is almost no speed difference between the vehicle and the target (for example, when the relative speed is 0 km/h), no Doppler component is generated in the received signal, so the phase of the received signal for each chirp is will all be the same. However, when there is a speed difference with the target (for example, when the relative speed is other than 0 km/h and there is a relative speed), the phase change according to the Doppler frequency between the received signals for each chirp occurs. Since this phase information is included in the peak information obtained by FFT processing the beat signal, if the peak information of the same target obtained from each beat signal is arranged in time series and the second FFT processing is performed, , the Doppler frequency is obtained from the phase information, and a peak appears at that frequency position. This peak frequency corresponds to the relative velocity.

By performing two-dimensional FFT on the beat signal in this manner, the distance and the relative velocity can be calculated. The FFT processing for obtaining the relative velocity is repeated for the number of range bins because the result of the first FFT processing is performed for each range bin.

As shown in FIG. 1, the radar device 1 mainly includes a transmitter 2, a receiver 3, and a signal processor 4. FIG.

The transmitter 2 includes a signal generator 21 and a transmitter 22 . The signal generation unit 21 generates a modulation signal having a plurality of continuous waveforms (one chirp) in which the voltage decreases at a constant rate from the reference value over time, reaches the minimum value, and immediately returns to the reference value, It feeds the transmitter 22 . The period of one chirp may be, for example, several tens of microseconds. The transmitter 22 frequency-modulates the continuous wave signal based on the modulated signal generated by the signal generator 21 , generates a transmission signal whose frequency changes over time, and outputs the transmission signal to the transmission antenna 23 .

The transmission antenna 23 outputs a transmission wave TW ahead of the own vehicle based on the transmission signal from the transmitter 22 . In the transmission wave TW output from the transmission antenna 23, as shown in FIG. 2, an idle time t0, which is a period during which TFT processing is not performed, is provided between each chirp. A transmission wave TW transmitted from the transmission antenna 23 to the front of the own vehicle is reflected by an object such as a person or another vehicle to become a reflected wave RW.

The receiving section 3 includes a plurality of receiving antennas 31 forming an array antenna and a plurality of individual receiving sections 32 connected to the plurality of receiving antennas 31 . In this embodiment, the receiver 3 includes, for example, four receiving antennas 31 and four individual receivers 32 . The four individual receivers 32 correspond to the four receiving antennas 31, respectively. Each receiving antenna 31 receives a reflected wave RW from an object and acquires a received signal, and each individual receiving section 32 processes the received signal obtained by the corresponding receiving antenna 31 .

Each individual receiver 32 includes a mixer 33 and an A/D converter 34 . A received signal obtained by the receiving antenna 31 is sent to the mixer 33 after being amplified by a low-noise amplifier (not shown). A transmission signal from the transmitter 22 of the transmission unit 2 is input to the mixer 33 , and the transmission signal and the reception signal are mixed in the mixer 33 . As a result, a beat signal having a beat frequency that is the difference between the frequency of the transmission signal and the frequency of the reception signal is generated. The beat signal generated by the mixer 33 is converted to a digital signal by the A/D converter 34 and then output to the signal processing device 4 .

The signal processing device 4 includes a microcomputer including a CPU (Central Processing Unit), a memory 41, and the like. The signal processing device 4 stores various data to be operated on in the memory 41, which is a storage device. The memory 41 is, for example, a RAM (Random Access Memory). The signal processing device 4 includes a transmission control section 42, a Fourier transform section 43, and a data processing section 44 as functions realized by software on a microcomputer. The transmission controller 42 controls the signal generator 21 of the transmitter 2 .

The Fourier transform unit 43 performs fast Fourier transform (FFT) on beat signals output from each of the plurality of individual receivers 32 . As a result, the Fourier transform unit 43 transforms the beat signal associated with the reception signal of each of the plurality of reception antennas 31 into a frequency spectrum, which is data in the frequency domain. The frequency spectrum obtained by Fourier transform section 43 is input to data processing section 44 .

The data processing unit 44 executes target data acquisition processing, and acquires target data related to targets in front of the vehicle based on the frequency spectrum of each of the plurality of receiving antennas 31 . The data processing unit 44 also outputs the target object data to the vehicle control ECU 51 and the like.

As shown in FIG. 1, the data processing unit 44 has a target data derivation unit 45, a target data processing unit 46, and a target data output unit 47 as main functions.

The target data derivation unit 45 derives target data related to the target based on the frequency spectrum obtained by the Fourier transform unit 43 .

The target data processing unit 46 performs various processes such as clustering on the derived target data. The target data processing unit 46 includes a clustering unit 46a, an extraction unit 46b, an identification unit 46c, and a tracking processing unit 46d. Details of the processing executed by the clustering unit 46a, the extraction unit 46b, the identification unit 46c, and the tracking processing unit 46d will be described later.

The target data output unit 47 outputs target data to the vehicle control ECU 51 and the like. Accordingly, the vehicle control ECU 51 and the like can use the target object data for ACC (Adaptive Cruise Control) and PCS (Pre-crash Safety System), for example.

<2. Operation of Signal Processing Device>
Next, the operation of the signal processing device 4 will be described. FIG. 3 is a flow chart showing the operation of the signal processing device 4. As shown in FIG. The signal processing device 4 periodically repeats the processing shown in FIG. 3 at regular time intervals.

The signal processing device 4 acquires a predetermined number of beat signals (step S10). Next, the Fourier transform unit 43 executes two-dimensional FFT on the beat signal (step S20).

By subjecting the beat signal to FFT processing, a peak appears at the position of the frequency corresponding to the distance to the target. Since the FFT extracts the reception level and phase information for each range bin set at predetermined frequency intervals, a peak appears in the frequency range bin corresponding to the target distance. Therefore, the distance to the target can be obtained by detecting the peak frequency. Since the FFT processing for obtaining this distance is performed for each beat signal, it is repeated by the number of beat signals, that is, the number of chirps. FIG. 4 shows the results of this FFT processing arranged in the direction of the range bins, the results of each beat signal arranged in the direction perpendicular to the range bins in a matrix form, and the value of each processing result (Spectrum [dB]) in the height direction. This is an example. FIG. 4 shows an example in which two peaks 61 and 62 occur.

Next, regarding the calculation of the relative velocity, in the FCM method, when the relative velocity of the target is not 0 km/h, the Doppler frequency is detected using the phase change that appears between the beat signals according to the Doppler frequency. and calculate the relative velocity. That is, if the relative velocity is 0 km/h, no Doppler component is generated in the received signal, and the phases of the received signals for each chirp are all the same. However, when there is a relative velocity between the object and the target, a phase change corresponding to the Doppler frequency occurs between the received signals for each chirp. Since this phase information is included in the peak information obtained by FFT processing the beat signal, if the peak information of the same target obtained from each beat signal is arranged in time series and the second FFT processing is performed, , the Doppler frequency is obtained from the phase information, and a peak appears at that frequency position. In this FFT processing, since phase information is extracted for each velocity bin set at predetermined frequency intervals according to velocity resolution, peaks appear in velocity bins of frequencies corresponding to the relative velocity of the target. Therefore, the relative velocity with respect to the target can be obtained by detecting the peak frequency. FIG. 5 shows the results of the second FFT processing arranged in the velocity bin direction, the results of the second FFT processing arranged in the range bin direction for each frequency point of distance in a matrix form, and each processing result in the height direction. It is an example showing a value (Spectrum [dB]). FIG. 5 shows an example in which two peaks 63 and 64 occur.

Returning to FIG. 3, in step S30 following step S20, the target data derivation unit 45 extracts peaks from the result of the two-dimensional FFT. Furthermore, the target data deriving unit 45 estimates the azimuth (angle) of the target based on the received signal received via the receiving antenna 31 . Then, the target data deriving unit 45 calculates the distance and relative velocity of the target from the peak range bin and velocity bin extracted in step S30 (step S50).

In step S60 following step S50, the clustering unit 46a clusters the target data into clusters based on predetermined measurement values of the target data. Specifically, the clustering unit 46a joins the target data in which the target positions obtained from the vertical position and the horizontal position included in the target data are close to each other (within a predetermined distance) to form the same cluster. . One cluster corresponds to one target.

In step S70 following step S60, the tracking processing unit 46d performs tracking processing for tracking a target that has been detected in the past. The method of tracking processing is not particularly limited. Here, as an example, tracking processing using an αβ filter will be described. FIG. 6 is a schematic diagram showing an outline of tracking processing using an αβ filter.

A current predicted value (predicted position and predicted speed) V2 is calculated from the previous smoothed value (smoothed position and smoothed speed) V1. Next, a gate G1 is set around the current predicted value V2, and the observed value V3, which is a correlation candidate (tracking candidate), is limited within the gate G1. Of the observed values V3 in the gate G1, the observed value closest to the current predicted value V2 is selected as the current correlation point. The current smoothed value V4 is calculated by smoothing the current correlation point and a predetermined number of past correlation points. The size of the gate G1 and the like are determined by the parameter values of the αβ filter.

Returning to FIG. 3, in step S80 following step S70, the target data processing unit 46 selects the detected target as a stationary target, a moving target moving in the direction from the rear to the front, and a target moving in the direction from the front to the rear. They are classified into three types of moving targets that move in a direction.

Next, the target data processing unit 46 removes target data corresponding to unnecessary objects (step S90). Examples of unnecessary objects include stationary objects that are higher than a predetermined height (for example, higher than the height of the vehicle), stationary objects that are lower than the height of the vehicle, and the like. can.

Next, the identification unit 46c identifies the type of target (step S100). For example, the identification unit 46c identifies whether the target is a pedestrian or another vehicle. The details of the identification processing in step S100 will be described later.

Next, the target data processing unit 46 generates a plurality of target data that can be estimated to be target data related to the same object, based on parameters of the target data (distance, relative speed, vertical position, horizontal position, etc.). into one group (step S110).

Finally, the target data output unit 47 sends the target data thus processed to the vehicle control ECU 51 or the like. The target data output unit 47 selects a predetermined number (for example, 10) of target data as output targets from the grouped target data (step S120). The target data output unit 47 preferentially selects target data related to targets close to the own vehicle in consideration of the distance and lateral position of the target data.

The target data selected as an output target in the above process is stored in the memory 41 and used as past target data in subsequent target data acquisition processes.

<3. Details of Identification Processing>
FIG. 7 is a flowchart illustrating an example of identification processing. When the identification process is started, the extraction unit 46b extracts instantaneous feature amounts (hereinafter referred to as "cluster feature amounts") of target data belonging to the same cluster for each cluster (step S110). Cluster features include, for example, the spread of the distance to the target, the spread of the vertical position, the spread of the horizontal position, the spread of the relative velocity, the spread of the received signal power (received power), and the like.

Next, the extraction unit 46b accumulates time-series information obtained by time-series acquisition of target data belonging to the same cluster as a window (step S120). The number of clusters forming one window is arbitrary as long as it is plural. In the following description, it is assumed that the number of clusters forming one window is three.

Next, the identification unit 46c determines whether or not accumulation of the first window is completed (step S130).

If the accumulation of the first window has been completed (YES in step S130), the extraction unit 46b extracts the window-based feature amount (hereinafter referred to as "window feature amount") for the latest window (step S140). ). Window features include, for example, the spread of the distance to the target within the window, the spread of the vertical position within the window, the spread of the horizontal position within the window, the spread of the relative velocity within the window, and the spread of the relative velocity within the window. and the spread of the received power.

For example, the window W1 shown in FIG. 8 is composed of a first scan cluster C1 having one detected peak information, and a second scan cluster C2 and a third scan cluster C3 each having two detected peak information. ing. The spread of the relative velocity within the window W1 shown in FIG. 8 is 1.3 [m/s].

Further, for example, the window W2 shown in FIG. 9 is composed of a cluster C4 of the first scan, a cluster C5 of the second scan, and a cluster C6 of the third scan each having one detected peak information. The spread of the relative velocity within the window W2 shown in FIG. 9 is 1.0 [m/s].

In step S150 following step S140, the identification unit 46c identifies the type of target based on the cluster feature amount and the window feature amount. In addition, in step S150, the identifying unit 46c may identify the type of the target only based on the window feature amount.

In step S160 following step S150, the extraction unit 46b performs sliding shift of the window. That is, the extraction unit 46b employs a sliding window method, and by configuring a window with the latest predetermined number of clusters, the window is slid by one scan as shown in FIG. ) to accumulate windows. For example, if the number of clusters constituting a window is 3, and the nth scan cluster is Cn, the cluster Cn in the current scan, the cluster Cn-1 in the previous scan, and the cluster Cn-2 in the scan two times before. configure windows; In the sliding window method, a time delay is required until the first window is obtained. After that, window feature extraction timings t1 to t4 in windows W11 to W14 of the k to k+3 scans shown in FIG. , window features can be extracted at regular intervals.

On the other hand, if the accumulation of the first window has not been completed (NO in step S130), the identification unit 46c identifies the target type based only on the cluster feature amount (step S170).

When the process of step S160 or step S170 ends, the identification process ends.

According to the operation of the flowchart shown in FIG. 7 described above, after the accumulation of the first window is completed, the identification unit 46c identifies the type of target based on the window feature amount. As a result, even when clusters having only one detected peak information continue, as in the example shown in FIG. 9, the type of target can be identified. Therefore, the accuracy of identifying a distant target, such as a pedestrian, which has a small radar reflection cross section (RCS) and is difficult to obtain a large amount of detection peak information, is improved.

Further, according to the operation of the flow chart shown in FIG. 7 described above, after the accumulation of the first window is completed, the identification unit 46c identifies the target type based on not only the window feature amount but also the cluster feature amount. As a result, it is possible to improve the identification accuracy of the type of target when a cluster has a plurality of pieces of detected peak information.

Further, according to the operation of the flow chart shown in FIG. 7 explained above, until the accumulation of the first window is completed, the identification unit 46c identifies the type of the target based on the cluster feature amount. As a result, the identification unit 46c may be able to identify the target type even before the accumulation of the first window is completed.

It should be noted that the signal processing device 4 may perform the operations of the flowchart shown in FIG. 11 instead of the flowchart shown in FIG. The flowchart shown in FIG. 11 is a flowchart obtained by removing the process of step S170 from the flowchart shown in FIG. According to the operation of the flowchart shown in FIG. 11, the identification unit 46c does not identify the target type until the accumulation of the first window is completed. Accordingly, the identification processing by the identification unit 46c is reduced to step S150 only, and the identification processing by the identification unit 46c can be simplified.

Although there is no particular limitation on how the identification result of the identification unit 46c is used, for example, depending on the type of target identified by the identification unit 46c, the parameter values used by the tracking processing unit 46d are changed. you can As a result, it is possible to improve the tracking processing performance. Further, for example, the identification result of the identification unit 46c is provided to the outside of the radar device 1, and AEB (Advanced Emergency Braking System) control is executed according to the type of target identified by the identification unit 46c. The identification result may be displayed on a HUD (Head-Up Display) or the like.

In this embodiment, the identifying unit 46c identifies the type of target by supervised machine learning such as support vector machine and random forest. Accuracy of identification can be improved by using supervised machine learning. FIG. 12 is a schematic diagram showing an overview of supervised machine learning. As shown in FIG. 12, the identification unit 46c creates a learning result (parameters for forming a discrimination plane) from the feature amount of the teacher data in advance, and classifies the class (target target) from the estimation target feature amount extracted by online processing. type). In the identification processing of step S150, the feature amounts of the teacher data are the cluster feature amount and at least two types of window feature amounts, and the estimation target feature amount is also the cluster feature amount and at least two types of window feature amounts. In the identification processing of step S170, the feature amounts of the teacher data are at least two types of cluster feature amounts, and the estimation target feature amount is also at least two types of cluster feature amounts.

Note that in supervised machine learning, the received power (representative received power of a cluster or window) and the distance to a target (representative distance of a cluster or window) may be included in each of the feature amount of teacher data and the feature amount to be estimated. . Each of the received power and the distance to the target may be either a cluster feature amount or a window feature amount.

FIG. 13 is a diagram showing distance characteristics of received power. The thick line in FIG. 13 is the distance characteristic of the received power assuming a pedestrian as the target, and the thin line in FIG. 13 is the distance characteristic of the received power assuming another vehicle as the target. As is clear from FIG. 13, the value of the received power fluctuates depending on the distance to the target, so different targets (pedestrians and other vehicles) at different distances from the radar device 1 may have the same value. . Therefore, if an attempt is made to simply identify the type of target based on the received power, it is necessary to identify the type of target for each distance range divided by the dotted lines shown in FIG. . On the other hand, as described above, in supervised machine learning, by including the received power and the distance to the target in each of the feature amount of the teacher data and the feature amount to be estimated, the division by the dotted line shown in FIG. 41 can be lowered.

<4. Others>
Various modifications can be made to the various technical features disclosed in this specification without departing from the gist of the technical creation in addition to the above-described embodiments. In addition, multiple embodiments and modifications shown in this specification may be implemented in combination to the extent possible.

For example, instead of the FCM method in which the Doppler shift is detected as the phase change between multiple chirp signals rather than the frequency of the beat signal, another method is used in which the Doppler shift is detected as the phase change between multiple modulated signals rather than the frequency of the beat signal. may be adopted. Another method that detects the Doppler shift not as the frequency of the beat signal but as the phase change between multiple modulated signals is the pulse Doppler method that detects the Doppler shift as the phase change between multiple pulse signals instead of the frequency of the beat signal. be done.

Further, for example, in the radar apparatus, instead of the FCM method, the pulse Doppler method, or the like, the FMCW method (Frequency Modulated Continuous Wave) or the like may be adopted.

In the above-described embodiment, the vehicle-mounted radar system has been described, but the present invention can also be applied to an infrastructure radar system installed on a road or the like, an aircraft surveillance radar system, and the like.

1 radar device 2 transmitter 3 receiver 4 signal processor 46a clustering unit 46b extraction unit 46c identification unit 46d tracking processing unit

Claims (5)

a derivation unit for deriving target object data including position information , relative velocity information, and received power information of the target object based on a received signal obtained by receiving a reflected wave of the transmitted wave reflected by the target object;
a clustering unit that clusters the target data based on the position information included in the target data derived by the deriving unit;
a tracking processing unit that associates each current cluster, which is a cluster generated by clustering in the clustering unit, with the cluster generated in the previous processing cycle;
A group of the clusters grouped chronologically for each current cluster by the tracking processing unit as a cluster group, and a predetermined number of clusters including the current cluster in the cluster group as a window, for each cluster group, acquiring time-series information of the target data from the predetermined number of clusters belonging to the window, extracting one or more first feature amounts based on the time-series information; an extraction unit that extracts one or more second feature amounts based on the target data belonging to the cluster;
For each current cluster, the target associated with the current cluster based on at least one of the one or more first feature amounts and the one or more second feature amounts is a pedestrian. an identification unit that identifies the type of the target including whether it is a vehicle or a vehicle;
with
During a period in which the extraction unit cannot extract the one or more first feature quantities because the accumulation of the predetermined number of clusters necessary for forming the window is not completed, the identification unit performs the one or more Identifying the type of the target based only on the second feature amount;
radar equipment. The extraction unit extracts a plurality of feature amounts as the one or more first feature amounts,
The radar device according to claim 1, wherein the identification unit identifies the type of target by supervised machine learning. The extraction unit includes at least two feature quantities including spread of power of the received signal within the window and spread of distance to the target within the window as the one or more first feature quantities. 3. The radar system according to claim 2, for extracting . The radar device according to any one of claims 1 to 3, wherein a value of a parameter used by said tracking processing section is changed according to the type of said target identified by said identifying section. a derivation step of deriving target data including position information, velocity information and received power information of the target based on a received signal obtained by receiving a reflected wave of the transmitted wave reflected by the target;
a clustering step of clustering the target data based on the position information included in the target data derived in the deriving step;
a tracking step of matching each current cluster, which is a cluster generated by clustering in the clustering step, with the cluster generated in the previous processing cycle;
A group of the clusters chronologically grouped for each current cluster by the tracking step is defined as a cluster group, and a predetermined number of clusters including the cluster in the cluster group are defined as a window, and the window is set for each cluster group. acquire time-series information of the target data from the predetermined number of clusters belonging to the current cluster, extract one or more first feature amounts based on the time-series information, and for each current cluster, an extraction step of extracting one or more second feature amounts based on the target data belonging to
For each current cluster, the target associated with the current cluster based on at least one of the one or more first feature amounts and the one or more second feature amounts is a pedestrian. an identification step of identifying the type of the target, including whether it is a vehicle or a vehicle;
with
During the period in which the one or more first feature values cannot be extracted in the extraction step because the accumulation of the predetermined number of clusters required for forming the window is not completed, the identification step includes the one or more identifying the target type based only on the second feature of
Signal processing method.

Images