JP7156817B2 - Target detection device and target detection method - Google Patents

Description

The present invention relates to a target object detection device and a target object detection method.

2. Description of the Related Art Conventionally, there is known a target detection device that transmits radio waves around a vehicle, for example, and detects a target based on reflected waves of the transmitted radio waves that are reflected by the target. For example, there is a target detection device that generates target data by applying time-series filtering to instantaneous data representing reflection points of the target (see, for example, Patent Document 1). Further, in such time-series filtering, for example, by assigning the previous target data to the current instantaneous data, continuity with the previous target data is obtained.

JP 2015-210157 A

However, the conventional techniques described above have room for further improvement in terms of improving target detection accuracy. Specifically, conventionally, one piece of target data is assigned to one piece of instantaneous data. Therefore, if the number of pieces of instantaneous data is smaller than that of target data, the so-called target data may compete for the piece of instantaneous data. there were. If such a scramble occurs, the target data will be replaced, and there is a risk that the reliability of target detection will decrease.

SUMMARY OF THE INVENTION It is an object of the present invention to provide a target detection device and a target detection method capable of improving target detection accuracy.

In order to solve the above-described problems and achieve the object, the target detection device according to the present invention includes a generator and a filter processor. The generation unit generates instantaneous data corresponding to the target based on a reflected wave of the transmitted radio wave reflected by the target. The filter processor generates target object data corresponding to the target by applying time-series filtering to the instantaneous data generated by the generator. Further, the filter processing unit assigns the current instantaneous data to prediction data of the current target data based on the previous target data, thereby ensuring continuity with the previous target data. An allocation unit is provided. The allocation unit can allocate a plurality of the prediction data to one of the current instantaneous data.

According to the present invention, target detection accuracy can be improved.

FIG. 1A is a diagram showing an overview of a target detection method according to an embodiment. FIG. 1B is a diagram illustrating an outline of a target detection method according to the embodiment; FIG. 2 is a block diagram showing the configuration of the radar device according to the embodiment. FIG. 3 is a process explanatory diagram from pre-processing of the signal processing unit to peak extraction processing in the signal processing unit. FIG. 4A is a process explanatory diagram of the angle estimation process. FIG. 4B is a process explanatory diagram (part 1) of the pairing process. FIG. 4C is a process explanatory diagram (part 2) of the pairing process. FIG. 5 is a diagram illustrating multiple allocation processing by the allocation unit. FIG. 6 is a diagram showing processing for generating a representative value of target data. FIG. 7 is a flowchart illustrating a processing procedure of processing executed by the target object detection device according to the embodiment; FIG. 8 is a flowchart illustrating a procedure of allocation processing executed by an allocation unit according to the embodiment; FIG. 9 is a flowchart showing a processing procedure of representative value generation processing executed by the target object data generation unit.

Hereinafter, embodiments of a target object detection device and a target object detection method disclosed in the present application will be described in detail with reference to the accompanying drawings. It should be noted that the present invention is not limited by this embodiment. In the following description, the target detection device is based on the FM-CW (Frequency Modulated Continuous Wave) system. It may be a method.

First, the outline of the target object detection method according to the embodiment will be described with reference to FIGS. 1A and 1B. 1A and 1B are diagrams showing an outline of a target detection method according to an embodiment. FIG. 1A shows an own vehicle MC equipped with a target detection device 1 according to the embodiment, and another vehicle LC located in front of the own vehicle MC.

As shown in FIG. 1A, the target detection device 1 is mounted, for example, in the front grille of the own vehicle MC, and detects targets (for example, other vehicles LC, etc.) existing in the traveling direction of the own vehicle MC. Note that the target detection device 1 may be mounted at other locations such as a windshield, a rear grille, left and right side portions (for example, left and right door mirrors), and the like.

Further, as shown in FIG. 1A, the target detection device 1 generates instantaneous data 100 corresponding to the target on the basis of the reflected waves of the radio waves transmitted around the host vehicle MC reflected by the other vehicle LC (step S1). The instantaneous data 100 includes, for example, information such as the relative speed toward the own vehicle MC, the distance between the own vehicle MC and the instantaneous data 100, and the angle of the instantaneous data 100. FIG.

Further, the target detection device 1 generates target data corresponding to the target by applying time-series filtering to the generated instantaneous data 100 (step S2). For time-series filtering, for example, known time-series filtering such as a particle filter or a Kalman filter can be adopted. In the following description, the case where the time-series filtering is a particle filter will be described as an example.

FIG. 1B shows a process of ensuring continuity in time-series filtering. FIG. 1B shows target data 50a and 50b of the previous cycle (hereinafter referred to as previous target data) and prediction data of the current cycle predicted from the previous target data 50a and 50b (hereinafter referred to as current prediction). data) 60a and 60b, and instantaneous data of the current cycle (hereinafter referred to as current instantaneous data) 100. FIG. The details of the prediction process for predicting the current prediction data 60a and 60b from the previous target data 50a and 50b will be described later.

In the time-series filtering, by assigning prediction data 60a and 60b to current instantaneous data 100, continuity with previous target data 50a and 50 is ensured.

Here, a conventional target object detection method will be described. In the conventional target detection method, one predictive data is assigned to one instantaneous data. For this reason, when the number of instantaneous data is smaller than the number of predicted data, there is a possibility that so-called predicted data will compete for instantaneous data. If such a scramble occurs, there is a possibility that the prediction data assigned to the instantaneous data will be replaced, and there is a possibility that the reliability of target detection will be lowered. As described above, conventionally, there is room for further improvement in terms of improving target detection accuracy.

Therefore, in the target object detection method according to the embodiment, it is possible to assign a plurality of prediction data 60a and 60b to one instantaneous data 100. FIG. Specifically, the target detection device 1 according to the embodiment enables allocation of a plurality of prediction data 60a and 60b to one instantaneous data 100 of this time.

Then, the target detection device 1 according to the embodiment generates two current target data corresponding to the two prediction data 60a and 60b, respectively, based on the allocated single current instantaneous data 100. FIG.

This prevents the prediction data 60a and 60b from competing for the instantaneous data 100, thereby preventing the prediction data 60a and 60b from being replaced. That is, according to the target object detection method according to the embodiment, it is possible to improve the detection accuracy of the target object.

Note that the example shown in FIG. 1B shows the case where two prediction data 60a and 60b are assigned to one instantaneous data 100, but the number of prediction data 60a and 60b to be assigned to one instantaneous data 100 is three. There may be more than one.

Further, hereinafter, when the plurality of prediction data 60a and 60b, the plurality of target data 50a and 50b, and the allocation ranges Ra and Rb are not distinguished, they are referred to as the prediction data 60, the target data 50, and the allocation range R, respectively. Sometimes.

Next, the configuration of the target object detection device 1 according to the embodiment will be described in detail with reference to FIG. 2 . FIG. 2 is a block diagram showing the configuration of the target object detection device 1 according to the embodiment. In addition, in FIG. 2, only the components necessary for explaining the features of the present embodiment are represented by functional blocks, and the description of general components is omitted.

In other words, each component illustrated in FIG. 2 is functionally conceptual and does not necessarily need to be physically configured as illustrated. For example, the specific forms of distribution and integration of each functional block are not limited to those shown in the figure, and all or part of them can be functionally or physically distributed in arbitrary units according to various loads and usage conditions.・It is possible to integrate and configure.

As shown in FIG. 2 , the target detection device 1 includes a transmitter 10 , a receiver 20 and a processor 30 . The target detection device 1 is connected to a vehicle control device 2 that controls the behavior of the host vehicle MC.

The vehicle control device 2 performs vehicle control such as PCS (Pre-crash Safety System) and AEB (Advanced Emergency Braking System) based on the target detection result by the target detection device 1 .

The transmitter 10 includes a signal generator 11 , an oscillator 12 and a transmission antenna 13 . The signal generation unit 11 generates a modulation signal for transmitting millimeter waves frequency-modulated with a triangular wave under the control of the transmission/reception control unit 31, which will be described later. The oscillator 12 generates a transmission signal based on the modulated signal generated by the signal generator 11 and outputs it to the transmission antenna 13 . Incidentally, as shown in FIG. 2, the transmission signal generated by the oscillator 12 is also distributed to the mixer 22, which will be described later.

The transmission antenna 13 converts a transmission signal from the oscillator 12 into a transmission wave, and outputs the transmission wave to the outside of the own vehicle MC. A transmission wave output from the transmission antenna 13 is a continuous wave frequency-modulated with a triangular wave. A transmission wave transmitted from the transmission antenna 13 to the outside of the own vehicle MC, for example, to the front is reflected by a target such as another vehicle LC and becomes a reflected wave.

The receiving section 20 includes a plurality of receiving antennas 21 forming an array antenna, a plurality of mixers 22 and a plurality of A/D converting sections 23 . Mixer 22 and A/D converter 23 are provided for each receiving antenna 21 .

Each receiving antenna 21 receives a reflected wave from a target as a received wave, converts the received wave into a received signal, and outputs the received signal to the mixer 22 . Although the number of receiving antennas 21 shown in FIG. 2 is four, the number may be three or less or five or more.

A received signal output from the receiving antenna 21 is input to the mixer 22 after being amplified by an amplifier (for example, a low noise amplifier) not shown. The mixer 22 mixes the distributed transmission signal and part of the received signal input from the receiving antenna 21 to remove unnecessary signal components, generates a beat signal, and outputs the beat signal to the A/D converter 23 . .

The beat signal is a difference wave between the transmitted wave and the reflected wave, and is the difference between the frequency of the transmitted signal (hereinafter referred to as "transmission frequency") and the frequency of the received signal (hereinafter referred to as "received frequency"). It has a different beat frequency. The beat signal generated by the mixer 22 is converted into a digital signal by the A/D conversion section 23 and then output to the processing section 30 .

The processing unit 30 includes a transmission/reception control unit 31 , a signal processing unit 32 and a storage unit 33 . The signal processor 32 includes a generator 32a and a filter processor 32b.

The storage unit 33 stores history data 33a. The history data 33a is information including the history of the target data 50 and the history of the instantaneous data 100 in a series of signal processing executed by the signal processing unit 32 .

The processing unit 30 is a microcomputer including, for example, a CPU (Central Processing Unit), a ROM (Read Only Memory) and a RAM (Random Access Memory) corresponding to the storage unit 33, registers, other input/output ports, etc. It controls the target detection device 1 as a whole.

The CPU of such a microcomputer functions as a transmission/reception control section 31 and a signal processing section 32 by reading out and executing programs stored in the ROM. The transmission/reception control unit 31 and the signal processing unit 32 can also be configured entirely by hardware such as ASIC (Application Specific Integrated Circuit) or FPGA (Field Programmable Gate Array).

The transmission/reception control section 31 controls the transmission section 10 including the signal generation section 11 and the reception section 20 . The signal processing unit 32 periodically executes a series of signal processing. Next, each component of the signal processing section 32 will be described.

The generator 32 a generates instantaneous data 100 . Specifically, the generator 32a generates the instantaneous data 100 by performing frequency analysis processing, peak extraction processing, and instantaneous data generation processing.

In the frequency analysis process, the beat signal input from each A/D converter 23 is subjected to Fast Fourier Transform (FFT) processing (hereinafter referred to as "FFT processing"). The result of such FFT processing is the frequency spectrum of the beat signal, which is the power value (signal level) for each frequency of the beat signal (for each frequency bin set at frequency intervals corresponding to the frequency resolution).

In the peak extraction process, a peak frequency is extracted as a result of the FFT process by the frequency analysis process. In the peak extraction process, peak frequencies are extracted for each of the "UP section" and the "DN section" of the beat signal, which will be described later.

In the instantaneous data generation process, an angle estimation process is executed to calculate the arrival angle and the power value of the reflected wave corresponding to each of the peak frequencies extracted in the peak extraction process. It should be noted that, at the time of execution of the angle estimation process, the arrival angle is an angle at which it is estimated that the target exists, so it may be referred to as an "estimated angle" below.

Also, in the instantaneous data generation process, a pairing process is executed to determine the correct combination of the peak frequencies of the "UP section" and the "DN section" based on the calculated estimated angle and power value.

Further, in the instantaneous data generation process, the distance of each target from the own vehicle MC and the relative speed toward the own vehicle MC are calculated from the determined combination result. In the instantaneous data processing, the calculated estimated angle, distance, and relative velocity of each target are output as instantaneous data 100 for the latest cycle (latest scan) to the filter processing unit 32b, and the history data 33a of the storage unit 33 are output. remember as

3 to 4C show the flow of processing from the pre-processing of the signal processing unit 32 up to this point in the signal processing unit 32 for the sake of easy understanding of the explanation. FIG. 3 is a process explanatory diagram from the pre-processing of the signal processing unit 32 to the peak extraction processing in the generation unit 32a.

Also, FIG. 4A is a process explanatory diagram of the angle estimation process. 4B and 4C are process explanatory diagrams (part 1) and (part 2) of the pairing process. Note that FIG. 3 is divided into three regions by two thick downward white arrows. In the following, such regions are described in order as an upper stage, a middle stage, and a lower stage.

As shown in the upper part of FIG. 3, the transmission signal fs(t) is transmitted from the transmission antenna 13 as a transmission wave, is reflected by the target, arrives as a reflected wave, and is received by the reception antenna 21 as the reception signal fr(t ) is received as

At this time, as shown in the upper part of FIG. 3, the received signal fr(t) is delayed by the time difference T with respect to the transmitted signal fs(t) according to the distance between the host vehicle MC and the target. Due to this time difference T and the Doppler effect based on the relative velocities of the host vehicle MC and the target, the beat signal has a frequency fup in an "UP section" where the frequency rises and a frequency fdn in the "DN section" where the frequency drops. is obtained as a repeated signal (see the middle part of FIG. 3).

The lower part of FIG. 3 schematically shows the results of the FFT processing of the beat signal in the frequency analysis processing for each of the "UP section" side and the "DN section" side.

As shown in the lower part of FIG. 3, after the FFT processing, waveforms in the frequency domains on the "UP section" side and the "DN section" side are obtained. In the peak extraction process, the peak frequency of the waveform is extracted.

For example, in the case of the example shown in the lower part of FIG. 3, a peak extraction threshold is used, peaks Pu1 to Pu3 are determined as peaks, and peak frequencies fu1 to fu3 are extracted on the "UP section" side.

Also, on the "DN section" side, the peaks Pd1 to Pd3 are determined as peaks by the same peak extraction threshold, and the peak frequencies fd1 to fd3 are extracted.

Here, the frequency component of each peak frequency extracted by the peak extraction process may be mixed with reflected waves from a plurality of targets. Therefore, in the instantaneous data generation process, an angle estimation process for performing azimuth calculation is performed for each peak frequency, and the presence of a target corresponding to each peak frequency is analyzed.

Note that the azimuth calculation in the instantaneous data generation process can be performed using a known direction-of-arrival estimation technique such as ESPRIT (Estimation of Signal Parameters via Rotational Invariance Techniques).

FIG. 4A schematically shows an azimuth calculation result of instantaneous data generation processing. In the instantaneous data generation process, the estimated angle of each target (each reflection point) corresponding to each of the peaks Pu1 to Pu3 is calculated from each of the peaks Pu1 to Pu3 of the azimuth calculation result. Also, the magnitude of each of the peaks Pu1 to Pu3 is the power value. In the instantaneous data generation processing, as shown in FIG. 4B, such angle estimation processing is performed for each of the "UP section" side and the "DN section" side.

Then, in the instantaneous data generation process, a pairing process is performed in which peaks having similar estimated angles and power values are combined in the azimuth calculation result. Further, from the result of the combination, in the instantaneous data generation process, the distance of each target (each reflection point) corresponding to the combination of each peak and the relative speed toward the own vehicle MC are calculated.

The distance can be calculated based on the relationship “distance∝(fup+fdn)”. The relative velocity can be calculated based on the relationship "velocity∝(fup-fdn)". As a result, as shown in FIG. 4C, a pairing processing result showing instantaneous data 100 of the estimated angle, distance and relative speed of each reflection point RP with respect to the own vehicle MC is obtained.

Returning to FIG. 2, the filter processing section 32b will be described. As shown in FIG. 2, the filtering unit 32b includes a prediction unit 321b, an allocation unit 322b, a weighting unit 323b, a resampling unit 324b, and a target object data generation unit 325b.

The filtering unit 32b generates target object data 50 corresponding to the instantaneous data 100 by applying a particle filter that allocates a predetermined number of particle data to the instantaneous data 100 generated by the generating unit 32a.

The prediction unit 321b performs prediction processing of sample points (particle data) in the particle filter. Specifically, when the latest cycle is time t and the distribution state X _t of the particle data at time t, the prediction unit 321b sets the probability based on the distribution state X _t−1 at time t−1 of the previous cycle Arrange (sample) the N particle data based on the density function. That is, in the prediction process, the prediction unit 321b distributes the particle data from the particle data at time t−1 to a region where the instantaneous data 100 is likely to appear at time t.

Note that the prediction unit 321b distributes particle data in a predetermined distribution state for the instant data 100 corresponding to the new target, since there is no particle data for the previous period.

The prediction unit 321b also generates prediction data 60 corresponding to the current target data 50 based on the previous target data 50 . Specifically, the prediction unit 321b generates the prediction data 60 based on the actual moving direction of the previous target object data 50 and the relative speed to the moving direction.

The allocation unit 322b performs a process of allocating the instantaneous data 100 in the latest period to the latest particle data, which is the prediction result of the prediction unit 321b. Specifically, the allocation unit 322b assigns the current instantaneous data 100 within a predetermined allocation range R corresponding to the current prediction data 60 generated by the prediction unit 321b to the prediction data 60 and the prediction data 60 assign the particle data corresponding to

If there is instantaneous data 100 that does not exist within the allocation range R of any target data 50, the allocation unit 322b treats the instantaneous data 100 as a new target.

Also, the allocation unit 322b can allocate a plurality of prediction data 60 to one instantaneous data 100 of this time. Here, with reference to FIG. 5, the multiple allocation process will be specifically described.

FIG. 5 is a diagram showing multiple allocation processing by the allocation unit 322b. The upper part of FIG. 5 shows two predicted data 60a and 60b and four instantaneous data 100a to 100d. Of the four instantaneous data 100a-100d, three instantaneous data 100a-100c are within the allocation ranges Ra and Rb of the prediction data 60a and 60b, and one instantaneous data 100d is outside the allocation ranges Ra and Rb. Suppose there is.

Also, the lower part of FIG. 5 shows the cost value for each combination of the two prediction data 60a, 60b and the four instantaneous data 100a to 100d. A cost value is an index calculated by the allocation unit 322b, and is an evaluation value obtained by evaluating the relationship between the prediction data 60 and the instantaneous data 100. FIG. Specifically, the cost value is calculated based on, for example, the distance to the own vehicle MC, the angle, the relative speed, and the like. In the example shown in FIG. 5, the cost value is set to be lower for combinations of instantaneous data 100 and predicted data 60 having closer distances, angles, and relative velocities. Since the instantaneous data 100d is outside the allocation ranges Ra and Rb of the prediction data 60a and 60b, the cost value is set to the maximum value, which is the upper limit.

As shown in the lower part of FIG. 5, the allocation unit 322b first allocates the instantaneous data 100 to the prediction data 60 according to the first allocation condition based on the cost value. Specifically, the allocation unit 322b allocates the instantaneous data 100 so that the prediction data 60 with the lowest cost value satisfies the first allocation condition. In the example shown in the lower part of FIG. 5, the allocation unit 322b allocates the prediction data 60a to the three instantaneous data 100a to 100c under the first allocation condition.

That is, under the first allocation condition, none of the instantaneous data 100a to 100d is allocated to the predicted data 60b. In this case, the allocation unit 322b assigns the prediction data 60b, which has not been allocated to the instantaneous data 100a to 100c under the first allocation condition, to the instantaneous data 100a to 100c under the second allocation condition.

Specifically, the allocation unit 322b redundantly allocates the instantaneous data 100a having the lowest cost value among the cost values of the prediction data 60b so as to satisfy the second allocation condition. That is, the allocation unit 322b allocates two prediction data 60a and 60b to one instantaneous data 100a.

In this way, by performing multiple assignment processing based on the cost value considering the distance, angle, and relative speed of the instantaneous data 100 and the predicted data 60, a plurality of predicted data 60 are appropriately assigned to one instantaneous data 100. be able to. In addition, the instantaneous data 100 can be allocated to all of the prediction data 60 by performing the allocation process under the second allocation condition for the prediction data 60 not allocated under the first allocation condition.

Returning to FIG. 2, the weighting unit 323b will be described. The weighting unit 323b weights each of the particle data corresponding to the current instantaneous data 100 and the current prediction data 60 in the allocation relationship by the allocation unit 322b.

For example, the weighting unit 323b increases the weight of particles close to the instantaneous data 100 of this time, and decreases the weight of particles far from the instantaneous data 100 of this time, among the particle data of this time. Note that "near" and "distant" here refer to "near" and "distant" Mahalanobis distances.

Further, when a plurality of prediction data 60 are assigned to one instantaneous data 100, the weighting unit 323b weights particle data of each of the plurality of prediction data 60 based on the single instantaneous data 100. FIG.

Next, the resampling unit 324b rearranges (resamples) the particle data based on the respective weights of the current particle data. Specifically, the resampling unit 324b moves the particle data with the smaller weight closer to the instantaneous data 100 (with the larger weight). More specifically, the resampling unit 324b rearranges particle data whose weight is less than a predetermined threshold into particle data whose weight is greater than or equal to a predetermined threshold. As a result, the target data 50 generated by the target data generation unit 325b, which will be described later, can be brought closer to the instantaneous data 100 than the predicted data 60 is.

Note that, when a plurality of prediction data 60 are assigned to one instantaneous data 100, the resampling unit 324b determine the threshold for Specifically, the resampling unit 324b determines the threshold value based on the Mahalanobis distance between each of the plurality of prediction data 60 and the instantaneous data 100. FIG.

For example, the resampling unit 324b lowers the above-described threshold for prediction data 60 with a long Mahalanobis distance from the instantaneous data 100, and raises the above-described threshold for prediction data 60 with a short Mahalanobis distance. For example, the resampling unit 324b may determine how much the threshold should be lowered (or raised) according to the difference in the Mahalanobis distances of the multiple pieces of prediction data 60 .

That is, the resampling unit 324b makes the target data 50 corresponding to the prediction data 60 with a short Mahalanobis distance closer to the instantaneous data 100, and the target data 50 corresponding to the prediction data 60 with a long Mahalanobis distance , makes the instantaneous data 100 less accessible.

As a result, for example, when the noise component included in the instantaneous data 100 is relatively large, all of the plurality of target data 50 generated by the target object data generation unit 325b, which will be described later, come close to the instantaneous data 100. can be prevented, the continuity of the target object data 50 can be prevented from becoming unstable.

The target data generation unit 325b recalculates the probability density function based on the current particle data rearranged by the resampling unit 324b, and generates target data 50 based on the recalculated center of gravity of the probability density function. do. The target data generator 325b generates the target data 50 based on the center of gravity of the probability density function, but may generate the target data 50 based on the average of the probability density functions, for example.

Also, the target data generator 325b treats the instantaneous data 100 to which no particle data has been assigned as a new target, and outputs it as it is as target data. Namely. In the case of a new target, the target data generator 325b outputs instantaneous data 100=target data.

Further, when a plurality of prediction data 60 are assigned to one piece of instantaneous data 100, the target data generator 325b generates target data 50 corresponding to each of the plurality of prediction data 60. FIG. Thereby, even if the number of prediction data 60 is larger than that of instantaneous data 100, continuity can be ensured for all prediction data 60. FIG.

In addition, when a plurality of prediction data 60 are assigned to one instantaneous data 100, the target data generation unit 325b generates a plurality of targets corresponding to each of the plurality of prediction data 60 when a predetermined condition is satisfied. For the data 50, a representative value of the target data 50 is generated. Then, the target object data generator 325b outputs the representative value to the vehicle control device 2. FIG.

Here, the generation processing of the representative value of the target data 50 by the target data generation unit 325b will be described with reference to FIG. FIG. 6 is a diagram showing the process of generating the representative value of the target data 50. As shown in FIG.

In FIG. 6, generation processing of target object data 50 at time t corresponding to the latest cycle, time t-1 immediately before time t, and time t-2 immediately before time t-1 is performed. showing. Also, FIG. 6 shows a case where one instantaneous data 100 is continuous with a plurality of identical target data 50a, 50b (prediction data 60a, 60b) over the past two cycles. It should be noted that the parenthesis of each symbol indicates the corresponding time. That is, the instantaneous data 100(t) indicates the instantaneous data 100 obtained at time t.

As shown in FIG. 6, the target data generation unit 325b generates a plurality of , a representative value 50ab of the target data 50a and 50b is generated. For example, the target data generator 325b may set the average value of the plurality of target data 50a and 50b as the representative value 50ab, or may set one of the plurality of target data 50a and 50b as the representative value 50ab.

In other words, the target data generation unit 325b increases the probability count that the target is the same according to the number of consecutive cycles of the allocation relationship between the instantaneous data 100 and the plurality of prediction data 60, and the probability count reaches a predetermined threshold value. In this case, it is assumed that there is a high probability that the plurality of prediction data 60 are derived from the same target, and the generated plurality of target data 50 are grouped together as one representative value 50ab. Thereby, it is possible to avoid continuous generation of multiple target data 50 for the same target.

Note that the target data generator 325b may adjust the range of increase in the probability count according to the state of the target data 50, for example. For example, the target data generation unit 325b does not increase the probability count, or does not increase the probability count when the respective movement vectors (direction of movement and relative speed) of the plurality of prediction data 60 assigned to one instantaneous data 100 are different. Minimize the number of counts. That is, when the motion vectors of the plurality of prediction data 60 are different, there is a high possibility that they are different targets, so even if one instantaneous data 100 is shared, the representative value 50ab is not generated. This can prevent erroneous generation of the representative value 50ab from a plurality of target data 50 derived from different targets.

Note that the target data generation unit 325b may increase the number of increments of the probability count more than usual when the movement vectors of the plurality of prediction data 60 assigned to one instantaneous data 100 are similar. . As a result, the plurality of target object data 50 can be quickly collected into the representative value 50ab, so that the same target object can be detected more quickly.

Further, for example, when the past positions of a plurality of target data 50 are separated by a predetermined distance or more, the target data generation unit 325b does not increase the probability count even if one piece of instantaneous data 100 is shared. to negligible. This can prevent erroneous generation of the representative value 50ab when, for example, different targets come close to each other and share one instantaneous data 100 .

Next, a processing procedure of processing executed by the target object detection device 1 according to the embodiment will be described with reference to FIG. 7 . FIG. 7 is a flowchart showing a processing procedure of processing executed by the target object detection device 1 according to the embodiment.

As shown in FIG. 7, first, the generation unit 32a generates instantaneous data 100 corresponding to the target based on the reflected wave of the transmitted radio wave reflected by the target (step S101).

Subsequently, the prediction unit 321b of the filter processing unit 32b performs prediction processing for predicting the current particle data based on the previous particle data, and prediction corresponding to the current target data 50 based on the previous target data 50. Prediction processing for predicting data 60 is performed (step S102). In addition, when the previous particle data does not exist, the filter processing unit 32b predicts the current particle data based on the relative velocity of the previous instantaneous data 100 .

Subsequently, the allocation unit 322b allocates the current instantaneous data 100 to the current particle data (step S103). A specific processing procedure of the allocation unit 322b will be described later with reference to FIG.

Subsequently, the assignment unit 322b determines whether there is a new target based on the presence or absence of instantaneous data 100 to which the particle data of this time has not been assigned (step S104). If the instantaneous data 100 is a new target (step S104, Yes), the assigning unit 322b assigns predetermined particle data (for example, initial state particle data) to the instantaneous data 100 corresponding to the new target. is set (step S105).

Subsequently, if the instantaneous data 100 is not a new target (step S104, No), the weighting unit 323b weights the current particle data based on the instantaneous data 100 (step S106).

Subsequently, the resampling unit 324b resamples the current particle data based on the weighting by the weighting unit 323b (step S107). Subsequently, the target data generation unit 325b updates the probability density function of the current resampled particle data, generates the target data 50 based on the probability density function (step S108), and ends the process. . Note that the generation processing of the representative value 50ab of the target data 50 by the target data generation unit 325b will be described later with reference to FIG.

Next, a processing procedure of allocation processing executed by the allocation unit 322b according to the embodiment will be described with reference to FIG. FIG. 8 is a flowchart illustrating the procedure of allocation processing executed by the allocation unit 322b according to the embodiment.

As shown in FIG. 8, first, the allocation unit 322b calculates a cost value for each of the prediction data 60 and the instantaneous data 100 existing within the allocation range R (step S201).

Subsequently, the allocation unit 322b performs allocation processing according to the first allocation condition based on the calculated cost value (step S202). For example, the allocation unit 322b allocates the prediction data 60 with the lowest cost value as the prediction data 60 that satisfies the first allocation condition for each piece of instantaneous data 100 .

Then, the allocation unit 322b determines whether there is any prediction data 60 that cannot be allocated under the first allocation condition (step S203). If there is prediction data 60 that cannot be allocated under the first allocation condition (step S203, Yes), the allocation unit 322b performs redundant allocation under the second allocation condition (step S204), and ends the allocation process. .

On the other hand, if there is no predictive data 60 that cannot be assigned according to the first assignment condition (step S203, No), the assigning unit 322b terminates the assigning process because assignment of all predictive data 60 has been completed.

Next, a processing procedure of generation processing of the representative value 50ab executed by the target object data generation unit 325b according to the embodiment will be described with reference to FIG. FIG. 9 is a flow chart showing a processing procedure of generation processing of the representative value 50ab executed by the target object data generation unit 325b.

As shown in FIG. 9, the target data generation unit 325b determines whether or not the allocation unit 322b has allocated a plurality of prediction data 60 to one instantaneous data 100 (step S301).

When a plurality of prediction data 60 are assigned to one instantaneous data 100 (step S301, Yes), the target data generator 325b determines that the combination of the instantaneous data 100 and the plurality of prediction data 60 in the assignment relationship is the same. A probability count indicating the probability of being a target is incremented (step S302).

Subsequently, the target object data generation unit 325b determines whether or not the probability count exceeds a predetermined threshold (step S303). When the probability count is greater than or equal to the predetermined threshold (step S303, Yes), the target data generator 325b generates a representative value 50ab of the target data 50 (step S304), and ends the process.

On the other hand, in step S301, when the target data generator 325b fails to allocate a plurality of prediction data 60 to one instantaneous data 100 (step S301, No), the target data generation unit 325b generates target data corresponding to each of the prediction data 60. 50 is generated (step S305), and the process ends.

Also, in step S303, the target data generation unit 325b shifts the process to step S305 when the probability count is less than the predetermined threshold (step S303, No).

As described above, the target detection device 1 according to the embodiment includes the generator 32a and the filter processor 32b. The generation unit 32a generates instantaneous data 100 corresponding to the target based on the reflected wave of the transmitted radio wave reflected by the target. The filtering unit 32b generates target object data 50 corresponding to the target by applying time-series filtering to the instantaneous data 100 generated by the generating unit 32a. Further, the filter processing unit 32b allocates the current instantaneous data 100 to the prediction data 60 of the current target data 50 based on the previous target data 50, thereby ensuring continuity with the previous target data 50. and an allocation unit 322b that takes The allocation unit 322b can allocate a plurality of prediction data 60 to one instantaneous data 100 of this time. As a result, target detection accuracy can be improved.

In the above-described embodiment, the target object detection device 1 is provided in a vehicle, but it may of course be provided in a moving object other than a vehicle, such as a ship or an aircraft.

Further, in the above-described embodiment, ESPRIT is used as an example of the direction-of-arrival estimation method used by the target detection device 1, but the method is not limited to this. For example, DBF (Digital Beam Forming), PRISM (Propagator method based on an Improved Spatial-smoothing Matrix), MUSIC (Multiple Signal Classification), etc. may be used.

Further, in the above-described embodiment, the filter processing unit 32b uses a particle filter, but the time-series filtering used is not limited to the particle filter. It may be time series filtering.

Further effects and modifications can be easily derived by those skilled in the art. Therefore, the broader aspects of the invention are not limited to the specific details and representative embodiments so shown and described. Accordingly, various changes may be made without departing from the spirit or scope of the general inventive concept defined by the appended claims and equivalents thereof.

1 radar device 32 signal processing unit 32a generation unit 32b filter processing unit 33 storage unit 33a history data 50, 50a, 50b target data 60, 60a, 60b prediction data 100 instantaneous data 321b prediction unit 322b allocation unit 323b weighting unit 324b resampling Part 325b Target object data generation part LC Other vehicle MC Own vehicle

Claims (5)

Receiving the reflected wave of the transmitted radio wave, and repeatedly generating instantaneous data in each processing cycle , including the distance and relative velocity from the reflection point of the target that reflected the reflected wave, and an estimated angle indicating the direction of arrival of the reflected wave. a generator that
An estimated angle indicating a distance and a relative speed to the target and an azimuth at which the target is positioned, which are associated with the instantaneous data, by subjecting the instantaneous data generated by the generating unit to time-series filtering. a filter processing unit that generates target data including
The filter processing unit is
Prediction for generating, from the target data generated in the previous processing cycle, prediction data that is a prediction value of the target data that the target represented by the target data is predicted to have in the current processing cycle a data generator;
an index calculation unit that calculates an evaluation index that evaluates the relationship between each of the instantaneous data within a predetermined allocation range from the position indicated by the prediction data and the prediction data;
Allocating the instantaneous data to any of the predictive data according to a first allocation condition using the evaluation index, and assigning the instantaneous data allocated to any of the predictive data according to the first allocation condition to the an allocation unit that redundantly allocates one of the prediction data to which the instantaneous data is not allocated according to a second allocation condition using an evaluation index ;
a target data generation unit that generates the target data for each of the plurality of prediction data to which the instantaneous data is allocated by the allocation unit;
A target detection device characterized by: The time series filtering is
a particle filter that allocates a predetermined number of particle data to the instantaneous data;
The filter processing unit is
a weighting unit that gives a predetermined weight to the predetermined number of particle data based on the instantaneous data;
a resampling unit that rearranges the particle data for which the weight given by the weighting unit is less than a predetermined threshold among the predetermined number of particle data,
The resampling unit
Determining the predetermined threshold based on a relative relationship between the instantaneous data and each of the plurality of prediction data when the plurality of prediction data are assigned to the one current instantaneous data. The target detection device according to claim 1 , characterized by: The filter processing unit is
When the one current instantaneous data has continuity with the same plurality of previous target data over a predetermined past period or more, the current target data is obtained from the plurality of previous target data. The target detection device according to any one of claims 1 and 2 , wherein a representative value of data is generated. The evaluation index uses a cost value that becomes a lower value for a combination of the instantaneous data and the predicted data that have closer values for the distance, the relative speed, and the estimated angle.
The target detection device according to any one of claims 1 to 3, characterized by: Receiving the reflected wave of the transmitted radio wave, and repeatedly generating instantaneous data in each processing cycle , including the distance and relative velocity to the reflection point of the target that reflected the reflected wave, and an estimated angle indicating the direction of arrival of the reflected wave. a generation step to
Time-series filtering is applied to the instantaneous data generated by the generating step, thereby obtaining a distance and relative velocity to the target, which are associated with the instantaneous data, and an estimated angle indicating the azimuth of the target. a filtering step of generating target data comprising
The filtering step includes
Prediction for generating prediction data, which is a predicted value of the target data that the target represented by the target data is predicted to have in the current processing cycle, from the target data generated in the previous processing cycle. a data generation step;
an index calculation step of calculating an evaluation index that evaluates the relationship with the predicted data for each of the instantaneous data within a predetermined allocation range from the position indicated by the predicted data;
Allocating the instantaneous data to any of the prediction data according to a first allocation condition using the evaluation index, and assigning the instantaneous data allocated to any of the prediction data according to the first allocation condition to the an assigning step of redundantly assigning the instantaneous data to any of the predictive data to which the instantaneous data is not assigned according to a second assignment condition using an evaluation index ;
and a target data generation step of generating the target data for each of the plurality of prediction data to which the instantaneous data is assigned due to the affirmation of assignment.
A target detection method characterized by:

Images