JP7658100B2 - Road information collection system, road information collection method, and road information collection program - Google Patents

Description

This invention relates to technology for detecting objects using radar.

The configuration described in Patent Document 1 includes a sensor that emits millimeter waves. The sensor emits millimeter waves at obstacles such as railroad crossings. A signal processing unit obtains the reflection intensity from the reflected waves obtained by this irradiation. A determination unit determines the presence or absence of an obstacle from background information and fluctuations in reflection intensity.

JP 2005-233615 A

However, in the configuration of Patent Document 1, false detection (including failure to detect) occurs due to the presence of a metal structure near an obstacle (target of irradiation) or due to the effects of reflection, attenuation, diffraction, etc. The results of this false detection are used to determine the presence or absence of an obstacle, so there is a risk that accurate detection results will not be obtained.

Therefore, the object of the present invention is to suppress false positives and improve detection accuracy.

According to one example of the present disclosure, the road information collection system includes a number of sensors that are installed near a target area on a road and detect objects located in the target area by irradiating radio waves, and a determination unit that determines a false detection as occurring when at least one of the multiple sensors fails to detect an object in the direction of the radio wave irradiation.

With this configuration, even when multiple sensors detect an object, false detections from each sensor can be suppressed, improving detection accuracy.

According to one example of the present disclosure, multiple sensors in a road information collection system may be installed so that at least some of their illumination directions face each other.

This configuration can reduce blind spots where objects cannot be detected.

According to one example of the present disclosure, the feature extraction unit of the road information collection system may compare the reception strengths of multiple sensors with the positions corresponding to the reception strengths to determine whether the objects are the same or different.

In this configuration, if the reception strengths received from multiple sensors are the same, it can be determined that they are the same object, and if the reception strengths received from multiple sensors are different, it can be determined that they are different objects. From this, if the reception strengths of the same object are different, it can be determined that the objects are overlapping.

According to one example of the present disclosure, the feature extraction unit of the road information collection system may determine the type of object based on the object's moving speed.

In this configuration, the strength of the signal received from the sensor varies depending on the object's moving speed. This property makes it possible to determine the type of object.

According to one example of the present disclosure, the integrated processing unit of the road information collection system may determine that the detection is correct if it determines that the objects overlap.

With this configuration, even if objects are overlapping and hidden, if it is possible to identify the object based on its position, etc., it can be detected correctly, improving detection accuracy.

According to one example of the present disclosure, the road information collection system may include a transmission unit that transmits the result of a positive detection determined by the integrated processing unit to a higher-level system.

In this configuration, only positive detections are sent, providing highly accurate information to the host system.

According to one example of the present disclosure, the integrated processing unit of the road information collection system may count the results of false positives and output a warning if the number of false positives exceeds a threshold.

In this configuration, the information can be managed to identify causes of likely false positives.

According to one example of the present disclosure, the multiple sensors of the road information collection system may be radio radar.

This configuration provides a sensor that is highly resistant to all climates.

This invention can reduce false positives and improve detection accuracy.

1 is a schematic diagram of a road information collection system according to a first embodiment of the present invention. 1 is a schematic diagram (front view) of a road information collection system according to a first embodiment of the present invention; 5A, 5B, and 5C are diagrams showing information reflected from an object in the road information collection system according to the first embodiment of the present invention. 1 is a block diagram of a road information collection system according to a first embodiment of the present invention. 2 is a flowchart of the road information collection system according to the first embodiment of the present invention. 2 is a flowchart of the road information collection system according to the first embodiment of the present invention. 10 is a flowchart of a road information collection system according to a second embodiment of the present invention.

Below, we will explain how to implement the present invention with reference to several figures.

・Application examples
First, an example to which the present invention is applied will be described with reference to Fig. 1. Fig. 1 is a schematic diagram of a road information collection system according to a first embodiment of the present invention. The road information collection system 10 includes a first detection device 100A and a second detection device 100B. The first detection device 100A includes a first sensor 101A and a communication device 102A. The second detection device 100B includes a second sensor 101B and a communication device 102B.

The first detection device 100A and the second detection device 100B are installed on the roadside. More specifically, the road 20 has side edges 21 and 22, and the first detection device 100A is installed at side edge 21 and the second detection device 100B is installed at side edge 22.

The first sensor 101A and the second sensor 101B continuously illuminate the target area 400, which is a planar area including the side edge 21 and the side edge 22 of the road 20. In other words, the first sensor 101A and the second sensor 101B are installed in a direction that allows them to illuminate the target area 400. At this time, the first sensor 101A and the second sensor 101B continuously illuminate the pedestrian 200 and the automobile 300 from when they enter the target area 400 while traveling on the road 20 until they exit the target area 400.

The target area 400 defines, for example, an area on the road 20 where an accident or the like may occur. Also, it is not limited to being installed on roads, but may be installed at railroad crossings, etc., and the detected results can be provided as information to trains.

The first sensor 101A and the second sensor 101B detect the presence of an object (in this case, a pedestrian 200 and a car 300) by emitting radio waves to the object and receiving reflected waves from the object.

As shown in FIG. 1, when the pedestrian 200 and the automobile 300 are viewed from the first sensor 101A (first detection device 100A), the pedestrian 200 moves to a position where he is hidden by the automobile 300. At this time, the first sensor 101A receives only the reflected waves from the automobile 300. Meanwhile, the second sensor 101B receives the reflected waves from both the pedestrian 200 and the automobile 300.

The first sensor 101A outputs the received reflected wave as a first detection result to the communication device 102A. The communication device 102A transmits the first detection result to a first feature extraction unit (not shown in FIG. 1). Similarly, the second sensor 101B outputs the received reflected wave as a second detection result to the communication device 102B. The communication device 102B transmits the second detection result to a second feature extraction unit (not shown in FIG. 1).

The first feature extraction unit uses the first detection result to calculate the distance from the first sensor 101A to the automobile 300 and the moving speed of the automobile 300. Similarly, the second feature extraction unit uses the second detection result to calculate the distance from the second sensor 101B to the pedestrian 200 and the automobile 300 and the moving speed of the pedestrian 200 and the automobile 300.

The first feature extraction unit and the second feature extraction unit output the results of identifying pedestrians 200 and automobiles 300 in the target area 400 (hereinafter, feature identification results) to an integration processing unit (not shown in FIG. 1).

The integration processing unit determines that the pedestrian 200 and the automobile 300 are present in the target area 400 based on the positional relationship between the pedestrian 200 and the automobile 300 determined from the feature identification results. As a result, when the target area 400 is viewed from the first detection device 100A side, the integration processing unit determines that the pedestrian 200 is hidden by the automobile 300, but that the pedestrian 200 is present behind the automobile 300. As a result, the integration processing unit transmits to the upper system a positive detection indicating that the pedestrian 200 and the automobile 300 are present in the target area 400.

The higher-level system is a system that determines whether a dangerous situation has occurred, such as a pedestrian 200 falling, in the target area 400, which is an area defined on a road.

Next, a case will be described where the first sensor 101A cannot receive reflected waves from the target area 400. If the first sensor 101A cannot receive reflected waves from the target area 400, but the second sensor 101B can receive reflected waves from the target area 400, the integrated processing unit determines that the reflected waves received from the second sensor 101B are a false detection.

A more detailed explanation will be given. The first sensor 101A does not receive the reflected wave, and the second sensor 101B receives the reflected wave. The second sensor 101B outputs the received reflected wave as a second detection result to the communication device 102B. The communication device 102B transmits the second detection result to the second feature extraction unit 150B. The second feature extraction unit 150B uses the second detection result to calculate the distance from the second sensor 101B to the pedestrian 200 and the automobile 300, and the movement speeds of the pedestrian 200 and the automobile 300. The second feature extraction unit 150B outputs the feature identification result from the second sensor 101B to the integration processing unit.

At this time, the integration processing unit determines based on the characteristic identification result that the result was received only from the second sensor 101B, and determines that the result is a false positive. The integration processing unit discards that result, and does not transmit the false positive result to the higher-level system.

In this way, the integrated processing unit discards false detection results and transmits only correct detection results to the host system via the transmission unit. The host system can use only correct detection results as detection results, improving detection accuracy.

Configuration example 1
FIG. 1 is a schematic diagram of a road information collection system according to a first embodiment of the present invention. FIG. 2 is a schematic diagram (front view) of a road information collection system according to a first embodiment of the present invention. FIG. 3(A), FIG. 3(B), and FIG. 3(C) are diagrams showing information reflected from an object in the road information collection system according to the first embodiment of the present invention. FIG. 4 is a block diagram of the road information collection system according to the first embodiment of the present invention. FIG. 5 is a flowchart of the road information collection system according to the first embodiment of the present invention. FIG. 6 is a flowchart of the road information collection system according to the first embodiment of the present invention.

A more specific configuration example will be explained using the configuration of the road information collection system 10 in Figure 1 and the configuration in Figure 2 described above.

As shown in Figures 1 and 2, the road information collection system 10 includes a first detection device 100A and a second detection device 100B. The first detection device 100A includes a first sensor 101A and a communication device 102A. The second detection device 100B includes a second sensor 101B and a communication device 102B.

The first detection device 100A and the second detection device 100B are installed along the side of the road. More specifically, the road 20 has side edges 21 and 22, and the first detection device 100A is installed at side edge 21 and the second detection device 100B is installed at side edge 22.

The first sensor 101A and the second sensor 101B are, for example, radio wave radars. The first sensor 101A and the second sensor 101B continuously irradiate a target area 400, which is a planar area including the side edge 21 and the side edge 22 of the road 20. Specifically, as shown in FIG. 2, the irradiation area 105A of the first sensor 101A and the irradiation area 105B of the second sensor 101B are set in a direction that allows them to irradiate the target area 400.

As shown in Figures 1 and 2, a pedestrian 200 and a car 300 are present within the illumination area 105A of the first sensor 101A. A pedestrian 200 and a car 300 are present within the illumination area 105B of the second sensor 101B. However, when viewed from the first sensor 101A side, the pedestrian 200 is hidden by the car 300. Therefore, the first sensor 101A cannot detect the pedestrian 200.

Figures 3(A), 3(B), and 3(C) are distribution diagrams and graphs showing the results of the first sensor 101A and the second sensor 101B receiving reflected waves from a pedestrian 200 and a car 300.

The distribution diagram shown in FIG. 3(A) shows a point cloud 500A indicating the results reflected from the first sensor 101A, and a point cloud 500B indicating the results reflected from the second sensor 101B. The portion where point cloud 500A and point cloud 500B are output together indicates a car 300. The portion where only point cloud 500B is output indicates a pedestrian 200.

Figure 3 (B) is a graph showing the reception strength of the reflected wave received by the first sensor 101A. The first sensor 101A detects the automobile 300 at a distance where the reception strength P1 peaks (a position based on the first sensor 101A). Figure 3 (C) is a graph showing the reception strength of the reflected wave received by the second sensor 101B. The second sensor 101B detects the pedestrian 200 and the automobile 300. In Figure 3 (C), the automobile 300 is detected at a distance where the reception strength P1 peaks (a position based on the second sensor 101B), and the pedestrian 200 is detected at a distance where the reception strength P2 peaks (a position based on the second sensor 101B).

When viewed from the first sensor 101A side, the pedestrian 200 is hidden by the automobile 300, and the first sensor 101A detects only the automobile 300. The second sensor 101B detects both the pedestrian 200 and the automobile 300.

Using the results above, we will explain the block diagram of the road information collection system according to the first embodiment of the present invention in Figure 4.

(In case of correct detection)
First, a case will be described where detection results are received from the first sensor 101A and the second sensor 101B and a positive detection is determined. The first sensor 101A outputs the received reflected wave to the communication device 102A as a first detection result. The communication device 102A transmits the first detection result to the first feature extraction unit 150A. The second sensor 101B outputs the received reflected wave to the communication device 102B as a second detection result. The communication device 102B transmits the second detection result to the second feature extraction unit 150B. The first sensor 101A and the second sensor 101B repeatedly perform detection in the target area 400.

The first feature extraction unit 150A uses the first detection result to calculate the distance from the first sensor 101A to the automobile 300 and the moving speed of the automobile 300. The first feature extraction unit 150A stores the detection results in a list (first correct detection list).

Similarly, the second feature extraction unit 150B uses the second detection result to calculate the distance from the second sensor 101B to the pedestrian 200 and the automobile 300, and the movement speeds of the pedestrian 200 and the automobile 300. The second feature extraction unit 150B lists the detection results and stores them (second correct detection list).

The first feature extraction unit 150A and the second feature extraction unit 150B may be configured to include a database that can identify objects (in this case, pedestrian 200 and automobile 300) using features such as their size (vehicle length), moving speed, and reception strength. Furthermore, the first feature extraction unit 150A and the second feature extraction unit 150B may be configured to have a function of predetermining or learning the results of identifying features and objects.

The first feature extraction unit 150A and the second feature extraction unit 150B output the first and second correct detection lists of the pedestrians 200 and the automobiles 300 in the target area 400 to the integration processing unit 160. Note that the first and second correct detection lists are lists of, for example, the positions, speeds, and vehicle lengths of the pedestrians 200 and the automobiles 300. The speed also includes the state when the pedestrians 200 and the automobiles 300 are stopped.

The integrated processing unit 160 uses the first correct detection list and the second correct detection list to identify the pedestrian 200 and the automobile 300. For example, a positional relationship is used for this identification. At this time, the first detection device 100A determines that the pedestrian 200 is hidden by the automobile 300 in the target area 400.

In other words, the integration processing unit 160 determines that the first detection device 100A cannot detect the pedestrian 200, but the second detection device 100B detects the pedestrian 200 and the automobile 300, and therefore determines that the pedestrian 200 and the automobile 300 are present in the target area 400. The integration processing unit 160 transmits this to the upper system via the transmission unit 170, stating that the first detection device 100A and the second detection device 100B have detected the target area 400 normally (correct detection).

(In case of false positive)
Next, a case where a detection result is received from either the first sensor 101A or the second sensor 101B and is determined to be an erroneous detection will be described.

The first detection result cannot be received from the first sensor 101A. On the other hand, the second sensor 101B receives reflected waves from the pedestrian 200 and the car 300 in the target area 400.

At this time, the second sensor 101B outputs the received reflected wave as the second detection result to the communication device 102B. The communication device 102B transmits the second detection result to the second feature extraction unit 150B.

The second feature extraction unit 150B uses the second detection result to calculate the distance from the second sensor 101B to the pedestrian 200 and the automobile 300, and the movement speed of the pedestrian 200 and the automobile 300. The second feature extraction unit 150B outputs the feature identification result from the second sensor 101B to the integration processing unit 160.

At this time, the integration processing unit 160 determines from the characteristic identification result that the result was received only from the second sensor 101B and judges the result to be a false detection. The integration processing unit 160 discards the result and does not transmit it to the higher-level system.

Based on the above configuration, the process flow of the road information collection system 10 will be explained using the flowcharts in Figures 5 and 6. First, the process up to when the first sensor 101A and the second sensor 101B identify an object will be explained using Figure 5.

(First sensor 101A)
The first sensor 101A irradiates the automobile 300 with millimeter waves and receives the reflected waves (S101). The first sensor 101A outputs a first detection result to the communication device 102A. The communication device 102A transmits the first detection result to the first feature extraction unit 150A (S102).

The first feature extraction unit 150A uses the first detection result to calculate the distance from the first sensor 101A to the automobile 300 and the moving speed of the automobile 300, and identifies the automobile 300 (S103).

(Second sensor 101B)
The second sensor 101B irradiates the pedestrian 200 and the automobile 300 with millimeter waves and receives the reflected waves (S101). The second sensor 101B outputs a second detection result to the communication device 102B. The communication device 102B transmits the second detection result to the second feature extraction unit 150B (S102).

The second feature extraction unit 150B uses the second detection result to calculate the distance from the second sensor 101B to the pedestrian 200 and the automobile 300, and the movement speeds of the pedestrian 200 and the automobile 300, and identifies the pedestrian 200 and the automobile 300 (S103).

The processing of the integration processing unit 160 will be described based on the results of FIG. 5 with reference to FIG.
(In case of correct detection)
The integration processing unit 160 receives the first detection result from the first feature extraction unit 150A (S111), and receives the second detection result from the second feature extraction unit 150B (S112).

If the first detection result and the second detection result can be detected (S113: Yes), the integration processing unit 160 determines that the first detection device 100A is hiding the pedestrian 200 behind the automobile 300 in the target area 400 based on the positional relationship between the pedestrian 200 and the automobile 300 obtained from the feature identification result, and determines that the detection is correct (S114).

As a result, the integrated processing unit 160 correctly detects that a pedestrian 200 and a vehicle 300 are present in the target area 400, and transmits this information to the upper system via the transmission unit 170 (S115).

(In case of false positive)
The integration processing unit 160 receives the first detection result from the first feature extraction unit 150A (S111). The integration processing unit 160 also receives the second detection result from the second feature extraction unit 150B (S112). At this time, the first detection result does not include the detected result.

If the first detection result cannot be detected (S113: No), the integration processing unit 160 determines from the characteristic identification result that the result was received only from the second sensor 101B, and determines the result as a false detection (S116). The integration processing unit 160 discards the result and does not transmit it to the higher-level system (S117).

In this way, only positive detection results can be sent to the higher-level system, and false detection information is not sent to the higher-level system. In other words, the higher-level system can obtain only positive detection results as information, thereby improving detection accuracy.

In the above configuration, an example in which two detection devices, a first detection device 100A and a second detection device 100B, are installed has been described. However, a configuration in which more detection devices are installed may also be used. In this case, the accuracy of detecting the target area 400 is further improved.

Configuration example 2
Next, a specific configuration of the road information collection system will be described with reference to Fig. 7. Fig. 7 is a flowchart of the road information collection system according to the second embodiment of the present invention.

The second embodiment differs in the processing that is performed after a false positive is determined. The rest of the difference is the same as in the first embodiment, and a description of the similar parts will be omitted.

The processing of the integration processing unit 160 will be described with reference to FIG.
(In case of correct detection)
The integration processing unit 160 receives the first detection result from the first feature extraction unit 150A (S211), and receives the second detection result from the second feature extraction unit 150B (S212).

If the first detection result and the second detection result can be detected (S213: Yes), the integration processing unit 160 determines that the first detection device 100A is hiding the pedestrian 200 behind the automobile 300 in the target area 400 based on the positional relationship between the pedestrian 200 and the automobile 300 obtained from the feature identification result, and determines that the detection is correct (S214).

As a result, the integrated processing unit 160 correctly detects that a pedestrian 200 and a vehicle 300 are present in the target area 400, and transmits this information to the upper system via the transmission unit 170 (S215).

(In case of false positive)
The integration processing unit 160 receives the first detection result from the first feature extraction unit 150A (S211). The integration processing unit 160 also receives the second detection result from the second feature extraction unit 150B (S212). At this time, the first detection result does not include the detected result.

If the first detection result cannot be detected (S213: No), the integrated processing unit 160 determines from the characteristic identification result that the result was received only from the second sensor 101B, and determines that the result is a false detection (S216).

The integration processing unit 160 counts the number of false positives (S217). The integration processing unit 160 determines whether the number of false positives exceeds a predefined threshold (S218). If the number of false positives does not exceed the threshold (S218: No), the integration processing unit 160 increments the number of false positives.

If the number of false positives exceeds the threshold (S218: Yes), the integrated processing unit 160 outputs a warning (S219).

In this way, by counting the number of times false positives occur, it is possible to identify situations in which false positives are more likely to occur, such as time, location, or weather.

On the other hand, the integrated processing unit 160 allows the higher-level system to obtain only positive detection results as information, thereby improving detection accuracy.

Furthermore, the correspondence between the configuration according to the present invention and the configuration according to the above-mentioned embodiment can be described as follows.
<Additional Notes>
A plurality of sensors (100A, 100B) are installed near a target area (400) on a road (20) and detect objects (200, 300) located in the target area (400) by irradiating the objects with radio waves;
an integration processing unit (160) that determines a false detection when at least one of the plurality of sensors (100A, 100B) cannot detect the object (200, 300) in the irradiation direction of the radio wave;
A road information collection system (10).

P1...reception intensity P2...reception intensity 10...road information collection system 20...road 21, 22...side edge 100A...first detection device 100B...second detection device 101A...first sensor 101B...second sensor 102A...communication device 102B...communication device 105A...illumination area 105B...illumination area 150A...first feature extraction unit 150B...second feature extraction unit 160...integration processing unit 170...transmission unit 200...pedestrian 300...car 400...target area 500A, 500B...point cloud

Claims (9)

a first detection device and a second detection device each having a sensor that is installed near a target area on a road and receives a reflected wave of a radio wave irradiated to the target area ;
a feature extraction unit that identifies features of an object located within the target area based on reflected waves of radio waves irradiated onto the target area and received by the first detection device, and identifies features of an object located within the target area based on reflected waves of radio waves irradiated onto the target area and received by the second detection device;
an integration processing unit that determines that an object located within the target region has been erroneously detected when the feature extraction unit determines that no object has been identified as a feature for the first detection device and that one or more objects have been identified as a feature for the second detection device ;
Equipped with
The first detection device and the second detection device have different directions of irradiating radio waves with respect to the target area,
When the number of objects whose features have been identified by the feature extraction unit for the first detection device is smaller than the number of objects whose features have been identified by the feature extraction unit for the second detection device, the integration processing unit determines that the objects located within the target area have not been erroneously detected if an object whose features have been identified by the feature extraction unit for the second detection device and whose features have not been identified by the feature extraction unit for the first detection device is hidden by another object whose features have been identified by the feature extraction unit for the first detection device when viewed from the first detection device side.
Road information collection system. The road information collection system according to claim 1 , wherein the first detection device and the second detection device are installed facing each other with the target area therebetween . The feature extraction unit identifies a position and a moving speed of an object within the target area as features of the object .
3. The road information collection system according to claim 1 or 2. The road information collection system according to claim 3, wherein the feature extraction unit determines the type of the object based on the moving speed of the object. 5. The road information collection system according to claim 1, wherein the integrated processing unit causes the transmission unit to transmit to a higher-level system a detection result that it determines that an object located within the target area has not been erroneously detected, and discards a detection result that it determines that an object located within the target area has been erroneously detected without causing the transmission unit to transmit the result to the higher-level system. 6. The road information collection system according to claim 1 , wherein the integration processing unit counts the number of times that an object located within the target area is determined to have been erroneously detected , and outputs a warning when this number exceeds a threshold value. 7. The road information gathering system according to claim 1, wherein the sensor is a radio wave radar. a computer of an information processing device to which a first detection device and a second detection device, each of which is installed near a target area on a road and has a sensor that receives a reflected wave of an electric wave irradiated to the target area , are connected;
a feature extraction step of identifying features of an object located within the target area based on the reflected waves of radio waves irradiated onto the target area by the first detection device and received by the second detection device, and identifying features of an object located within the target area based on the reflected waves of radio waves irradiated onto the target area and received by the second detection device;
an integration processing step of determining that an object located within the target region has been erroneously detected when no object has been identified as a feature for the first detection device and one or more objects have been identified as a feature for the second detection device in the feature extraction step;
Run
The first detection device and the second detection device have different directions of irradiating radio waves with respect to the target area,
The integration processing step is a step of judging that an object located within the target area has not been erroneously detected if the number of objects whose features have been identified for the first detection device in the feature extraction step is smaller than the number of objects whose features have been identified for the second detection device in the feature extraction step, and an object whose features have been identified for the second detection device in the feature extraction step but whose features have not been identified for the first detection device in the feature extraction step is hidden behind another object whose features have been identified for the first detection device in the feature extraction step when viewed from the first detection device side.
Methods for collecting road information. A first detection device and a second detection device, each of which is installed near a target area on a road and has a sensor that receives a reflected wave of an electric wave irradiated to the target area, are connected to a computer of an information processing device.
a feature extraction step of identifying features of an object located within the target area based on the reflected waves of radio waves irradiated onto the target area by the first detection device and received by the second detection device, and identifying features of an object located within the target area based on the reflected waves of radio waves irradiated onto the target area by the second detection device;
an integration processing step of determining that an object located within the target region has been erroneously detected when no object has been identified as a feature for the first detection device and one or more objects have been identified as a feature for the second detection device in the feature extraction step;
Run the command ,
The first detection device and the second detection device have different directions of irradiating radio waves with respect to the target area,
The integration processing step is a step of judging that an object located within the target area has not been erroneously detected if the number of objects whose features have been identified for the first detection device in the feature extraction step is smaller than the number of objects whose features have been identified for the second detection device in the feature extraction step, and an object whose features have been identified for the second detection device in the feature extraction step but whose features have not been identified for the first detection device in the feature extraction step is hidden behind another object whose features have been identified for the first detection device in the feature extraction step when viewed from the first detection device side.
Road information collection program.

Images