Radio wave sensor and analysis method

Abstract

This radio wave sensor is provided with: a detection unit for detecting an object present on a road; and an analysis unit for analyzing, by first analysis processing, first detection data obtained by the detection unit detecting a first object present in a first region of the road, and analyzing, by second analysis processing, second detection data obtained by the detection unit detecting a second object present in a second region of the road.

Description

Radio wave sensor and analysis method

[0001] This application claims priority to Japanese Patent Application No. 2024-063315, filed April 10, 2024, and incorporates by reference the entire contents of said Japanese application.

[0002] Patent Document 1 discloses a radio wave sensor that detects objects within a target area including a crosswalk.

[0003] International Publication No. 2021 / 181981

[0004] A radio wave sensor according to one aspect of the present disclosure includes a detection unit that detects an object present on a road, and an analysis unit that analyzes first detection data obtained by the detection unit detecting a first object present on a first area of ​​the road using a first analysis process, and analyzes second detection data obtained by the detection unit detecting a second object present on a second area of ​​the road using a second analysis process.

[0005] FIG. 1 is a diagram illustrating an example of use of a radio wave sensor according to an embodiment. FIG. 2 is a diagram illustrating an example of setting a detection area in an inherent coordinate system. FIG. 3 is a diagram illustrating an example of a detection area. FIG. 4 is a perspective view illustrating an example of the external configuration of an infrastructure radio wave sensor according to an embodiment. FIG. 5 is a block diagram illustrating an example of the hardware configuration of an infrastructure radio wave sensor according to an embodiment. FIG. 6 is a functional block diagram illustrating an example of functions of an infrastructure radio wave sensor according to an embodiment. FIG. 7 is a diagram illustrating movement of an object for each area. FIG. 8 is a diagram illustrating an example of a tracking process when a target object is hidden by an obstruction. FIG. 9 is a diagram illustrating an example of tracking an object in a mask area. FIG. 10 is a flowchart illustrating an example of operation of the infrastructure radio wave sensor according to an embodiment.

[0006] <Problem to be Solved by the Present Disclosure> On roads, the behavior (movement direction, movement speed, etc.) of moving objects (vehicles, pedestrians, etc.) differs in each area of ​​the roadway, sidewalk, crosswalk, intersection, etc. Radio wave sensors are installed on roads for the purpose of traffic monitoring, but traffic conditions cannot be monitored accurately unless the detection data of the radio wave sensors is properly analyzed for each monitored area.

[0007] <Effects of the Present Disclosure> According to the present disclosure, accurate monitoring of traffic conditions becomes possible.

[0008] <Outline of Embodiments of the Present Disclosure> Below, an outline of embodiments of the present disclosure will be listed and described.

[0009] (1) The radio wave sensor according to this embodiment includes a detection unit that detects an object present on a road, and an analysis unit that analyzes, through a first analysis process, first detection data obtained by the detection unit detecting a first object present on a first area of ​​the road, and analyzes, through a second analysis process, second detection data obtained by the detection unit detecting a second object present on a second area of ​​the road. This allows the detection data to be appropriately analyzed through an analysis process appropriate for each area, enabling accurate monitoring of traffic conditions.

[0010] (2) In the above (1), the first analysis process may include a first tracking process for tracking the first object, and the second analysis process may include a second tracking process for tracking the second object. This allows objects to be appropriately tracked by a tracking process appropriate for each area, enabling accurate monitoring of traffic conditions.

[0011] (3) In the above (2), the first tracking process may include estimating the current position of the first object based on past positions of the first object detected by the detection unit a first period going back from the present, and the second tracking process may include estimating the current position of the second object based on past positions of the second object detected by the detection unit a second period going back from the present that is shorter than the first period. As a result, for example, in a first region that is less susceptible to noise, the tracking sensitivity can be increased by extending the period going back (first period), thereby enabling high-precision tracking of the object. In a second region, the tracking sensitivity can be reduced by shortening the period going back (second period), thereby enabling less susceptibility to noise.

[0012] (4) In any one of (1) to (3) above, the first analysis process may include an identification process for identifying the type of the first object, and the second analysis process may not include an identification process for identifying the type of the second object. This allows, for example, in a first region where multiple types of objects are mixed, to identify the type of object and perform analysis according to the type of object. For example, in a second region where a single type of object is present, the processing load can be reduced by not identifying the type of object.

[0013] (5) In the above (4), the identification process may be a process of identifying the type of the first object based on a moving direction of the first object, thereby making it possible to accurately identify the type of object in a first region in which the moving direction differs for each type of object.

[0014] (6) In the above (4) or (5), the first analysis process may be a process of estimating the current position of a first type of object identified by the identification process based on past positions of the first object detected by the detection unit a third period prior to the present, and estimating the current position of a second type of object identified by the identification process based on past positions of the first object detected by the detection unit a fourth period prior to the present that is shorter than the third period. As a result, for example, by setting the period prior to the present for the first type of object to a third period suitable for the first type of object, the first type of object can be appropriately tracked. For example, by setting the period prior to the present for the second type of object to a fourth period suitable for the second type of object, the second type of object can be appropriately tracked.

[0015] (7) In any one of (4) to (6) above, the first area may be an area including a pedestrian crossing, and the second area may be an area where pedestrians wait to cross the pedestrian crossing. This makes it possible to accurately track pedestrians crossing the pedestrian crossing and pedestrians waiting to cross the pedestrian crossing.

[0016] (8) In the above (2), the second tracking process may be a process of tracking the first object moving from outside the second region into the second region and discarding detection data of a third object constantly present in the second region. This makes it possible to track the first object moving from outside the second region into the second region in the second region where objects constantly exist, while excluding the third object constantly present in the second region from the tracking target.

[0017] (9) The analysis method according to this embodiment includes the steps of: detecting an object present on a road based on radio waves emitted from a radio wave sensor toward the road; analyzing, by a first analysis process, first detection data obtained by detecting a first object present on a first area of ​​the road; and analyzing, by a second analysis process, second detection data obtained by detecting a second object present on a second area of ​​the road. This allows the detection data to be appropriately analyzed by an analysis process appropriate for each area, thereby enabling accurate monitoring of traffic conditions.

[0018] The present disclosure can be realized not only as a radio wave sensor having the above-described characteristic configuration and an analysis method having characteristic processing steps, but also as a computer program for causing a computer to execute the characteristic processing, as a traffic monitoring system including a radio wave sensor, or as a semiconductor integrated circuit in which part of the radio wave sensor is implemented.

[0019] <Details of Embodiments of the Present Disclosure> Hereinafter, details of embodiments of the present disclosure will be described with reference to the drawings. Note that at least some of the embodiments described below may be combined in any manner.

[0020] 1 is a diagram showing an example of use of a radio wave sensor according to an embodiment. The radio wave sensor 10 according to this embodiment is a radio wave sensor for traffic monitoring (hereinafter also referred to as an "infrastructure radio wave sensor"), and detects an object at a crosswalk 20. The infrastructure radio wave sensor 10 is, for example, a millimeter-wave radar.

[0021] The infrastructure radio wave sensor 10 is attached to a structure 50 provided on a road. The structure 50 is several meters tall, and the infrastructure radio wave sensor 10 is installed several meters above the ground. The structure 50 includes, for example, a pole 51 and an arm 52 provided near the top end of the pole 51, and the infrastructure radio wave sensor 10 is attached to the arm 52.

[0022] The infrastructure radio wave sensor 10 emits radio waves (millimeter waves) onto the crosswalk 20 and receives the reflected waves to detect an object on the crosswalk 20. More specifically, the infrastructure radio wave sensor 10 can detect the distance from the infrastructure radio wave sensor 10 to an object on the crosswalk 20, the speed of the object, and the horizontal angle (azimuth angle) of the object's location relative to the radio wave emission axis.

[0023] The crosswalk 20 is provided on a roadway 60 near an intersection 70. At the intersection 70, the roadway 60 intersects with a roadway 65. Sidewalks 63a and 63b are provided adjacent to the roadway 60. The roadway 60 includes on-coming lanes 61a and 61b through which vehicles enter the intersection 70 and on-coming lanes 62a and 62b through which vehicles exit the intersection 70. Hereinafter, the direction of travel of vehicles in the on-coming lanes 61a and 61b will be referred to as the "x1 direction," and the direction of travel of vehicles in the on-coming lanes 62a and 62b will be referred to as the "x2 direction." The direction of travel of vehicles traveling from the lower left to the upper right in the drawing on the roadway 65 will be referred to as the "y1 direction," and the direction of travel of vehicles traveling from the upper right to the lower left in the drawing on the roadway 65 will be referred to as the "y2 direction."

[0024] The oncoming lane 61a, which is close to the sidewalk 63a, is a lane for going straight and turning left. That is, a vehicle traveling on the oncoming lane 61a in the x1 direction either goes straight through the intersection 70 or turns left (changing its direction of travel in the y1 direction) to enter the roadway 65. The oncoming lane 61b, which is away from the sidewalk, is a lane for going straight and turning right. That is, a vehicle traveling on the oncoming lane 61b in the x1 direction either goes straight through the intersection 70 or turns right (changing its direction of travel in the y2 direction) to enter the roadway 65. The oncoming lane 62a, which is close to the sidewalk, is used by vehicles traveling on the roadway 60 in the x2 direction to pass the intersection 70 and vehicles traveling on the roadway 65 in the y1 direction to turn left at the intersection 70. The outgoing lane 62b, which is away from the sidewalk, is used by vehicles traveling straight along the roadway 60 in the x2 direction and passing through the intersection 70, and by vehicles traveling along the roadway 65 in the y2 direction and turning right at the intersection 70.

[0025] [2. Detection Area] The infrastructure radio wave sensor 10 sets a detection area 30, which is a range on the road for detecting objects. The detection area 30 is set as part of the radio wave irradiation area 40 of the infrastructure radio wave sensor 10. In other words, the radio wave irradiation area 40 covers the detection area 30. In order for the infrastructure radio wave sensor 10 to monitor the traffic conditions across the entire crosswalk 20, it is preferable to set a detection area 30 that includes the entire crosswalk 20. Note that the radio wave irradiation area 40 is a range in which an object reflects the radio waves irradiated by the infrastructure radio wave sensor 10 and the infrastructure radio wave sensor 10 can detect the object based on the reflected waves from the object, and does not include a range in which the infrastructure radio wave sensor 10 cannot detect an object even if it can irradiate radio waves. However, the radio wave irradiation area 40 is not limited to this and may be the entire range in which the infrastructure radio wave sensor 10 can irradiate radio waves.

[0026] For example, the infrastructure radio wave sensor 10 used to measure the number of pedestrians and bicycles (including riders; hereinafter, pedestrians and bicycles will be simply referred to as "pedestrians") crossing the crosswalk 20 or to control traffic signals installed at the crosswalk 20 is required to detect not only pedestrians on the crosswalk 20 but also pedestrians waiting to cross on the sidewalks 63a and 63b adjacent to the crosswalk 20. For this reason, for example, the detection area 30 includes not only the area of ​​the crosswalk 20 but also areas on the sidewalks 63a and 63b where pedestrians wait to cross. In other words, the detection area 30 is an area extending from the crosswalk 20 on both sides in the longitudinal direction of the crosswalk 20 (the direction in which pedestrians move on the crosswalk 20).

[0027] A coordinate space for detecting an object is set in the infrastructure radio wave sensor 10. Hereinafter, the unique coordinate system set in the infrastructure radio wave sensor 10 will also be referred to as the "unique coordinate system."

[0028] In the infrastructure radio wave sensor 10, the position of an object is expressed as coordinate values ​​in a unique coordinate system. A detection area 30 is set in the unique coordinate system of the infrastructure radio wave sensor 10. Fig. 2 is a diagram for explaining an example of setting a detection area in the unique coordinate system.

[0029] 2, the point indicated by the reference symbol 31O is a point on the ground surface that is a vertically downward projection of the installation position of the infrastructure radio wave sensor 10. The line indicated by the reference symbol 31Y is an intersection line (hereinafter also referred to as the "projection center axis") between the ground and a vertical plane that includes the radio wave emission axis of the infrastructure radio wave sensor 10 (the normal direction of the radio wave transmission / reception surface of the infrastructure radio wave sensor 10), and the line indicated by the reference symbol 31X is a line on the ground surface that intersects with the projection center axis 31Y at point 31O.

[0030] For example, the inherent coordinate system is a virtual coordinate system set in the infrastructure radio wave sensor 10, and is a two-dimensional coordinate system corresponding to the earth's surface. The inherent coordinate system is defined by an X axis and a Y axis. The origin O of the inherent coordinate system corresponds to the real point 31O. The Y axis of the inherent coordinate system corresponds to the real projection center axis 31Y. The X axis of the inherent coordinate system corresponds to the real line 31X.

[0031] In the infrastructure radio wave sensor 10, a virtual detection area 300 corresponding to the detection area 30 in real space is set in a specific coordinate system.

[0032] 3 is a diagram showing an example of a detection area. The detection area 30 includes zebra areas 31 a and 31 b that are the area of ​​the crosswalk 20, waiting areas 32 a and 32 b where pedestrians (including people and bicycles driven by people) waiting to cross the crosswalk 20, and a center reservation area 33.

[0033] The waiting areas 32a and 32b are provided on both sides of the longitudinal direction (width direction of the roadway 60) of the detection area 30. In the example of Fig. 3, the waiting area 32a is set on the sidewalk 63a adjacent to the incoming lane 61a, and the waiting area 32b is set on the sidewalk 63b adjacent to the outgoing lane 62a.

[0034] The median strip area 33 is an area of ​​the crosswalk 20 that overlaps with the median strip 64 .

[0035] Zebra areas 31a and 31b are areas of the crosswalk 20 excluding the center divider area 33. Zebra area 31a is an area of ​​the crosswalk 20 that overlaps with the oncoming lanes 61a and 61b and is adjacent to the waiting area 32a. Zebra area 31b is an area of ​​the crosswalk 20 that overlaps with the offcoming lanes 62a and 62b and is adjacent to the waiting area 32b. The center divider area 33 is located between zebra areas 31a and 31b.

[0036] A detection area 300 in the specific coordinate system is divided into zebra areas 301a and 301b, waiting areas 302a and 302b, and a median area 303. The zebra area 301a corresponds to the zebra area 31a in real space, and the zebra area 301b corresponds to the zebra area 31b in real space. The waiting area 302a corresponds to the waiting area 32a in real space, and the waiting area 302b corresponds to the waiting area 32b in real space. The median area 303 corresponds to the median area 33 in real space.

[0037] Within the radio wave irradiation area 40, there may be stationary objects or objects that remain in the same place for a long period of time (hereinafter also referred to as "non-moving objects"). Examples of non-moving objects include trees, shrubs, utility poles, traffic lights, guardrails, traffic signs, buildings, and stopped vehicles. Such non-moving objects are detected by the infrastructure radio wave sensor 10 and appear as noise in the detection results. For this reason, the infrastructure radio wave sensor 10 sets a mask area that excludes non-moving objects from detection targets.

[0038] 3, a mask area 34a is set for the pole of a traffic signal 65a on the sidewalk 63a, and a mask area 34b is set for the pole of a traffic signal 65b on the sidewalk 63b. Furthermore, a mask area 34c is set for a plant 65c in the median strip 64.

[0039] The mask areas are set in the intrinsic coordinate system. Mask area 304a in the intrinsic coordinate system corresponds to mask area 34a in real space, and mask area 304b in the intrinsic coordinate system corresponds to mask area 34b in real space. Mask area 304c in the intrinsic coordinate system corresponds to mask area 34c in real space. Mask areas 304a, 304b, and 304c exclude traffic lights 65a and 65b and a shrub 65c, which are non-moving objects, from detection targets.

[0040] 3. Hardware Configuration of Infrastructure Radio Wave Sensor FIG. 4 is a perspective view showing an example of the external configuration of the infrastructure radio wave sensor 10 according to the embodiment. As shown in FIG. 4, the infrastructure radio wave sensor 10 includes a housing 140 having a transmitting / receiving surface 140a on one side for transmitting and receiving radio waves. The housing 140 houses a transmitting / receiving unit 104 and a detection processing unit 120. The transmitting / receiving unit 104 includes a transmitting antenna 105a and multiple (e.g., four) receiving antennas 106a. The infrastructure radio wave sensor 10 transmits modulated radio waves from the transmitting antenna 105a through the transmitting / receiving surface 140a. The modulated waves hit an object and are reflected, and the receiving antenna 106a receives the reflected waves. The transmitting / receiving unit 104 and the detection processing unit 120 perform signal processing on the transmitted wave signal and the received wave signal to detect the distance to the object, the line-of-sight velocity of the object, and the azimuth angle at which the object is located. Here, the azimuth angle is the angle of the position of the object relative to the projection center axis.

[0041] 5 is a block diagram showing an example of the hardware configuration of the infrastructure radio wave sensor 10 according to the embodiment. The infrastructure radio wave sensor 10 includes a transmitting / receiving unit 104 and a detection processing unit 120.

[0042] The detection processing unit 120 includes a processor 101 , a non-volatile memory 102 , and a volatile memory 103 .

[0043] The transmitting / receiving unit 104 includes a transmitting circuit 105 and a receiving circuit 106 .

[0044] The transmission circuit 105 includes a transmission antenna 105a. The number of transmission antennas 105a is not limited to one, and may be multiple. The transmission circuit 105 generates a modulated wave and transmits the generated modulated wave from the transmission antenna 105a.

[0045] The receiving circuit 106 includes a receiving antenna 106a. Multiple (four in the figure) receiving antennas 106a are provided to detect the azimuth angle of an object. The receiving circuit 106 performs signal processing on the received reflected waves. The reflected wave data generated by the signal processing is provided to the processor 101. The processor 101 analyzes the reflected wave data to detect the position (distance and azimuth angle) and radial velocity of the object.

[0046] The volatile memory 103 is a semiconductor memory such as an SRAM (Static Random Access Memory) or a DRAM (Dynamic Random Access Memory). The non-volatile memory 102 is a flash memory, a hard disk, a ROM (Read Only Memory), or the like. The non-volatile memory 102 stores an analysis program 110, which is a computer program, and data used to execute the analysis program 110. Each function of the infrastructure radio wave sensor 10 is achieved by the processor 101 executing the analysis program 110. The analysis program 110 can be stored in a recording medium such as a flash memory, a ROM, or a CD-ROM. The processor 101 can identify the position (distance and azimuth) and line-of-sight velocity of an object using the analysis program 110.

[0047] The processor 101 is, for example, a CPU (Central Processing Unit). However, the processor 101 is not limited to a CPU. The processor 101 may also be a GPU (Graphics Processing Unit). The processor 101 may also be, for example, an ASIC (Application Specific Integrated Circuit) or a programmable logic device such as an FPGA (Field Programmable Gate Array). In this case, the ASIC or programmable logic device is configured to be able to execute the same processing as the analysis program 110.

[0048] The communication interface 108 can communicate with an external device. For example, the communication interface 108 is a wireless communication interface, and can transmit detection result data (hereinafter also referred to as "detection result data") to a server operated by a road traffic information center or the like, a control device for controlling traffic signals installed at an intersection including the crosswalk 20, or a vehicle (or an on-board device installed in) around the infrastructure radio wave sensor 10.

[0049] The non-volatile memory 102 stores setting information 111 including information on the above-described virtual detection area 300, zebra areas 301a and 301b, waiting areas 302a and 302b, and median strip area 303. The setting information 111 further includes information on mask areas 304a, 304b, and 304c.

[0050] 6 is a functional block diagram showing an example of the functions of the infrastructure radio wave sensor according to the embodiment. When the processor 101 executes the analysis program 110, the infrastructure radio wave sensor 10 functions as a detection unit 121 and an analysis unit 122.

[0051] The transmitter / receiver 104 repeats a transmission cycle in which it continuously transmits a plurality of chirp signals. A chirp signal is a radio wave signal whose frequency changes over time at a constant rate of change.

[0052] The detection unit 121 detects the position (the distance from the infrastructure radio wave sensor 10 to the object and the azimuth angle of the location where the object is located) and speed (line of sight speed) of the object based on the reflected waves that are generated when radio waves are irradiated onto the object and reflected by the object.

[0053] Specifically, the detection unit 121 generates reflected wave data that indicates information including the signal level of the reflected wave for each position irradiated with the radio wave. The transmission circuit 105 transmits a transmission signal, which is a modulated wave, from the transmission antenna 105a. The transmission signal from the transmission antenna 105a hits an object and is reflected. The receiving antenna 106a receives the reflected wave from the object. The receiving circuit 106 combines the modulated wave signal output from the transmission circuit 105 with the reflected wave signal received by the receiving antenna 106a to generate an intermediate frequency signal (hereinafter referred to as an "IF signal"). The detection unit 121 performs a fast Fourier transform (FFT) on the IF signal to obtain information on distance, speed, and azimuth angle. The detection unit 121 generates reflected wave data based on the obtained distance and azimuth angle information.

[0054] The detection unit 121 extracts reflection points, which are peak points included in the reflected wave data. The reflected wave data includes data indicating the waveform of the reflected wave for distance and data indicating the waveform of the reflected wave for angle. The detection unit 121 extracts peak points from each of the waveform of the reflected wave for distance and the waveform of the reflected wave for angle. The detection unit 121 determines the reflection points by associating the peak points in the reflected wave for distance with the peak points in the reflected wave for angle.

[0055] Radio waves emitted from the infrastructure radio wave sensor 10 may be reflected by multiple objects simultaneously. The detection unit 121 groups reflection points on the same object. The detection unit 121 identifies the position of each object by treating each group as one object. The position of the object is expressed as coordinate values ​​in a unique coordinate system (hereinafter also referred to as "unique coordinate values"). Specifically, the detection unit 121 determines a representative value of the reflection points belonging to the same group and sets the determined representative value as the position of the object. For example, the representative value is the center of gravity. However, the position of the object may be a representative value other than the center of gravity of multiple reflection points. For example, the representative value may be the average value of the reflection points or the median value of the reflection points.

[0056] The detection unit 121 outputs detection data including the position of the object detected as described above. The detection data includes not only the position information of the object (coordinate values ​​in the unique coordinate system) but also the velocity information of the object.

[0057] The analysis unit 122 determines an area that includes the detected object based on the position of the object detected by the detection unit 121. Specifically, the analysis unit 122 determines in which of the zebra areas 301a and 301b, the waiting areas 302a and 302b, the median area 303, and the mask areas 304a, 304b, and 304c the unique coordinate values ​​of the object are included. For example, the zebra areas 301a and 301b, the waiting areas 302a and 302b, the median area 303, and the mask areas 304a, 304b, and 304c are each a polygon (a rectangle, in one example), and the setting information 111 includes the unique coordinate values ​​of the vertices of the zebra areas 301a and 301b, the waiting areas 302a and 302b, the median area 303, and the mask areas 304a, 304b, and 304c. The analysis unit 122 determines whether or not the detected object is included in each of the zebra areas 301 a and 301 b, the waiting areas 302 a and 302 b, the median strip area 303, and the mask areas 304 a, 304 b, and 304 c, which are determined by the unique coordinate values ​​of the vertices. The analysis unit 122 determines the area in which the object is included for each of the detected objects.

[0058] The analysis unit 122 determines an analysis process for the object detection data based on the area containing the detected object. For example, a detection data analysis process is associated with each of the zebra areas 301a and 301b, the waiting areas 302a and 302b, the median area 303, and the mask areas 304a, 304b, and 304c. The analysis unit 122 analyzes, by a first analysis process, first detection data obtained by the detection unit 121 detecting a first object present in a first area of ​​the road. The analysis unit 122 analyzes, by a second analysis process, second detection data obtained by the detection unit 121 detecting a second object present in a second area of ​​the road. For example, the analysis unit 122 analyzes the object detection data by an analysis process corresponding to the area in which the object is present (one or more of the zebra areas 301a, 301b, waiting areas 302a, 302b, median strip area 303, and mask areas 304a, 304b, 304c).

[0059] The analysis process includes a tracking process for tracking detected objects. Specifically, the analysis unit 122 assigns an ID to each object detected by the detection unit 121. The detection unit 121 outputs detection data on the position and speed of the object at regular time intervals. The analysis unit 122 identifies, among the objects detected this time, an object that is the same as the object detected previously. For example, the analysis unit 122 estimates the current position of object a based on the movement direction and speed of object a previously. Of the objects detected this time, the analysis unit 122 identifies the object closest to the position estimated from the movement direction and speed of object a previously as object a. An object identified as the same as the object detected previously inherits the ID of the object detected previously.

[0060] FIG. 7 is a diagram illustrating the movement of objects in each area. In the zebra areas 31a and 31b, pedestrians M1, M2, and M3 cross the crosswalk 20 and vehicles V travel on the roadway 60. The pedestrians M1, M2, and M3 move along the longitudinal direction y of the crosswalk 20. The vehicle V moves along the width direction of the crosswalk 20, i.e., the longitudinal direction of the roadway 60. Furthermore, the movement speeds of the pedestrians M1, M2, and M3 are low (several kilometers per hour), while the movement speed of the vehicle V is high (several tens of kilometers per hour). That is, in the zebra areas 31a and 31b, multiple types of objects (pedestrians and vehicles) move in multiple directions at multiple speeds. In particular, when the traffic signal for the crosswalk 20 is lit in the color allowing passage (blue) (when passage on the crosswalk 20 is permitted), the pedestrian M1 and the vehicle V turning right or left from the roadway 65 may be present in the zebra area 31b of the outgoing lanes 62a, 62b at the same time. When the traffic signal for the crosswalk 20 is lit in the color indicating a stop (red) (when passage on the crosswalk 20 is prohibited), only the vehicle V moves in the zebra areas 31a, 31b.

[0061] Meanwhile, in the waiting areas 32a and 32b, pedestrians M4, M5, and M6 are present, but no vehicle V is present. In the waiting areas 32a and 32b, pedestrians M4, M5, and M6 move in various directions. That is, in the waiting areas 32a and 32b, one type of object (pedestrians) moves in multiple directions. When the traffic signal for the crosswalk 20 is lit in the proceeding color (blue) (when passage through the crosswalk 20 is permitted), pedestrians M4 and M6 who are attempting to cross the crosswalk 20 move in the waiting areas 32a and 32b toward the crosswalk 20 (i.e., the longitudinal direction y of the crosswalk 20). Pedestrian M5 who is not attempting to cross the crosswalk 20 moves in a direction different from the direction toward the crosswalk 20 in the waiting area 32a. When the traffic signal for crossing the crosswalk 20 is lit in a stop indication color (red) (when crossing the crosswalk 20 is prohibited), pedestrians M4 and M6 who are about to cross the crosswalk 20 stop in waiting areas 32a and 32b.

[0062] A pedestrian M7 is present in the median strip area 33, but no vehicle V is present. In the median strip area 33, the pedestrian M7 moves in various directions. That is, in the median strip area 33, one type of object (a pedestrian) moves in multiple directions. When the traffic signal for the crosswalk 20 is lit in the proceeding color (green) (when passage through the crosswalk 20 is permitted), the pedestrian M7, who is about to cross the crosswalk 20, moves in the longitudinal direction y of the crosswalk 20 in the median strip area 33. For example, if the traffic signal changes from the proceeding color (blue) to the stop color (red) while the pedestrian M7 is crossing the crosswalk 20, the pedestrian M7 will wait in the median strip area 33. In this case, the pedestrian M7, who is about to cross the crosswalk 20, will stop in the median strip area 33.

[0063] For example, in the waiting area 32a close to the installation position of the infrastructure radio wave sensor 10, pedestrians M4 and M5 are unlikely to be obscured by large vehicles or the like. Therefore, the waiting area 32a is an area where the infrastructure radio wave sensor 10 can easily detect pedestrians M4 and M5. The analysis unit 122 tracks objects (pedestrians M4 and M5) present in the waiting area 32a, which is less susceptible to noise, using a highly sensitive tracking process (hereinafter also referred to as a "first pedestrian tracking process").

[0064] For example, in the waiting area 32b and the median strip area 33, which are far from the installation position of the infrastructure radio wave sensor 10, pedestrians M6 and M7 may be hidden by large vehicles or the like. Therefore, the waiting area 32b and the median strip area 33 are areas where it is difficult for the infrastructure radio wave sensor 10 to detect pedestrians M6 and M7. The analysis unit 122 tracks objects (pedestrians M6 and M7) present in the waiting area 32b and the median strip area 33, which are susceptible to the influence of noise, using a tracking process with low sensitivity (hereinafter also referred to as a "second pedestrian tracking process"). Hereinafter, the sensitivity of the tracking process will also be referred to as the "tracking sensitivity."

[0065] FIG. 8 is a diagram for explaining an example of a tracking process when a target object is hidden by an obstruction.

[0066] The detection unit 121 periodically outputs detection data at a fixed cycle. The analysis unit 122 periodically performs tracking processing at the above cycle (hereinafter also referred to as the "processing cycle"). In the following description, it is assumed that an object position T0 is detected in processing cycle C0, and the subsequent processing cycles are C1, C2, C3, C4, C5, and C6 in order. If an object (pedestrian) is detected at position TP0 in processing cycle C0 and the target object is hidden by an obstruction OB in the next processing cycle C1, the analysis unit 122 estimates the object's position EP1 in processing cycle C1 based on the object's position TP0, movement direction, and movement speed detected in processing cycle C0. In processing cycle C1, the target object is hidden by the obstruction OB, so the object is not detected near the estimated position EP1. Therefore, the analysis unit 122 sets the tracking status of the target object to a failed state, indicating that tracking failed.

[0067] In the next processing cycle C2, the target object is still hidden by the obstruction OB. Therefore, the analysis unit 122 estimates the object's position EP2 in processing cycle C2 based on the object's position TP0, movement direction, and movement speed detected in processing cycle C0. The tracking state remains in the failed state in processing cycle C2 as well.

[0068] Similarly, in processing cycles C3, C4, and C5, the target object is hidden by an obstruction OB. Therefore, in processing cycles C3, C4, and C5, the analysis unit 122 estimates object positions EP3, EP4, and EP5 based on the object position TP0, movement direction, and movement speed detected in processing cycle C0. During this time, the tracking state remains in a failed state.

[0069] In processing cycle C6, the target object moves outside the obstruction OB. This allows the detection unit 121 to detect the position TP6 of the target object. For example, the analysis unit 122 estimates the position of the object in processing cycle C6 (i.e., the present) based on the position TP0, movement direction, and movement speed of the object detected in processing cycle C0. Since the object is detected at position TP6, which is close to the estimated position, the analysis unit 122 determines position TP6 as the position of the target object. In this case, the tracking state is set to a successful state, indicating that tracking was successful.

[0070] As described above, in the tracking process, the current position of an object is estimated based on the positions of the object detected in the past. In the tracking process, the number of processing cycles to go back in time (hereinafter also referred to as the "number of retroactive cycles") is set. In the above example, in processing cycle C6, the current position of the object (position in processing cycle C6) is estimated based on the position TP0 of the object detected in processing cycle C0, which is six cycles earlier. In other words, in the above example, the number of retroactive cycles is six or more.

[0071] In the tracking process, a failure state is allowed for a predetermined number of processing cycles. In the above example, the failure state is allowed to be maintained for five cycles, C1 to C5. The number of retroactive cycles is equal to one less than the number of processing cycles in which the failure state is allowed to continue (hereinafter also referred to as the "allowable number of continued cycles"). For example, in the above example, if the number of retroactive cycles is 6, the allowable number of continued cycles is 5.

[0072] In other words, the number of retroactive cycles is the period of time that goes back to the past (hereinafter also referred to as the "retroactive period"). If the processing cycle is T (ms) and the number of retroactive cycles is N, the retroactive period is T x N (ms).

[0073] The tracking sensitivity mentioned above is a retrospective period. In other words, the longer the retrospective period, the higher the tracking sensitivity, and the shorter the retrospective period, the lower the tracking sensitivity. For example, the retrospective period of the first pedestrian tracking process in the waiting area 302a is longer than the retrospective period of the second pedestrian tracking process in the waiting area 302b and the median strip area 303. In this case, the waiting area 302a corresponds to the "first area," and the first pedestrian tracking process corresponds to the "first tracking process." The waiting area 302b and the median strip area 303 correspond to the "second area," and the second pedestrian tracking process corresponds to the "second tracking process." The retrospective period in the first pedestrian tracking process corresponds to the "first period," and the retrospective period in the second pedestrian tracking process corresponds to the "second period."

[0074] Returning to FIG. 6 , as described above, the zebra areas 31a and 31b are areas where multiple types of objects (pedestrians and vehicles) are present. On the other hand, the waiting areas 32a and 32b and the median strip area 33 are areas where only a single type of object (pedestrians) is present. The analysis unit 122 identifies the type of object detected in the zebra areas 301a and 301b. That is, the analysis process includes an identification process for identifying the type of detected object. The analysis unit 122 does not identify the type of object detected in the waiting areas 302a and 302b and the median strip area 303.

[0075] In the identification process, the type of object is identified based on the movement direction of the detected object. For example, the analysis unit 122 identifies the type of object moving along the Y direction as a "pedestrian" and the type of object moving along the X direction as a "vehicle." Here, "moving along the Y direction" means, for example, that when the movement direction of the object is resolved into the X direction component and the Y direction component, the Y direction component is larger than the X direction component, and "moving along the X direction" means, for example, that the X direction component of the movement direction of the object is larger than the Y direction component.

[0076] In another example, the analysis unit 122 may identify the type of object based on the moving speed of the object. For example, a speed threshold T is set, and if the moving speed v is greater than T, the type of object may be identified as a "vehicle," and if the moving speed v is equal to or less than T, the type of object may be identified as a "pedestrian." In yet another example, the analysis unit 122 may identify the type of object based on the moving direction and moving speed of the object.

[0077] The analysis unit 122 performs a tracking process on an object identified as a "pedestrian" by the identification process, and performs a tracking process on an object identified as a "vehicle" by the identification process. For example, the tracking sensitivity in the tracking process for a "pedestrian" is different from the tracking sensitivity in the tracking process for a "vehicle." More specifically, in the infrastructure radio wave sensor 10 for pedestrian detection, the tracking sensitivity in the tracking process for a "pedestrian," which is the detection target, is higher than the tracking sensitivity in the tracking process for a "vehicle," which is not the detection target (hereinafter also referred to as "vehicle tracking process"). For example, the tracking sensitivity for a "pedestrian" in the zebra areas 301a and 301b may be different from or the same as the tracking sensitivity in the waiting area 302a. For example, the tracking sensitivity for a "pedestrian" in the zebra areas 301a and 301b may be different from or the same as the tracking sensitivity in the waiting area 302b and the median strip area 303. The zebra areas 31a and 31b have fewer obstructions, making it easier to detect objects. Therefore, in a specific example, the tracking sensitivity for "pedestrians" in the zebra areas 301a and 301b is the same as the tracking sensitivity in the waiting area 302a. That is, the analysis unit 122 executes a first pedestrian tracking process for "pedestrians" in the zebra areas 301a and 301b. In this case, the zebra areas 301a and 301b correspond to a "first area," and the waiting areas 302a and 302b and the median strip area 303 correspond to a "second area." Pedestrians correspond to a "first type of object," and vehicles correspond to a "second type of object." The retrospective period for "pedestrians" in the zebra areas 301a and 301b corresponds to a "third period," and the retrospective period for "vehicles" in the zebra areas 301a and 301b corresponds to a "fourth period."

[0078] As described above, non-moving objects are constantly detected in the mask areas 304 a, 304 b, and 304 c. Therefore, the analysis unit 122 executes a special tracking process (hereinafter also referred to as a “third pedestrian tracking process”) in the mask areas 304 a, 304 b, and 304 c.

[0079] 9 is a diagram for explaining an example of tracking an object in a mask area. In this embodiment, part of the detection data of the object included in the mask areas 304a, 304b, and 304c is not discarded.

[0080] 9 shows an example of object tracking in the mask area 304a. The third pedestrian tracking process tracks an object moving from outside the mask area 304a into the mask area 304a. The third pedestrian tracking process discards detection data of an object that is constantly present in the mask area 304a.

[0081] In the mask area 304a, a traffic light 65a, which is a non-moving object, is steadily detected at a position TP20.

[0082] On the other hand, a target object (pedestrian) is detected at a position TP10 in a processing cycle C10, and at a position TP11 in a processing cycle C11. Both positions TP10 and TP11 are outside the mask area 304a.

[0083] In the next processing cycle C12, the target object moves to position TP12 within the mask area 304a. In this embodiment, the analysis unit 122 does not discard the detection data of the object detected at position TP12 within the mask area 304a in the processing cycle C12. Furthermore, in the next processing cycle C13, the target object also moves to position TP13 within the mask area 304a. The analysis unit 122 does not discard the detection data of the object detected at position TP13 within the mask area 304a in the processing cycle C13.

[0084] In processing cycle C14, the target object is detected at position TP14 outside the mask area 304a. In the third pedestrian tracking process, as described above, tracking of an object moving within the mask area 304a is performed. This allows the object to be tracked without failure, even if, for example, the object's direction of movement changes within the mask area 304a.

[0085] Meanwhile, in the third pedestrian tracking process, the analysis unit 122 discards the detection data of a non-moving object that is constantly detected at position TP20 in the mask area 304a, thereby removing noise that is constantly detected in the mask area 304a.

[0086] Furthermore, in the third pedestrian tracking process, the analysis unit 122 does not start tracking of an object newly detected in the mask areas 304a, 304b, and 304c. For example, in the third pedestrian tracking process, the analysis unit 122 discards detection data of an object newly detected in the mask areas 304a, 304b, and 304c. This makes it possible to prevent a non-moving object that is constantly present in the mask areas 304a, 304b, and 304c from being erroneously detected as a new object.

[0087] In some cases, a pedestrian and a non-moving object may be detected as a single object in the mask areas 304a, 304b, and 304c. In this case, in the third pedestrian tracking process, the analysis unit 122 may treat the object detected in the mask areas 304a, 304b, and 304c as a tracking target. In other words, in this case, the detection data of the object in the mask areas 304a, 304b, and 304c is not discarded. In another example, in the third pedestrian tracking process, the analysis unit 122 may not treat the object detected in the mask areas 304a, 304b, and 304c as a tracking target. In other words, in this case, the detection data of the object in the mask areas 304a, 304b, and 304c is discarded.

[0088] For example, the analysis unit 122 may track an object detected at a position other than the detection area 300 and the mask areas 304a, 304b, and 304c. Specifically, the analysis unit 122 tracks an object detected on the sidewalks 63a and 63b (the corresponding positions in the inherent coordinate system) other than the waiting areas 302a and 302b. In yet another example, the analysis unit 122 tracks an object detected on the roadway 60 (the corresponding positions in the inherent coordinate system) other than the zebra areas 301a and 301b. For example, in the tracking process for an object detected outside the detection area 300 (hereinafter also referred to as the "outside area tracking process"), the tracking sensitivity can be lower than that in the first pedestrian tracking process because the object is not a detection target. In one specific example, the tracking sensitivity in the outside area tracking process is lower than that in the second pedestrian tracking process. As another example, the tracking sensitivity in the outside area tracking process may be the same as the tracking sensitivity in the second pedestrian tracking process.

[0089] 6 , the analysis unit 122 outputs information about the detected objects. Specifically, the analysis unit 122 outputs the position, speed, and ID of the objects (pedestrians and vehicles). The information (detection result data) output from the analysis unit 122 is provided to the communication interface 108. The communication interface 108 transmits the detection result data to, for example, a server, a traffic signal control device, or surrounding vehicles.

[0090] 5. Operation of the Infrastructure Radio Wave Sensor The operation of the infrastructure radio wave sensor 10 according to the embodiment will be described below. The processor 101 executes the following operations using the analysis program 110. Fig. 10 is a flowchart showing an example of the operation of the infrastructure radio wave sensor according to the embodiment.

[0091] The radio wave sensor 10 emits radio waves from the transmitting antenna 105a. The radio waves transmitted from the transmitting antenna 105a are reflected by objects present within the radio wave irradiation area 40. The reflected waves are received by the receiving antenna 106a.

[0092] The transmitter / receiver 104 outputs reflected wave data including a transmission signal (modulated wave signal) from the transmitting antenna 105a and a reception signal (reflected wave signal) by the receiving antenna 106a to the detection processing unit 120. The processor 101 receives the reflected wave data (step S101).

[0093] The processor 101 combines the modulated wave signal and the reflected wave signal to generate an IF signal. The processor 101 extracts reflection points from the waveform of the reflected wave. The processor 101 performs a fast Fourier transform on the IF signal to measure the distance to an object, the radial velocity of the object, and the azimuth angle of the object. The processor 101 groups the reflection points, treats each group of reflection points as an object, and detects the position, moving direction, and speed of the object (step S102).

[0094] The processor 101 selects one object (detection data thereof) from the detected objects (step S103).

[0095] The processor 101 determines whether the unique coordinate values ​​of the selected object are included in any of the zebra areas 301a, 301b, waiting areas 302a, 302b, median strip area 303, and mask areas 304a, 304b, 304c, and determines the area in which the selected object is located (step S104).

[0096] If the selected object is present in the zebra areas 301a and 301b ("zebra areas 301a and 301b" in step S104), the processor 101 identifies the type of the selected object based on the movement direction of the detected object (step S105).

[0097] If the type of the selected object is identified as a "vehicle" ("vehicle" in step S106), the processor 101 executes a vehicle tracking process (step S107). After the vehicle tracking process, the processor 101 proceeds to step S112.

[0098] If the type of the selected object is identified as a "pedestrian" ("pedestrian" in step S106), the processor 101 executes a first pedestrian tracking process (step S108). After the first pedestrian tracking process, the processor 101 proceeds to step S112.

[0099] On the other hand, if the selected object is present in the waiting area 302a ("waiting area 302a" in step S104), the processor 101 executes a first pedestrian tracking process (step S108). After the first pedestrian tracking process, the processor 101 proceeds to step S112.

[0100] If the selected object is present in the waiting area 302b or the median strip area 303 ("waiting area 302b or median strip area 303" in step S104), the processor 101 executes a second pedestrian tracking process (step S109). After the second pedestrian tracking process, the processor 101 proceeds to step S112.

[0101] If the selected object is present in the mask areas 304a, 304b, and 304c ("mask areas 304a, 304b, and 304c" in step S104), the processor 101 executes a third pedestrian tracking process (step S110). After the third pedestrian tracking process, the processor 101 proceeds to step S112.

[0102] If the selected object is outside the detection area 300 and the mask areas 304a, 304b, and 304c ("outside the detection area and mask area" in step S104), the processor 101 executes an outside-area tracking process (step S111). After the outside-area tracking process, the processor 101 proceeds to step S112.

[0103] The processor 101 determines whether all detected objects have been selected (step S112). If an unselected object remains (NO in step S112), the processor 101 returns to step S103 and selects an unselected object.

[0104] If all of the detected objects have been selected (YES in step S112), the processor 101 causes the communication interface 108 to transmit (output) the detection results, including the object's position (coordinate values ​​in the unique coordinate system), movement direction, movement speed, and object ID, to an external device (a server, a traffic light control device, or a surrounding vehicle) (step S113). After outputting the detection results, the processor 101 returns to step S101. As a result, the radio wave sensor 10 outputs the detection results at the processing cycle.

[0105] [6. Modifications] In the above-described embodiment, the infrastructure radio wave sensor 10 is a radio wave sensor for monitoring pedestrian passage at the crosswalk 20, but is not limited to this. For example, the infrastructure radio wave sensor 10 may be a radio wave sensor for monitoring vehicle passage on a roadway, or a radio wave sensor for monitoring vehicle and pedestrian passage at an intersection.

[0106] [7. Supplementary Note] The embodiments disclosed herein are illustrative in all respects and are not restrictive. The scope of the present disclosure is defined by the claims, not the above-described embodiments, and includes meanings equivalent to the claims and all modifications within the scope thereof.

[0107] 10 Radio wave sensor (infrastructure radio wave sensor) 101 Processor 102 Non-volatile memory 103 Volatile memory 104 Transmitting / receiving unit 105 Transmitting circuit 105a Transmitting antenna 106 Receiving circuit 106a Receiving antenna 108 Communication interface 110 Analysis program 111 Setting information 120 Detection processing unit 121 Detection unit 122 Analysis unit 140 Housing 140a Transmitting / receiving surface 20 Crosswalk 25 Roadway 26A, 26B Sidewalk 30 Detection area 31O Point 31Y Projection center axis 31X Straight line 31a, 31b Zebra area 32a, 32b Waiting area 33 Central reservation area 34a, 34b, 34c Mask area 300 Detection area 301a, 301b Zebra area 302a, 302b Waiting area 303 Median strip area 304a, 304b, 304c Mask area 40 Radio wave irradiation area 50 Structure 51 Pole 52 Arm 60, 65 Roadway 61a, 61b Entering lane 62a, 62b Exiting lane 63a, 63b Sidewalk 64 Median strip 65a, 65b Traffic signal 65c Plants 70 Intersection M1, M2, M3, M4, M5, M6, M7 Pedestrian EP1, EP2, EP3, EP4, EP5 Estimated position TP0, TP6, TP10, TP11, TP12, TP13, TP14, TP20 Position

Claims

1. A radio wave sensor comprising: a detection unit that detects an object present on a road; and an analysis unit that analyzes, through a first analysis process, first detection data obtained by the detection unit detecting a first object present on a first area of ​​the road, and that analyzes, through a second analysis process, second detection data obtained by the detection unit detecting a second object present on a second area of ​​the road.

2. The radio wave sensor according to claim 1, wherein the first analysis process includes a first tracking process for tracking the first object, and the second analysis process includes a second tracking process for tracking the second object.

3. The radio wave sensor of claim 2, wherein the first tracking process includes a process of estimating the current position of the first object based on the past position of the first object detected by the detection unit a first period prior to the present, and the second tracking process includes a process of estimating the current position of the second object based on the past position of the second object detected by the detection unit a second period prior to the present that is shorter than the first period.

4. A radio wave sensor as described in any one of claims 1 to 3, wherein the first analysis process includes an identification process for identifying the type of the first object, and the second analysis process does not include an identification process for identifying the type of the second object.

5. The radio wave sensor according to claim 4, wherein the identification process is a process of identifying the type of the first object based on the direction of movement of the first object.

6. The radio wave sensor of claim 4 or claim 5, wherein the first analysis process is a process of estimating the current position of a first type of object identified by the identification process based on past positions of the first object detected by the detection unit in the past three periods going back from the present, and estimating the current position of a second type of object identified by the identification process based on past positions of the first object detected by the detection unit in the past four periods going back from the present that are shorter than the third period.

7. The radio wave sensor according to any one of claims 4 to 6, wherein the first area is an area including a crosswalk, and the second area is an area for pedestrians to wait before crossing the crosswalk.

8. The radio wave sensor according to claim 2, wherein the second tracking process is a process of tracking the first object moving from outside the second area into the second area and discarding detection data of a third object constantly present in the second area.

9. An analysis method comprising the steps of: detecting an object present on a road based on radio waves emitted from a radio wave sensor toward the road; analyzing first detection data obtained by detecting a first object present in a first area of ​​the road using a first analysis process; and analyzing second detection data obtained by detecting a second object present in a second area of ​​the road using a second analysis process.

Images