JP2023156543A - Detection processor and method for detection - Google Patents

Abstract

To detect a still object by a radio sensor.SOLUTION: The detection processor performs still object detection processing including: acquiring detection data showing the result of detecting a moving object in a detection area by a radio sensor; and performing analysis based on the detection data for each one of a plurality of division areas in a detection area; and estimating one of the division areas which the moving object is avoiding, as the location of a still object existence.SELECTED DRAWING: Figure 2

Description

The present disclosure relates to a detection processing device and a detection method.

Patent Document 1 discloses a radio wave sensor installed on a road. A radio wave sensor emits radio waves to a preset detection area to detect moving objects such as pedestrians or vehicles.

JP 2017-90138 Publication

It is also desirable to use radio wave sensors to detect not only moving objects such as pedestrians or vehicles, but also stationary objects such as fallen objects on the road. However, although radio wave sensors are suitable for detecting moving objects, they are not suitable for detecting stationary objects. In other words, radio wave sensors are suitable for detecting moving objects because they can observe the moving speed and displacement of a detection target using the Doppler effect. On the other hand, since the reflected waves of the radio waves emitted from the radio wave sensor return from all kinds of stationary objects, the radio wave sensor is not suitable for detecting specific stationary objects such as fallen objects on the road.

Therefore, it is desirable to be able to detect stationary objects while using a radio wave sensor suitable for detecting moving objects.

One aspect of the present disclosure is a sensing processing device. The disclosed detection processing device acquires detection data indicating a detection result of a moving object within a detection area by a radio wave sensor, and performs an analysis based on the detection data for each of a plurality of divided areas within the detection area. , and a process of estimating a divided area that the moving object avoids among the plurality of divided areas as a stationary object existing position.

Another aspect of the disclosure is a sensing method. The disclosed detection method acquires detection data indicating the detection results of a moving object within a detection area by a radio wave sensor, and performs an analysis based on the detection data for each of a plurality of divided areas within the detection area. The method includes estimating a divided area that the moving object avoids among the divided areas as a stationary object existing position.

According to the present disclosure, stationary objects can be detected.

FIG. 1 is a perspective view of a road on which radio wave sensors are installed. FIG. 2 is a configuration diagram of the radio wave sensor. FIG. 3 is a flowchart of moving object detection processing. FIG. 4 is an explanatory diagram of generation of detection data by a radio wave sensor. FIG. 5 is a flowchart of stationary object detection processing according to the first embodiment. FIG. 6 is a data structure diagram of histogram data. FIG. 7 is an explanatory diagram of histogram data generation. FIG. 8 is a diagram showing an example of detected position histogram data. FIG. 9 is a flowchart of stationary object detection processing according to the second embodiment. FIG. 10 is an explanatory diagram of histogram data generation. FIG. 11 is a diagram showing an example of lane change position histogram data.

[Description of embodiments of the present disclosure]

(1) The detection processing device according to the embodiment includes a process of acquiring detection data indicating a detection result of a moving object within a detection area by a radio wave sensor, and a process based on the detection data for each of a plurality of divided areas within the detection area. Through analysis, the apparatus is configured to perform a stationary object detection process including a process of estimating a divided area that the moving object avoids among the plurality of divided areas as a stationary object existing position. According to the detection processing device according to the embodiment, a divided area that a moving object avoids can be estimated as a stationary object existing position.

(2) Preferably, the detection data includes a position at which the moving object is detected within the detection area. In this case, it is possible to analyze the position that the moving object is avoiding from the position where the moving object was detected.

(3) Performing the analysis based on the detection data for each of the plurality of divided areas may include determining the detection frequency of the moving object in each of the plurality of divided areas during the target period. In this case, the position avoided by the moving object can be analyzed from the detection frequency of the moving object in each of the plurality of divided areas.

(4) Performing the analysis based on the detection data for each of the plurality of divided areas means that the detection frequency in the first divided area included in the plurality of divided areas and the first divided area among the plurality of divided areas are The method may include analyzing the detection frequency in one or more second divided areas other than the area. In this case, the position that the moving object is avoiding can be analyzed based on the detection frequency in the first divided area and the detection frequency in the second divided area.

(5) Performing the analysis based on the detection data for each of the plurality of divided areas means comparing the detection frequency in the first divided area with a representative value of the detection frequency in each of the plurality of second divided areas. This may include: In this case, the position that the moving object is avoiding can be analyzed by comparing the detection frequency in the first divided area and the representative detection frequency in the second divided area.

(6) Performing the analysis based on the detection data for each of the plurality of divided areas may include determining the lane change frequency of the moving object in each of the plurality of divided areas during the target period. Preferably, the lane change frequency indicates the frequency with which the moving object moving along the movement lane changes the movement lane. In this case, the position avoided by the moving object can be analyzed from the lane change frequency of the moving object in each of the plurality of divided areas.

(7) Performing an analysis based on the detection data for each of the plurality of divided areas includes determining the lane change frequency in the first divided area included in the plurality of divided areas and the lane change frequency in the first divided area included in the plurality of divided areas. The method may include analyzing the lane change frequency in one or more second divided areas other than the divided area. In this case, the position avoided by the moving object can be analyzed based on the lane change frequency in the first divided area and the lane change frequency in the second divided area.

(8) Performing the analysis based on the detection data for each of the plurality of divided areas means that the lane change frequency in the first divided area is determined as a representative value of the lane change frequency in each of the plurality of second divided areas. It can include contrasting. In this case, the position that the moving object is avoiding can be analyzed by comparing the lane change frequency in the first divided area and the representative value of the lane change frequency in the second divided area.

(9) The detection method according to the embodiment acquires detection data indicating a detection result of a moving object within a detection area by a radio wave sensor, and performs an analysis based on the detection data for each of a plurality of divided areas within the detection area. In this way, a divided area that the moving object avoids among the plurality of divided areas is estimated as a stationary object existing position. According to the detection method according to the embodiment, a divided area avoided by a moving object can be estimated as a stationary object location.

[Details of embodiments of the present disclosure]

FIG. 1 shows a road 70 on which a radio wave sensor 10 is installed. The radio wave sensor 10 detects objects within the detection area 201 by reflecting the transmitted radio waves. When the detection area 201 is set on the road 70 on which the vehicle 50 travels, the moving object to be detected is, for example, the vehicle 50, and the detection area 201 is set on a crosswalk or other road where pedestrians pass. If so, the detected moving object is, for example, a pedestrian. Although the radio wave sensor 10 is installed so as to radiate radio waves to the front of the vehicle 50 as shown in FIG. 1, it may be installed so as to radiate the radio waves to the back of the vehicle 50.

The radio wave sensor 10 of the embodiment is configured as a millimeter wave radar sensor, and is used for traffic monitoring of the vehicle 50. The radio wave sensor 10 of the embodiment can not only detect a moving object such as a vehicle 50 traveling on a road 70, but also a stationary object 60 such as a fallen object on the road 70. In the embodiment, a moving object such as the vehicle 50 is detected based on received data of a reflected wave from the moving object, and a stationary object is detected by estimation using detection data indicating the detection result of the moving object.

As shown in FIG. 2, the radio wave sensor 10 includes a transmitter 11 that transmits radio waves for object detection. In the embodiment, the signal transmitted as a radio wave from the transmitter 11 is a chirp signal. A chirp signal is a signal whose frequency changes over time.

The radio wave sensor 10 includes a receiver 13 that receives reflected waves of transmitted radio waves. The receiver 13 outputs a received signal of the reflected wave. The radio sensor 10 determines the distance from the radio sensor 10 to the object and the angle of existence (azimuth) of the object. The two-dimensional position coordinates of the object are determined from the distance and angle.

The radio wave sensor 10 includes a signal processing device 15 . The signal processing device 15 performs signal processing on the received signal of the reflected wave obtained from the receiver 13 and outputs received data to the processing device 100 .

The radio wave sensor 10 includes a processing device 100. Processing device 100 controls transmitter 11, receiver 13, and signal processing device 15. The processing device 100 acquires the received data and determines the two-dimensional position coordinates of the object. The processing device 100 is configured by a computer including a processor 110 and a storage device 120, for example. Processor 110 is, for example, a CPU as shown in FIG. The storage device 120 includes, for example, a primary storage device and a secondary storage device. The primary storage device is, for example, a RAM. The secondary storage device is, for example, a hard disk drive (HDD) or a solid state drive (SSD).

A computer program 150 is stored in the storage device 120. The computer program 150 of the embodiment causes the processor 110 of the processing device 100 to execute the moving object detection process 113. The computer program 150 includes a program code for causing the processor 110 to execute the moving object detection process 113. The computer program 150 also causes the processor 110 of the processing device 100 to execute the stationary object detection process 114. Computer program 150 includes program code for causing processor 110 to execute stationary object detection processing 114. Processor 110 reads and executes computer program 150 stored in storage device 120.

The storage device 120 stores detection data 121 obtained by the moving object detection process 113. Further, the storage device 120 stores histogram data 122 that is generated in the stationary object detection process 114 and used for stationary object detection.

Note that either one or both of the moving object detection processing 113 and the stationary object detection processing 114 according to the embodiment may be executed by a hardware logic circuit. Further, either one or both of the moving object detection processing 113 and the stationary object detection processing 114 according to the embodiment may be executed by the external processing device 20 of the radio wave sensor 10.

The radio wave sensor 10 includes a communication interface 130 connected to a network 140 in order to communicate with the external processing device 20. The communication interface 130 is for wireless communication or wired communication. The radio wave sensor 10 can send and receive data to and from the external processing device 20 via the communication interface 130. The external processing device 20 may be, for example, a server on a network, another radio sensor or a roadside sensor other than the radio sensor, or a device connected to the radio sensor 10. .

FIG. 3 shows the procedure of the moving object detection process 113 executed by the radio wave sensor 10. In step S21, the transmitter 11 of the radio wave sensor 10 emits a transmission wave for detecting an object. In step S22, the receiver 13 of the radio wave sensor receives a reflected wave generated by reflection from an object from the emitted transmission wave. A received signal of the reflected wave received by the receiver 13 is given to a signal processing device 15 . The signal processing device 15 provides the received data to the processing device 100. The received data is, for example, digital data indicating reflected waves.

In step S23, the processing device 100 executes a distance profile calculation process. The distance to the object is calculated based on the time required from the timing of the transmitted wave to the timing of receiving the reflected wave from the object. This required time is the time required for the radio wave to travel back and forth between the radio wave sensor 10 and the object. Therefore, the processing device 100 can determine the distance from the radio wave sensor 10 to the object from this required time. In distance profile calculation processing, the processing device 100 calculates a distance profile by performing fast Fourier transform (FFT) on received data.

In step S24, the processing device 100 executes a process of estimating the arrival direction of the reflected wave based on the received data of the reflected wave. The direction of arrival (angle of arrival) is calculated based on the phase difference of reflected waves between the plurality of antenna elements included in the receiver 13.

In the arrival direction estimation process of step S24, reflection intensity data of reflected waves is calculated for each distance zone from the radio wave sensor 10. The reflection intensity data shows reflection intensity spectra for each of a plurality of distance zones.

In step S25, the processing device 100 determines, for example, the maximum point in the reflection intensity spectrum to be the peak point corresponding to the detected object. In step S26, the processing device 100 detects the angle at which the peak point exists as the angle at which the object exists. Then, the processing device 100 obtains two-dimensional position coordinates indicating where the object is present in the detection area 201 as the detection data 121. In the embodiment, the two-dimensional position coordinates of the object are the detection results by the radio wave sensor 10.

In step S27, the processing device 100 stores the detection data 121 in the storage device 120. The radio wave sensor 10 obtains detection data 121 at every predetermined time interval, and the detection data 121 at each time is stored in the storage device 120. The stored plural pieces of detection data are used in stationary object detection processing 114.

FIG. 4 shows an example of the state of the road 70 at a certain time t and the detection data 121 obtained at that time t. As shown in FIG. 4, at time t, it is assumed that the first vehicle 50A, the second vehicle 50B, and the third vehicle 50C are traveling within the detection area 201 of the radio wave sensor 10. The road 70 has a first lane 70A that is a first travel lane, a second lane 70B that is a second travel lane, and a third lane 70C that is a third travel lane. The number of lanes that the road 70 has is not particularly limited. Here, the first vehicle 50A is traveling on the first lane 70A, the second vehicle 50B is traveling on the second lane 70B, and the third vehicle 50C is traveling on the third lane 70C. A stationary object 60 that is a fallen object from a vehicle is present on the second lane 70B.

As shown in FIG. 4, the detection data 121 obtained at time t includes a first detection position P _A (x, y) which is the detection position (two-dimensional position coordinates) of the first vehicle 50A, and a detection position of the second vehicle 50B. It has a second detection position P _B (x, y), which is the position of the third vehicle 50C, and a third detection position P _C (x, y), which is the detection position of the third vehicle 50C.

Since it is difficult for the radio wave sensor 10 to detect the stationary object 60, the detection position of the stationary object 60 is not included in the detection data 121 obtained at time t. In the stationary object detection process 114 according to the embodiment, the presence and position of the stationary object 60 is estimated from the detection data 121 that does not include the detected position of the stationary object 60. The stationary object detection process 114 performs analysis based on the detection data 121 for each of the plurality of divided areas in the detection area 201, and estimates the divided area that the vehicle avoids from among the plurality of divided areas as the stationary object location. . In this manner, in the stationary object detection process 114, a stationary object is indirectly detected by estimation using detection data indicating the detection result of a moving object.

FIG. 5 shows a stationary object detection process 114A according to the first embodiment, which is an example of the stationary object detection process 114. The stationary object detection process 114A includes a process of acquiring detection data 121 (step S31) and an analysis based on the detection data 121 for each of the plurality of divided areas, thereby detecting the divided area that the vehicle is avoiding among the plurality of divided areas. and a process of estimating the position of the stationary object (from step S32 to step S26).

In step S31, the processor 110 of the processing device 100 acquires the detection data 121 for the past T hours, which is the analysis target period, from the storage device 120. The detection data for the past T hours includes detection data 121 for each of a plurality of times included in the past T hours. In the stationary object detection process 114A, a stationary object is detected from a plurality of detection data 121 in the past T time.

In step S32, the processing device 100 generates histogram data 122 from the detection data 121 for the past T hours. In the stationary object detection process 114A according to the first embodiment, detection position histogram data 122A is generated as an example of the histogram data 122.

FIG. 6 shows the structure of the histogram data 122 generated in step S32. The histogram data 122 is composed of a set of frequencies D _ij for each of a plurality of divided areas (i, j) obtained by dividing the detection area 201 into a grid shape. Here, the detection area 201 is divided into three areas in the road width direction and eight areas in the road traveling direction, corresponding to the fact that the road 70 has three lanes 70A, 70B, and 70C. In the histogram data 122, the road width direction corresponds to the j direction (abscissa) in FIG. 6, and the road traveling direction corresponds to the i direction (ordinate) in FIG. Note that the method of setting the divided areas is not particularly limited.

In the histogram data 122 shown in FIG. 6, each divided area is indicated by a pair (i, j) of an ordinate i and an abscissa j. The frequency in the divided area (i, j) is indicated by D _ij . The frequency D _ij in the first embodiment indicates the detection frequency of a moving object in the divided area (i, j).

In step S32, it is identified to which divided area (i, j) the detected position of each object in the detection data 121 at each time t corresponds. For example, in the case of the detection data 121 at time t shown in FIG. 4, as shown in FIG. 7, the first detection position P _A (x, y) of the first vehicle 50A corresponds to the divided area (6, 3), The second detection position P _B (x, y) of the second vehicle 50B corresponds to the divided area (8, 2), and the third detection position P _C (x, y) of the third vehicle 50C corresponds to the divided area (2, It is identified that it corresponds to 1).

In this case, in the histogram data 122, the frequency D ₆₃ in the divided area (6, 3), the frequency D ₈₂ in the divided area (8, 2), and the frequency D ₂₁ in the divided area (2, 1) are each incremented by (+1 ) to be done. The above processing is performed for each of the plurality of detection data 121 for the past T hours, thereby generating the detected position histogram data 122A for the past T hours. FIG. 8 shows an example of the generated detection position histogram data 122A. Note that the number of pieces of detection data 121 for the past T hours used to generate the histogram data 122 is not particularly limited. When the number of pieces of detection data 121 for the past T hours is increased, the time interval between each detection data 121 becomes smaller. Conversely, when the number of pieces of detection data 121 for the past T hours is reduced, the time interval between each piece of detection data 121 increases. When the time interval between each detection data 121 becomes large and the vehicle 50 passes through a certain divided area at high speed, the vehicle 50 may not be detected in that divided area. Therefore, in the detection position histogram data 122A of FIG. 8, the total detection frequency in the three lanes 70A, 70B, and 70C differs depending on the position of i.

Since the vehicle 50 travels while avoiding stationary objects 60 on the road 70, analysis is performed based on the detection data 121 for each of the plurality of divided areas within the detection area 201, so that the vehicle 50 avoids the stationary objects 60 among the plurality of divided areas. It is possible to detect the divided area where a stationary object exists. For example, in FIG. 8, the frequency D ₅₂ of the divided area (5, 2) is 2, which is smaller than the frequencies D _ij of the other divided areas. Therefore, it can be seen that the vehicle 50 is traveling avoiding the divided area (5, 2).

Returning to FIG. 5, in step S33, the processor 110 of the processing device 100 stores the generated detected position histogram data 122A in the storage device 120.

In step S34, the processing device 100 detects all the divided areas (i, j) in the detected position histogram data 122A in order to detect the divided area that the vehicle 50 is avoiding, that is, the divided area where a stationary object exists. The determination shown in step S35 is made regarding the frequency of .

The determination shown in step S35 is made by comparing the frequency D _ij of the divided area (i, j) to be determined with the representative value of the frequency of one or more divided areas around the divided area to be determined. It will be done. Note that the divided area (i, j) to be determined is referred to as a first divided area D1, and one or more divided areas around the divided area to be determined is referred to as a second divided area D2.

The position of the second divided area D2 is determined according to the position of the first divided area D1, which is the determination target. The one or more second divided areas D2 can include, for example, one or more divided areas behind the first divided area D1. Here, the term "rear" refers to the direction opposite to the front when the traveling direction of the vehicle 50 or the lanes 70A, 70B, and 70C included in the detection area 201 is defined as the front. This is the direction in which the value decreases. For example, if the first divided area D1 is located at i=5, the rear area will be in the range of i=4 or less. It is preferable that the one or more second divided areas D2 include a divided area one area behind the first divided area D1. For example, if the first divided area D1 is located at a position where i=5, it is preferable that one or more second divided areas D2 include the divided area D2 where i=4. Furthermore, it is preferable that the one or more second divided areas D2 include a divided area directly behind the first divided area D1. For example, if the first divided area D1 is located at (i, j)=(5,2), it is preferable that one or more second divided areas D2 include the divided area (4,2). The one or more second divided areas D2 may include divided areas on the sides of the first divided area D1. Here, the side refers to a range in which divided areas having the same value of i exist.

In FIG. 8, the second divided area D2 includes three divided areas located one place behind the first divided area D1 and two divided areas located to the sides.

The representative value of the frequency of the plurality of second divided areas D2 is here an average value, but may be another representative value such as a median value. In the determination in step S35, the frequency of the first divided area D1 and the average frequency of the second divided area D2 are compared. If the frequency of the first divided area D1 is smaller than the average frequency of the second divided area D2, it is estimated that the vehicle 50 is moving avoiding the first divided area D1. In addition, in the determination formula shown in step S35, the average value of the frequency of the second divided area D2 is established so that the frequency of the first divided area D1 is sufficiently smaller than the average value of the frequency of the second divided area D2. The result obtained by subtracting a predetermined threshold value from the frequency is compared with the frequency of the first divided area. By subtracting the threshold value, the vehicle 50 avoids the first divided area D1 when the difference between the frequency of the first divided area D1 and the average frequency of the second divided area D2 is equal to or greater than the threshold value. It is estimated that they are moving.

The determination in step S35 is repeated with all divided areas as the first divided areas D1. If the determination formula in step S35 holds true for a certain first divided area D1, the processing device 100 detects that a stationary object exists in that first divided area D1 (step S36).

For example, in the case of FIG. 8, the frequency _D52 of the divided area D1 (first divided area (5, 2)) is sufficiently smaller than the average value of the five divided areas D2 around it, so The presence of a stationary object in the divided area (5, 2) is detected. In this way, according to the present embodiment, even if the detection data 121 in which no stationary object is detected is used, the presence of a stationary object can be estimated by the stationary object detection process 114A.

FIG. 9 shows a stationary object detection process 114B according to the second embodiment, which is another example of the stationary object detection process 114. The stationary object detection process 114B performs a process of acquiring detection data 121 (step S50) and an analysis based on the detection data 121 for each of the plurality of divided areas, thereby detecting the divided area that the vehicle is avoiding among the plurality of divided areas. and a process of estimating the position of the stationary object (steps S51 to S56).

In step S50, the processor 110 of the processing device 100 acquires the detection data 121 for the past T hours, which is the analysis target period, from the storage device 120. The detection data for the past T hours includes detection data 121 for each of a plurality of times included in the past T hours. In the stationary object detection process 114B according to the second embodiment, a lane change position, which is a lane change position, is determined from a plurality of detection data 121 in the past T time, and a stationary object is detected.

In step S51, the processing device 100 calculates the lane change position of the vehicle from the detection data 121 for the past T hours. Then, in step S52, the processing device 100 generates histogram data 122 from the lane change positions for the past T hours. In the stationary object detection process 114B according to the second embodiment, lane change position histogram data 122B is generated as an example of the histogram data 122.

The structure of the lane change position histogram data 122B is similar to the structure shown in FIG. 6. That is, the histogram data 122 is composed of a set of frequencies D _ij for each of a plurality of divided areas (i, j) obtained by dividing the detection area 201 into a grid shape. The frequency D _ij in the second embodiment indicates the lane change frequency of a moving object in the divided area (i, j).

As shown in FIG. 10, in step S51, vehicle trajectories T _A and T _B of each vehicle 50A and 50B are calculated from the detection data 121. In step S51, positions where vehicle trajectories T _A and T _B cross lanes are identified as lane change positions C _A , C _B , and C _C .

In step S52, it is identified which divided area (i, j) each lane change position C _A , C _B , C _C corresponds to. For example, as shown in FIG. 10, lane change position C _A corresponds to divided area (4, 2), lane change position C _B corresponds to divided area (3, 2), and lane change position C _C is identified as corresponding to the divided area (6, 1).

In this case, in the histogram data 122B, the frequency D ₄₂ in the divided area (4, 2), the frequency D ₃₂ in the divided area (3, 2), and the frequency D ₆₁ in the divided area (6, 1) are each incremented by (+1 ) to be done. By performing the above processing for each lane change position that occurred in the past T hours, lane change position histogram data 122B for the past T hours is generated. FIG. 11 shows an example of the generated lane change position histogram data 122B.

Since the vehicle 50 travels while avoiding stationary objects 60 on the road 70, by performing an analysis based on the detection data 121 for each of the plurality of divided areas within the detection area 201, the vehicle 50 avoids the stationary objects 60 among the plurality of divided areas. The divided area where the object is located can be detected as the location of a stationary object. For example, in FIG. 11, the frequency D ₄₂ of the divided area (4, 2) is 48, and the frequency D ₃₂ of the divided area (3, 2) located further behind it is 32. 2) (3,2) has a higher lane change frequency than the surrounding divided areas. Therefore, it can be seen that the vehicle is running avoiding the divided area (5,2) which is downstream of these divided areas (4,2) (3,2).

Returning to FIG. 9, in step S53, the processor 110 of the processing device 100 stores the generated lane change position histogram data 122B in the storage device 120.

In step S54, the processing device 100 detects all the divided areas (i, j ) is determined in step S55.

The determination shown in step S55 is made by comparing the frequency D _ij of the divided area (i, j) to be determined with the representative value of the frequency of one or more divided areas around the divided area to be determined. It will be done. In the second embodiment as well, the divided area (i, j) to be determined is referred to as a first divided area D1, and one or more divided areas around the divided area to be determined is referred to as a second divided area D2. .

The position of the second divided area D2 is determined according to the position of the first divided area D1, which is the determination target. The one or more second divided areas D2 are, for example, all the divided areas around the first divided area D1. For example, if the first divided area D1 is located at (i, j) = (4, 2), one or more second divided areas D2 are located at the divided area (3, 1) (3, 2) (3 , 3) (4, 1), (4, 3) (5, 1) (5, 2) (5, 3).

The representative value of the frequency of the plurality of second divided areas D2 is here an average value, but may be another representative value such as a median value. In the determination in step S55, the frequency of the first divided area D1 and the average value of the frequencies of the second divided area D2 are compared. If the frequency of the first divided area D1 is greater than the average frequency of the second divided area D2, it is estimated that the vehicle 50 is moving avoiding the downstream of the first divided area D1. In addition, in the determination formula shown in step S55, the average value of the frequency of the second divided area D2 is established so that the frequency of the first divided area D1 is sufficiently larger than the average value of the frequency of the second divided area D2. The result obtained by adding a predetermined threshold value to the frequency is compared with the frequency of the first divided area. By adding the threshold values, if the difference between the frequency of the first divided area D1 and the average frequency of the second divided area D2 is greater than or equal to the threshold value, the vehicle 50 can move downstream of the first divided area D1. It is presumed that they are avoiding it.

The determination in step S55 is repeated with all divided areas as the first divided areas D1. If the determination formula in step S35 holds true for a certain first divided area D1, the processing device 100 detects that a stationary object exists downstream of the first divided area D1 (step S56).

For example, in the case of FIG. 1, the frequency _D52 of the divided area D1 (first divided area (4, 2)) is sufficiently larger than the average value of the eight divided areas D2 around it, so It is detected that a stationary object exists in the divided area (5, 2) that is downstream of the divided area (4, 2). In this way, according to the present embodiment, even if the detection data 121 in which no stationary object is detected is used, the presence of a stationary object can be estimated by the stationary object detection process 114.

Note that the processing device 100 executes both the stationary object detection process 114A according to the first embodiment and the stationary object detection process 114B according to the second embodiment, and combines the processing results of both processes 114A and 114B to detect the stationary object. may be detected. By detecting a stationary object using the processing results of both processes 114 and 114B, detection accuracy is increased.

It should be noted that the embodiments disclosed herein are illustrative in all respects and should not be considered restrictive. The scope of the present invention is indicated by the claims rather than the above-mentioned meaning, and is intended to include meanings equivalent to the claims and all changes within the scope.

10: Radio wave sensor 11: Transmitter 13: Receiver 15: Signal processing device 20: External processing device 50: Vehicle 50A: First vehicle 50B: Second vehicle 50C: Third vehicle 60: Stationary object 70: Road 70A: No. 1st lane 70B: 2nd lane 70C: 3rd lane 100: Processing device 110: CPU
113: Moving object detection processing 114: Stationary object detection processing 114A: Stationary object detection processing 114B: Stationary object detection processing 120: Storage device 121: Detection data 122: Histogram data 122A: Detection position histogram data 122B: Lane change position histogram data 130 : Communication interface 140 : Network 150 : Computer program 201 : Detection area C _A : Lane change position C _B : Lane change position D1 : First divided area D2 : Second divided area D _ij : Frequency P _A : First detection position P _B : Second detection position _PC : Third detection position T _A : Vehicle trajectory T _B : Vehicle trajectory

Claims (9)

A process of acquiring detection data indicating a detection result of a moving object within a detection area by a radio wave sensor;
A process of estimating a divided area that the moving object avoids among the plurality of divided areas as a stationary object existing position by performing an analysis based on the detection data for each of the plurality of divided areas in the detection area;
A detection processing device configured to perform stationary object detection processing. The detection processing device according to claim 1, wherein the detection data includes a position where the moving object is detected within the detection area. According to claim 1 or claim 2, performing the analysis based on the detection data for each of the plurality of divided areas includes determining the detection frequency of the moving object in each of the plurality of divided areas during the target period. The detection processing device described. Performing the analysis based on the detection data for each of the plurality of divided areas means that the detection frequency in the first divided area included in the plurality of divided areas and the detection frequency in the first divided area among the plurality of divided areas other than the first divided area are analyzed for each of the plurality of divided areas. The detection processing device according to claim 3, further comprising analyzing the detection frequency using the detection frequency in one or more second divided areas. Performing the analysis based on the detection data for each of the plurality of divided areas includes comparing the detection frequency in the first divided area with a representative value of the detection frequency of each of the plurality of second divided areas. The detection processing device according to claim 4. Performing the analysis based on the detection data for each of the plurality of divided areas includes determining the lane change frequency of the moving object in each of the plurality of divided areas during the target period,
The detection processing device according to any one of claims 1 to 5, wherein the lane change frequency indicates how often the moving object moving along the movement lane changes the movement lane. The analysis based on the detection data for each of the plurality of divided areas includes the lane change frequency in the first divided area included in the plurality of divided areas and the frequency of lane changes in the first divided area included in the plurality of divided areas, and the analysis based on the detected data for each of the plurality of divided areas. The detection processing device according to claim 6, further comprising analyzing the lane change frequency in one or more of the second divided areas. Performing the analysis based on the detection data for each of the plurality of divided areas means comparing the lane change frequency in the first divided area with a representative value of the lane change frequency in each of the plurality of second divided areas. The detection processing device according to claim 7. Acquire detection data indicating the detection results of moving objects within the detection area by the radio wave sensor,
Detection comprising: estimating a divided area that the moving object avoids among the plurality of divided areas as a stationary object existing position by performing an analysis based on the detection data for each of the plurality of divided areas in the detection area. Method.

Images