JP7007049B2 - Obstacle detection device - Google Patents

Description

The present invention relates to an obstacle detection device that detects the position of an obstacle by exploring the vicinity of the own vehicle with a plurality of direct detection sensors and indirect detection sensors at predetermined measurement time intervals and measuring the distance to the obstacle.

Conventionally, the area around the own vehicle is explored by at least a plurality of sensors at predetermined measurement times to measure the distance to an obstacle, and one sensor is used as a direct detection sensor and the other is used as an indirect detection sensor, which is transmitted by the direct detection sensor. Based on the reflected wave received by the direct detection sensor, the reflected wave received by the indirect detection sensor, and the distance between these sensors, the lateral position of the obstacle is calculated using the principle of the triangulation method, and then the lateral position of the obstacle is calculated. Using one sensor as an indirect detection sensor and the other as a direct detection sensor, the horizontal positions of obstacles are calculated in the same manner, and when the two obtained horizontal positions do not match, both of the two horizontal positions An obstacle detection device has been proposed in which the lateral position detected in the vicinity of the middle one of the sensors is determined to be due to a virtual image and is invalidated (Patent Document 1).

JP-A-2015-4562 (see paragraphs 0025 to 0035 and FIGS. 1 to 7)

However, the device described in Patent Document 1 described above is effective when detecting one and the same obstacle, but when detecting, for example, two obstacles existing in the vicinity of the own vehicle, one of the obstacles is detected. Even if the position of an object can be detected, the position of the other obstacle may not be detected, which causes a problem of lack of reliability.

An object of the present invention is to reliably detect the positions of two obstacles existing in the vicinity of the own vehicle and improve the reliability.

In order to achieve the above object, the obstacle detection device of the present invention explores the vicinity of the own vehicle by a plurality of direct detection sensors and indirect detection sensors at predetermined measurement time intervals, measures the distance to the obstacle, and measures the distance to the obstacle. In the obstacle detection device that detects the position of the obstacle, the triangulation estimation position derivation means that estimates the position of the obstacle by the triangulation method based on the distance measurement data by each sensor and derives the triangulation estimation position. For each of the sensors, an arc centered on the position of the sensor at the current time and a radius of the first measurement distance to the obstacle measured by the sensor at the current time, and the measurement once from the current time. The point where the arc overlaps with the arc centered on the position of the sensor before the time and whose radius is the second measurement distance to the obstacle measured by the sensor one time before the current time before the measurement time is the obstacle. The time-series estimation position derivation means derived as the time-series estimation position, the triangulation estimation position derived by the triangulation estimation position derivation means, and the time-series estimation position derivation means for each sensor. A comparison means for comparing whether the distance to the time-series estimated position is larger or smaller than a predetermined threshold, and the triangulation estimated position and the time when the distance is smaller than the predetermined threshold by the comparison means. It is characterized in that the series estimation position is determined to be the same obstacle, and if it is large, the triangulation estimation position is a virtual image and the time series estimation position is determined to be an obstacle.

According to the present invention, the triangulation estimation position derived by the triangulation estimation position derivation means for each of the direct detection sensor and the indirect detection sensor, and the time series estimation position derivation means for each of the direct detection sensor and the indirect detection sensor are derived. The comparison means compares whether the distance from the time-series estimated position is larger or smaller than the predetermined threshold, and if the result of the comparison is smaller than the predetermined threshold, the triangulation estimated position and the time-series estimated position are If it is judged to be the same obstacle, and if it is large, the estimated position of the triangulation is a phantom image and the estimated position of the time series is judged to be an obstacle by the judgment means. Objects can be reliably detected, and the reliability of obstacle detection can be improved.

It is a block diagram of one Embodiment of the obstacle detection apparatus which concerns on this invention. It is operation explanatory drawing of one Embodiment. It is operation explanatory drawing of one Embodiment. It is a flowchart for operation explanation of one Embodiment.

An embodiment of the obstacle detection device according to the present invention will be described in detail with reference to FIGS. 1 to 4.

The obstacle detection device in this embodiment is configured as shown in FIG. That is, for example, a direct detection sensor (hereinafter referred to as a P sensor) 1a and an indirect detection sensor (hereinafter referred to as a Q sensor) 1b composed of an ultrasonic sensor mounted on the front portion of the own vehicle, and both of these. A control means 2 including an ECU (Electronic Control Unit) that controls sensors 1a and 1b to search the front of the own vehicle every predetermined measurement time Δt (for example, 100 ms) and calculate the distance to an obstacle. Triangulation estimation position derivation means 3 that estimates the position of an obstacle by the triangulation method based on the distance measurement data by each sensor 1a, 1b and control means 2 and derives the triangulation estimation position, and each sensor 1a, 1b , The time-series estimated position of the obstacle is derived by superimposing the triangulation area to the obstacle at the current measurement time t and the triangulation area to the obstacle at the measurement time t-1 one time before the present. The time-series estimated position derivation means 4, the triangulation estimation position derived by the triangulation estimation position derivation means 3, and the time-series estimation position derived by the time-series estimation position derivation means 4 for each sensor 1a, 1b. If the distance is smaller than the predetermined threshold by the comparison means 5 for comparing whether the distance is larger or smaller than the predetermined threshold, the triangulation estimation position and the time series estimation position are the same obstacle. If it is large, the triangulation estimated position is a phantom image, and the time-series estimated position is determined to be an obstacle.

By the way, the operation of the triangulation estimation position derivation means 3 is a known technique as described in Patent Document 1 (Japanese Unexamined Patent Publication No. 2015-4562) described above, and detailed description thereof will be omitted, and the principle will be simplified. To explain to.

Now, as shown in FIG. 2A, the P sensor 1a and the Q sensor 1b are attached to the front part of the own vehicle C at positions symmetrical with respect to the center line L in the front-rear direction, and both sensors 1a and 1b are used to attach the P sensor 1a and the Q sensor 1b. Control means 2 when the ultrasonic waves are transmitted to the front of the vehicle C at predetermined measurement times to detect the distances to the two first and second obstacles B1 and B2 existing in front of the own vehicle C. Therefore, the distances from the P sensor 1a to the first and second obstacles B1 and B2 are calculated based on the reflected waves from the obstacles B1 and B2 received by the P sensor 1a by the ultrasonic waves, and the Q sensor 1b is calculated. The distances from the Q sensor 1b to the first and second obstacles B1 and B2 are calculated based on the reflected waves from the first and second obstacles B1 and B2 received by. At this time, the P sensor 1a is a direct detection sensor that transmits both ultrasonic waves and the reflected wave, and the Q sensor 1b is an indirect detection sensor that only receives the reflected wave.

Next, when the Q sensor 1b is used as a direct detection sensor and the P sensor 1a is used as an indirect detection sensor by the control means 2, the first and second obstacles B1 and 2 from the P sensor 1a and the Q sensor 1b, respectively, are similarly used. The distance to each of B2 is calculated.

Then, the time until the reflected wave is received by both sensors 1a and 1b by the triangulation estimation position derivation means 3, the calculated distance, the installation interval of both sensors 1a and 1b, and the sound velocity of the first obstacle B1. The estimated triangulation position α, which is the lateral position with respect to the center line L, is calculated. As shown in FIG. 2B, the round-trip path length between the P sensor 1a and the first obstacle B1 is the P sensor 1a. When it is smaller than the round-trip path length between and the second obstacle B2, the P sensor 1a first receives the reflected wave from the first obstacle B1 and then receives the reflected wave from the second obstacle B2. do.

On the other hand, when the path length from the P sensor 1a to the Q sensor 1b via the first obstacle B1 is smaller than the path length from the P sensor 1a to the Q sensor 1b via the obstacle B2, the Q sensor 1b is first first. Since the reflected wave from the first obstacle B1 is received and then the reflected wave from the second obstacle B2 is received, as shown in FIG. 2 (b), both sensors 1a and 1b first receive the first obstacle. Since the reflected wave from B1 is received, the lateral position of the first obstacle B1, that is, the estimated triangulation position α can be detected by the principle of triangulation.

However, when the Q sensor 1b is a direct detection sensor and the P sensor 1a is an indirect detection sensor, the reciprocating path length between the Q sensor 1b and the second obstacle B2 is the same as that of the Q sensor 1b and the first obstacle B1. When it is smaller than the round-trip path length between, the Q sensor 1b first receives the reflected wave from the second obstacle B2, and then receives the reflected wave from the first obstacle B1. Further, when the path length from the Q sensor 1b to the P sensor 1a via the first obstacle B1 is smaller than the path length from the Q sensor 1b to the P sensor 1a via the second obstacle B2, the P sensor 1a is the first. The reflected wave from the first obstacle B1 is received, and then the reflected wave from the second obstacle B2 is received. At this time, the round-trip path length between the Q sensor 1b and the second obstacle B2 is determined. When it is smaller than the round-trip path length between the Q sensor 1b and the first obstacle B1, the Q sensor 1b first receives the reflected wave from the second obstacle B2 and then the reflection from the first obstacle B1. The wave is received, and as shown in FIG. 2C, the obstacles that the sensors 1a and 1b receive the reflected wave first do not match, and the obstacles that first receive the reflected wave do not match, and the sensors 1a and 1b are located near the intermediate position. The triangular survey estimated position β of the second obstacle B2 is detected, and this estimated position β detects a virtual image, and does not mean that the actual position of the second obstacle B2 is detected.

As described above, in the obstacle detection based only on the principle of the triangulation method, as shown in FIG. 2D, even if the first obstacle B1 can be detected as the estimated triangulation position α, the second obstacle B2 Cannot detect the actual second obstacle B2 by detecting that it is the estimated triangulation position β of the virtual image. Therefore, the time-series estimated position by the time-series estimation position deriving means 4 described above is compared with the triangulation estimation position by the triangulation estimation position deriving means 3 by the comparison means 5, and the distance between the two positions is larger than a predetermined threshold value. Whether it is small or small is determined by the determination means 6.

That is, as shown in FIG. 3A, the distance from the P sensor 1a measured at the current measurement time t to the first obstacle B1 by the control means 2 is the arc S1a (1 in FIG. 3A). A dotted chain line) is formed, and the arc S1b (one-dot chain line in FIG. 3A) is formed at the distance from the Q sensor 1b measured at the current measurement time t to the second obstacle B2 by the control means 2. Is formed, and the distance from the P sensor 1a measured at the measurement time t-1 before the one measurement time Δt (for example, 100 ms) to the first obstacle B1 is the arc S2a (two points in FIG. 3B). The chain line) is formed, and the arc S2b (two-dot chain line in FIG. 3B) is the distance from the Q sensor 1b measured at the measurement time t-1 before one measurement time to the second obstacle B2. ) Is formed.

Then, as shown in FIG. 3B, the point where the arcs S1a and the arcs S2a overlap at the measurement time t and the measurement time t-1 is derived as the time-series estimated position Ω of the first obstacle B1 and is also derived. The point where the arcs S1b and S2b overlap at the measurement time t and the measurement time t-1 is derived as the time series estimation position θ of the second obstacle B2, and is derived by the triangulation estimation position derivation means 3 by the comparison means 5. A comparison is made as to whether the distance (deviation) between the triangulation estimated position α of the first obstacle B1 and the time-series estimated position Ω by the time-series estimated position deriving means 4 is larger or smaller than the preset threshold value Da. Similarly, the distance (deviation) between the triangulation estimation position β of the second obstacle B2 derived by the triangulation estimation position derivation means 3 and the time series estimation position θ by the time series estimation position derivation means 4 is predetermined. A comparison is made as to whether it is larger or smaller than the set threshold Db. Here, the threshold values Da and Db may be the same value or may be different values.

Further, as a result of comparison, when the distance (deviation) between the triangulation estimation position α and the time series estimation position Ω is smaller than the preset threshold value Da, the determination means 6 is used to determine the triangulation estimation position α and the time series estimation position Ω. Are both determined to be the same first obstacle B1. On the other hand, regarding the triangulation estimation position β and the time series estimation position θ, since the triangulation estimation position β is a virtual image and not the actual second obstacle B2, the triangulation estimation position β and the time series estimation position θ The distance (deviation) of is larger than the preset threshold value Db, and the determination means 6 determines that the triangulation estimation position β and the time series estimation position θ are not the same second obstacle B2, and as a result The estimated triangulation position β is excluded as a phantom image, and the time-series estimated position θ is determined to be the second obstacle B2. In this way, both the first and second obstacles B1 and B2 are detected.

The detection processing of the first and second obstacles B1 and B2 by the P and Q sensors 1a and 1b described above will be described with reference to the flowchart of FIG. Now, as shown in FIG. 4, ultrasonic waves are transmitted from the P sensor 1a, the triangulation estimation position α of the first obstacle B1 is derived by the triangulation estimation position derivation means 3 (step S1), and the Q sensor 1b is derived. The ultrasonic wave is transmitted from, and the triangulation estimation position β of the second obstacle B2 is derived by the triangulation estimation position derivation means 3 (step S2), and then the time series estimation is performed based on the ultrasonic wave from the P sensor 1a. The time-series estimated position Ω of the first obstacle B1 is derived by the position deriving means 4 (step S3), and the time-series of the second obstacle B2 is derived by the time-series estimated position deriving means 4 based on the ultrasonic waves from the Q sensor 1b. The estimated position θ is derived (step S3).

Next, the comparison means 5 determines whether or not the distance (deviation) between the triangulation estimation position α and the time series estimation position Ω is smaller than the preset threshold value Da (step S5), and this determination result is obtained. If YES, the estimated triangulation position α and the time series estimated position Ω are determined to be the same first obstacle B1 (step S6), and if the determination result in step S5 is NO, the triangulation exclusion process, that is, triangulation is performed. The estimated position α and the time-series estimated position Ω are not the same first obstacle B1, but the triangulation estimated position α is determined to be a phantom image, and the time-series estimated position Ω is detected as the first obstacle B1 (step). S7).

Further, after the processing of steps S6 and S7, the comparison means 5 determines whether or not the distance (deviation) between the triangulation estimation position β and the time series estimation position θ is smaller than the preset threshold value Db. (Step S8) If this determination result is YES, it is determined that the triangulation estimation position β and the time series estimation position θ are the same second obstacle B2 (step S9), and the determination result in step S8 is NO. For example, the same imaginary image exclusion process as in step S7 described above, that is, the triangulation estimation position β and the time series estimation position θ are not the same second obstacle B2, and the triangulation estimation position β is determined to be a phantom image and the time series estimation position. It is detected that θ is the second obstacle B2 (step S10), and then the operation ends with the process of step S9 described above.

Therefore, according to the above-described embodiment, the triangulation estimation position derived from the triangulation estimation position derivation means 3 based on the distance measurement data of the P sensor 1a and the Q sensor 1b used as the direct detection sensor and the indirect detection sensor, and the triangulation estimation position. The comparison means 5 compares whether the distance from the time-series estimated position derived by the time-series estimated position deriving means 4 for each of the sensors 1a and 1b is larger or smaller than the predetermined threshold value, and as a result of the comparison, a predetermined value is obtained. If it is smaller than the threshold value, the triangulation estimated position and the time-series estimated position are determined to be the same obstacle, and if it is larger, the triangulation estimated position is a virtual image and the time-series estimated position is an obstacle. Since the determination is made, it is possible to reliably detect a plurality of obstacles that could not be detected only by the triangulation method in the past, and it is possible to improve the reliability of obstacle detection.

The present invention is not limited to the above-described embodiment, and various modifications can be made other than those described above as long as the present invention is not deviated from the gist thereof.

For example, in the above-described embodiment, as the direct detection sensor and the indirect detection sensor, one of the two P sensors 1a and the Q sensor 1b is measured as a direct detection sensor and the other is an indirect detection sensor, and then the direct detection sensor and the indirect detection sensor are measured. Although the case where the detection sensors are exchanged for measurement has been described, the number of sensors is not limited to two, and may be adapted by a combination of two of the three or more sensor configurations.

Further, in the above-described embodiment, the case where the sensors 1a and 1b are provided at the front portion of the own vehicle C has been described, but the present invention can be similarly implemented even when the sensors 1a and 1b are provided at the rear portion of the own vehicle C. can.

1a ... P sensor 1b ... Q sensor 3 ... Triangulation estimation position derivation means 4 ... Time series estimation position derivation means 5 ... Comparison means 6 ... Judgment means C ... Own vehicle α, β ... Triangulation estimation position Ω, θ ... Time series Estimated position

Claims (1)

In an obstacle detection device that detects the position of an obstacle by exploring the area around the vehicle with a plurality of direct detection sensors and indirect detection sensors at predetermined measurement time intervals and measuring the distance to the obstacle.
Triangulation estimation position derivation means that estimates the position of the obstacle by the triangulation method based on the distance measurement data by each sensor and derives the triangulation estimation position.
For each of the sensors, an arc centered on the position of the sensor at the current time and a radius of the first measurement distance to the obstacle measured by the sensor at the current time, and the measurement once from the current time. The point where the arc overlaps with the arc centered on the position of the sensor before the time and whose radius is the second measurement distance to the obstacle measured by the sensor one time before the current time is the obstacle. Time-series estimated position derivation means to be derived as the time-series estimated position of
The distance between the triangulation estimation position derived by the triangulation estimation position derivation means and the time series estimation position derived by the time series estimation position derivation means for each sensor is larger than a predetermined threshold value, respectively. A comparison method to compare whether it is small or small,
By the comparison means, if the distance is smaller than the predetermined threshold value, the triangulation estimation position and the time series estimation position are determined to be the same obstacle, and if the distance is larger, the triangulation estimation position is a virtual image and the time series estimation is performed. An obstacle detection device including a means for determining that a position is an obstacle.

Images