Detection method and sensor system

Abstract

Provided is a detection method for detecting an object by comparing a detection value of a radar sensor with a threshold value, the detection method comprising: a step for performing a first detection process for detecting the presence or absence of an object within a detection range using a first threshold value as the threshold value; and a step for performing, when an object is detected in the first detection process, a second detection process for detecting the presence or absence of an object in a specific range out of the detection range using a second threshold that affords a higher detection sensitivity than the first threshold value, the specific range being in the vicinity of the position of the detected object and on the radar sensor side.

Description

Technical Field

[0001] This specification discloses a detection method and a sensor system.

Background Art

[0002] Conventionally, it is known to detect surrounding objects by transmitting radio waves from a radar sensor mounted on a vehicle or the like, receiving the reflected waves, and comparing the generated detection values with a threshold value. For example, in Patent Document 1, normally, an object is detected using a standard threshold value, and when an object with a low reflection intensity such as a pedestrian is detected, the object is detected using a threshold value lower than the standard threshold value.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] As described above, by lowering the threshold value when an object with a low reflection intensity is detected, it is possible to appropriately detect an object with a low reflection intensity while preventing misdetection such as noise during normal times as much as possible. However, by lowering the threshold value, the tendency is for misdetection due to the influence of noise to increase. Therefore, further improvement is required to suppress the influence of noise in the state where the threshold value is lowered.

[0005] The main object of the present disclosure is to improve the detection accuracy by suppressing the influence of noise by appropriately changing the threshold value for object detection.

Means for Solving the Problems

[0006] The present disclosure has adopted the following means to achieve the above main object.

[0007] The detection method disclosed herein is: A detection method for detecting an object by comparing the detected value of a radar sensor with a threshold value, The steps include: executing a first detection process to detect the presence or absence of an object within the detection range using a first threshold as the threshold; If an object is detected in the first detection process, the second detection process is performed to detect the presence or absence of the object in a specific range within the detection range that is near the detection position of the object and on the radar sensor side, using a second threshold that has a higher detection sensitivity than the first threshold. The gist of this is that it includes the following:

[0008] In the detection method disclosed herein, if an object is detected in a first detection process using a first threshold, a second detection process is executed to detect the presence or absence of the object using a second threshold that has a higher detection sensitivity in a specific range. In the second detection process, instead of detecting the entire area near the object, a specific range is limited to the area near the object and on the radar sensor side. This makes it possible to accurately detect small parts associated with the object that are located on the radar sensor side while minimizing the influence of noise. Therefore, by appropriately changing the threshold for object detection, the influence of noise can be suppressed and detection accuracy can be improved. [Brief explanation of the drawing]

[0009] [Figure 1] A schematic diagram showing the configuration of a vehicle 10 equipped with a sensor system 20. [Figure 2] A schematic diagram showing the general configuration of the radar sensor 21. [Figure 3] A flowchart illustrating an example of data analysis processing. [Figure 4] A flowchart illustrating an example of the detection process. [Figure 5] An explanatory diagram showing an example of the CFAR processing overview. [Figure 6] An explanatory diagram showing an example of detection range A and detection sensitivity. [Figure 7] An explanatory diagram showing an example of detection range A, specific range As, and detection sensitivity. [Figure 8]A flowchart showing an example of data analysis processing for a modified example. [Figure 9] A flowchart illustrating an example of the re-detection process. [Figure 10] A flowchart showing an example of the process for detecting variations. [Modes for carrying out the invention]

[0010] Next, embodiments of the present disclosure will be described with reference to the drawings. Figure 1 is a schematic diagram showing the configuration of a vehicle 10 equipped with a sensor system 20. Figure 2 is a schematic diagram showing the configuration of a radar sensor 21.

[0011] Vehicle 10 is configured to transport goods while avoiding collisions with objects such as people and structures in its vicinity. This vehicle 10 may automatically travel to a destination along a predetermined route, or it may autonomously travel to a destination while arbitrarily changing the route according to the driving conditions. As shown in Figure 1, vehicle 10 comprises a vehicle body 12 equipped with a plurality of wheels 14 for driving, and a sensor system 20 having a radar sensor 21 provided at the front of the vehicle body 12. The sensor system 20 may also have a plurality of radar sensors 21 provided at the sides or rear of the vehicle body 12.

[0012] The vehicle body 12 comprises a drive unit 16 and a control unit 18 that controls the entire vehicle 10. The drive unit 16 includes a motor that rotates the wheels 14, a steering mechanism, and a battery that supplies power to the motor. The control unit 18 receives detection results from the sensor system 20 (radar sensor 21). Based on the input detection results, the control unit 18 outputs control signals to the drive unit 16 to drive while avoiding collisions with detected objects or to temporarily stop in order to avoid collisions.

[0013] The radar sensor 21 is configured as a sensor using, for example, a millimeter-wave radar, and detects the presence or absence of an object in the detection range A in front of the vehicle 10. Further, the radar sensor 21 adopts an FMCW (Frequency Modulated Continuous Wave) method in which a continuous wave of a frequency-modulated chirp signal is transmitted and received using radio waves as a detection method. Note that the detection method may be another method such as a pulse method.

[0014] As shown in FIG. 2, this radar sensor 21 includes a signal generation unit 22, a transmission antenna 23, a reception antenna 24, a mixer 25, an AD conversion unit 26, and a detection processing unit 27. The signal generation unit 22 generates and outputs a transmission signal Tx which is a continuous wave of a frequency-modulated signal. The transmission antenna 23 outputs a transmission wave based on the transmission signal Tx generated by the signal generation unit 22. The reception antenna 24 receives a reflected wave transmitted from the transmission antenna 23 and reflected by an object, and outputs a reception signal Rx. Note that the radar sensor 21 includes a plurality of transmission antennas 23 and reception antennas 24.

[0015] The mixer 25 inputs and mixes the transmission signal Tx and the reception signal Rx, and generates and outputs a signal IF whose frequency is reduced to an intermediate frequency by performing frequency conversion on the mixed signal. The AD conversion unit 26 outputs data obtained by converting the signal IF output from the mixer 25 into a digital signal. The detection processing unit 27 is a microcomputer including a CPU, a ROM, a RAM, etc. not shown in the figure. The detection processing unit 27 inputs the data (digital signal) output from the AD conversion unit 26 and executes FFT (Fast Fourier Transform) processing, CFAR (Constant False Alarm Rate) processing, etc. By the processing of the detection processing unit 27, the presence or absence of an object, the distance to the object, the relative speed, the angle, etc. are detected.

[0016] Next, the detection processing unit 27 of the radar sensor 21 will explain the process of analyzing data to detect an object. FIG. 3 is a flowchart showing an example of the data analysis process. In this data analysis process, it is executed by the detection processing unit 27 at every predetermined detection cycle such as, for example, several tens of milliseconds. The detection processing unit 27 first inputs data, which is a digital signal of the above-described signal IF, from the AD conversion unit 26 (S100). Next, the detection processing unit 27 performs FFT processing on the input data (S110), and then performs object detection processing (S120). The detection processing unit 27 performs well-known processes such as detecting the distance to the object by FFT processing on the input data, detecting the speed of the object by FFT processing on the result, and further detecting the angle by FFT processing on the result. Although the detection processing in S120 will be described later, for example, the presence or absence of an object is detected by performing a CFAR process of comparing the FFT-processed detection value with a threshold value.

[0017] Subsequently, the detection processing unit 27 determines whether or not an object has been detected in the detection processing of S120 (S130). If it is determined that no object has been detected, the result of no detection is stored in a storage unit such as a RAM (S140), and the data analysis process is terminated. On the other hand, if the detection processing unit 27 determines that an object has been detected, the result of detecting the object is stored in a storage unit such as a RAM together with position information such as the distance and angle of the object and speed information such as the relative speed (S150), and the data analysis process is terminated.

[0018] Next, the details of the detection process in S120 will be explained. Figure 4 is a flowchart showing an example of the detection process, and Figure 5 is an explanatory diagram showing an example of the overview of the CFAR process. Figure 5 illustrates cell-averaging CFAR processing, but is not limited to this. In CFAR processing, for example, the detected values ​​processed with FFT are used as the target cells (target values) for processing and are sequentially compared with a threshold Th, and an object is detected when it exceeds the threshold Th. For example, distance detected values ​​are used as the detected values ​​processed with FFT. Therefore, each cell in Figure 5 is arranged in order of distance from the radar sensor 21. Furthermore, the threshold Th in CFAR processing is set by a well-known method as follows. That is, the average value M is calculated by adding the sum of the values ​​of the n / 2 values ​​on the left and the sum of the values ​​of the n / 2 values ​​on the right of the multiple reference cells (reference values) on both sides of the target cell, and dividing the sum by the number of values ​​n. The threshold Th is set by multiplying this average value M by a coefficient k. In this embodiment, coefficient k1 and coefficient k2, which is smaller than coefficient k1, are stored as coefficient k in a storage unit such as ROM. Note that CFAR processing is not limited to being implemented by software, but may also be implemented by hardware.

[0019] In the detection process shown in Figure 4, the detection processing unit 27 first determines whether or not an object was detected in the previous data analysis process (S200). This determination is based on the result stored in either S140 or S150 of the previous data analysis process. If there is no previous result, such as at the start of the data analysis process, S200 determines that no object was detected. If the detection processing unit 27 determines that no object was detected, it executes a first detection process to detect the presence or absence of an object using the first threshold Th1 as the threshold Th for the CFAR process (S210).

[0020] In this embodiment, a first threshold Th1 is set by multiplying the average value M by the coefficient k1. If the value of the target cell exceeds the first threshold Th1, it is determined that detection has occurred, and if it is less than or equal to the first threshold Th1, it is determined that no detection has occurred. The coefficient k1 is set so that this first threshold Th1 is a value suitable for detecting relatively large parts of the human body, such as the torso. Furthermore, since the coefficient k1 is a larger value than the coefficient k2, a larger threshold Th (first threshold Th1) is set than when the average value M is multiplied by the coefficient k2. As a result, the value of the target cell is less likely to exceed the threshold Th, resulting in lower detection sensitivity. When the detection processing unit 27 executes the first detection process in S210, it determines whether the processing of all data (target cells) has been completed (S220). If it determines that the processing is not complete, it returns to S200 and repeats the process. Also, when the detection processing unit 27 determines in S220 that the processing of all data has been completed, it terminates the detection process.

[0021] Figure 6 is an explanatory diagram showing an example of detection range A and detection sensitivity. In Figure 6, the previous result is shown as no detection. In this case, as shown in the figure, detection processing is performed using a low-sensitivity first threshold Th1 throughout the entire detection range A. Therefore, it is possible to detect relatively large bodies such as torsos while preventing false detections due to the influence of noise, making it possible to appropriately detect when a person enters the detection range A.

[0022] Furthermore, if the detection processing unit 27 determines in S200 of Figure 4 that an object was detected in the previous data analysis process, it determines whether a specific range As, which is a part of the detection range A, has already been set (S230). If the detection processing unit 27 determines that a specific range As has not been set, it sets a specific range As of a certain size near the detection position of the previously detected object and on the radar sensor 21 side (S240). The detection position of the object is based on position information such as the distance and angle of the object stored in S150 of the previous data analysis process. Next, the detection processing unit 27 determines whether the position of the target of processing this time, i.e., the target cell, is within the specific range As (S250). If the detection processing unit 27 determines that the target of processing is not within the specific range As, it proceeds to S210 and executes the first detection process using the first threshold Th1. On the other hand, if the detection processing unit 27 determines that the target of processing is within the specific range As, it executes a second detection process to detect the presence or absence of an object using the second threshold Th2 as the threshold Th (S260), and then proceeds to S220.

[0023] In this embodiment, a second threshold Th2 is set by multiplying the average value M by the coefficient k2. Detection is determined when the value of the target cell exceeds the second threshold Th2, and detection is determined when it is less than or equal to the second threshold Th2. The coefficient k2 is set so that this second threshold Th2 is a value suitable for detecting relatively small objects such as hands and arms. Furthermore, since the coefficient k2 is a smaller value than the coefficient k1, a second threshold Th2 smaller than the first threshold Th1 is set. As a result, the value of the target cell is more likely to exceed the threshold Th, resulting in high detection sensitivity.

[0024] Figure 7 is an explanatory diagram showing an example of detection range A, specific range As, and detection sensitivity. In Figure 7, the result from the previous test shows the case where a person's torso B is detected. As shown in the figure, the specific range As is set near the torso B and on the radar sensor 21 side. Therefore, if the processing target is within the specific range As, the second detection process is performed using the second threshold Th2, and if the processing target is outside the specific range As, the first detection process is performed using the first threshold Th1. In the specific range As, where the detection sensitivity is high, relatively small parts such as a person's hands and arms can be appropriately detected. That is, when the torso B is detected, it is possible to appropriately detect that the person's hands and arms are located on the radar sensor 21 side, and prevent the hands and arms from coming into contact with the vehicle 10. Also, since the first detection process is performed outside the specific range As, the influence of noise can be minimized. Note that within the specific range As, the influence of noise may occur, but since this is a limited process to enhance safety when the torso B is detected, it is unlikely to become a major problem. From these points, it is possible to accurately detect the hands and arms of the detected person (torso B) while suppressing the influence of noise.

[0025] Here, we will clarify the correspondence between the components of this embodiment and the components of this disclosure. In this embodiment, the radar sensor 21 corresponds to a radar sensor, and the detection processing unit 27 corresponds to a detection processing unit.

[0026] In the sensor system 20 described above, when an object is detected in the first detection process using the first threshold Th1, the presence or absence of the object is detected in the second detection process using the second threshold Th2, such that the detection sensitivity is high in the vicinity of the detected object and in a specific range As on the radar sensor 21 side. Therefore, it is possible to accurately detect small parts associated with the detected object that are located on the radar sensor 21 side while minimizing the influence of noise.

[0027] Furthermore, in the first detection process, a first threshold Th1 is used to enable the detection of a person's torso B, and in the second detection process, a second threshold Th2 is used to enable the detection of a person's hand or arm. Therefore, while the person's torso B is detected, relatively small parts such as hands and arms can be detected accurately while suppressing the effects of noise.

[0028] Furthermore, if an object is detected in the first detection process, the second detection process is executed in a specific range As in the next detection cycle. This prevents redundant processing in the same detection cycle while allowing for the appropriate detection of small parts associated with the object in the next detection cycle.

[0029] It goes without saying that this disclosure is not limited in any way to the embodiments described above, and can be implemented in various forms as long as they fall within the technical scope of this disclosure.

[0030] In the embodiment described above, when an object was detected in the first detection process, the second detection process was executed in a specific range As in the next detection cycle. However, the embodiment is not limited to this, and the following modified example may also be used. Figure 8 is a flowchart showing an example of the data analysis process in the modified example. In the following modified example, the same steps as in the embodiment are given the same step numbers and their explanations are omitted.

[0031] In the modified example shown in Figure 8, when the detection processing unit 27 determines that an object has been detected in S130, it executes the re-detection process shown in Figure 9 (S145). The re-detection process in Figure 9 is executed similarly, except that a part of the detection process in S120 (Figure 4) (S200, S210) is not performed. That is, the detection processing unit 27 sets the specific range As if it has not already been set (S230, S240), and if the object to be processed is within the specific range As, it executes the second detection process (S250, 260). In this modified example, the second detection process is executed for the specific range As of the object within the same detection cycle as the data analysis process that detected the object. However, depending on the detection position and relative velocity of the person (torso B), if the second detection process is executed after waiting for the next detection cycle, the risk of proximity or collision with hands, arms, etc. may increase. In other words, if safety is a priority, the time spent waiting for the next detection cycle cannot be overlooked. Therefore, by performing a re-detection process (second detection process) within the same detection cycle, as shown in the modified example, hands, arms, etc., can be properly detected, thereby improving safety.

[0032] On the other hand, if the re-detection process is performed within the same detection cycle as in the modified example, the detection process will be performed twice in S120 and S145, which may lead to delays in data analysis and processing. For this reason, the detection processing unit 27 may, for example, perform the re-detection process if the detected position is within a predetermined distance from the radar sensor 21, based on the detected position of the object (body B), but may not perform the re-detection process if it is outside the predetermined distance.

[0033] In this embodiment, the specific range As is defined as a certain range near the object's detection position and on the radar sensor 21 side, but it is not limited to this. Figure 10 is a flowchart showing an example of a modified detection process. In this detection process, if the detection processing unit 27 determines in S230 that the specific range As has not been set, it calculates a value obtained by multiplying the relative velocity of the object based on the previous detection result by the detection period (time) (S235). The relative velocity of the object is based on the velocity information stored in S150 of the previous data analysis process. The value obtained by multiplying the relative velocity by the detection period corresponds to the distance the object has traveled. Next, the detection processing unit 27 sets the specific range As in the vicinity of the object's detection position and on the radar sensor 21 side according to the value calculated in S235 (S240b), and then executes the subsequent processing. Here, in S240b, the specific range As should be set to include at least the range of the distance the object (torso B) has traveled as calculated. Alternatively, the specific range As may be set by adding the range that hands, arms, etc., can reach to the range of the travel distance. In this modified version, the second detection process is performed within an appropriate specific range As that reflects the object's movement distance, thereby further improving detection accuracy while more effectively suppressing the effects of noise.

[0034] In this embodiment, a first threshold Th1 was used in the first detection process to enable the detection of a person's torso B, and a second threshold Th2 was used in the second detection process to enable the detection of a person's hands and arms, but the system is not limited to this. In the first detection process, a threshold capable of detecting a relatively large portion of an object should be used, and in the second detection process, a threshold that provides a higher detection sensitivity than the first detection process should be used. Furthermore, in the detection process, it was determined that an object had been detected when the threshold was exceeded in the CFAR process, but the system is not limited to this, and any process that detects the presence or absence of an object by comparing the detected value of the radar sensor 21 with a threshold may be used.

[0035] In this embodiment, the radar sensor 21 is exemplified as being used in a vehicle 10, but it is not limited to this, and may be used in other mobile devices other than the vehicle 10, or in various types of work robots other than the vehicle 10.

[0036] Herein, the detection method of the present disclosure may be as follows. In the detection method of the present disclosure, the first threshold may be used in the first detection process so as to enable the detection of a human torso as an object, and the second threshold may be used in the second detection process so as to enable the detection of a human hand or arm as an object. In this way, while the human torso has been detected, relatively small parts such as hands and arms can be detected with high accuracy while suppressing the influence of noise.

[0037] In the detection method of this disclosure, the detection process can be executed at predetermined detection cycles, and if an object is detected in the first detection process, the second detection process may be executed in the specified range in the next detection cycle. In this way, by reflecting the previous detection result in the detection process of the next detection cycle, small parts attached to the object can be appropriately detected.

[0038] In the detection method of this disclosure, the detection process can be executed at predetermined detection cycles, and if an object is detected in the first detection process, the second detection process may be executed in the specific range at the same detection cycle. In this way, small parts attached to the object can be quickly detected without waiting for the next detection cycle.

[0039] In the detection method of the present disclosure, the first detection process can detect the relative velocity of the object with respect to the radar sensor from the detected value when the object is detected, and the second detection process may determine the specific range based on a value obtained by multiplying the relative velocity by a predetermined detection period. In this way, the second detection process is performed in an appropriate specific range that reflects the distance the object has moved, so that the detection accuracy can be improved while more effectively suppressing the effects of noise.

[0040] The sensor system disclosed herein is A sensor system that detects an object by comparing the detected value of a radar sensor with a threshold, A first detection process that detects the presence or absence of an object within the detection range using a first threshold as the threshold, If an object is detected in the first detection process, the detection processing unit performs a second detection process to detect the presence or absence of an object in a specific range within the detection range that is near the detection position of the object and on the radar sensor side, using a second threshold that has a higher detection sensitivity than the first threshold. This is the gist of it.

[0041] The sensor system of this disclosure, similar to the detection method described above, can improve detection accuracy by appropriately changing the threshold for object detection, thereby suppressing the effects of noise. In this sensor system, various embodiments of the detection method may be employed, or configurations that realize each step of the detection method may be added. [Industrial applicability]

[0042] This disclosure is applicable to technical fields such as object detection using radar sensors. [Explanation of Symbols]

[0043] 10 Vehicle, 12 Body, 14 Wheels, 16 Drive unit, 18 Control unit, 20 Sensor system, 21 Radar sensor, 22 Signal generation unit, 23 Transmitting antenna, 24 Receiving antenna, 25 Mixer, 26 AD conversion unit, 27 Detection processing unit, A Detection range, As Specific range, B Body.

Claims

1. A detection method for detecting an object by comparing the detected value of a radar sensor with a threshold value, The steps include: executing a first detection process to detect the presence or absence of an object within the detection range using a first threshold value set by multiplying a predetermined value by a first coefficient as the threshold value; If an object is detected in the first detection process, a specific range is set on the radar sensor side with respect to the detection position where the object was detected, such that it is included in the distance from the radar sensor and the azimuth angle of the detection position. Within the specific range of the detection range, a second detection process is performed to detect the presence or absence of an object using a second threshold value, which is set by multiplying a predetermined value by a second coefficient smaller than the first coefficient, so that the detection sensitivity is higher than the first threshold value. Includes, In the first detection process, the first threshold is used so that a human torso can be detected as an object. In the second detection process, the second threshold is used so that a human hand or arm can be detected as an object. Detection method.

2. The detection method according to Claim 1, The predetermined value is calculated as the average value obtained by referencing multiple data points centered on itself as reference values ​​when the detected values ​​of the radar sensor are arranged by distance. Detection method.

3. A detection method according to claim 1 or 2, The detection process can be executed at predetermined detection intervals. If an object is detected in the first detection process, the second detection process will be executed in the specified range in the next detection cycle. Detection method.

4. A detection method according to claim 1 or 2, The detection process can be executed at predetermined detection intervals. If an object is detected in the first detection process, the second detection process is executed in the specified range with the same detection cycle. Detection method.

5. A detection method according to any one of claims 1 to 4, In the first detection process, the relative velocity of the object with respect to the radar sensor can be detected from the detected value when the object is detected. In the second detection process, the specific range is determined based on a value obtained by multiplying the relative speed by a predetermined detection period. Detection method.

6. A sensor system that detects an object by comparing the detected value of a radar sensor with a threshold, A first detection process that detects the presence or absence of an object within the detection range using a first threshold value set by multiplying a predetermined value by a first coefficient, The detection processing unit performs the following steps: When an object is detected in the first detection process, it sets a specific range on the radar sensor side of the detection position where the object was detected, such that it is included in the distance from the radar sensor and the azimuth angle of the detection position, and in the specific range within the detection range, it detects the presence or absence of the object using a second threshold value, which is set by multiplying a predetermined value by a second coefficient smaller than the first coefficient, so that the detection sensitivity is higher than the first threshold value. In the first detection process, the first threshold is used so that a human torso can be detected as an object. In the second detection process, the second threshold is used so that a human hand or arm can be detected as an object. Sensor system.

Images