JP5943981B2 - Object detection device, parking assist device, and object detection method - Google Patents

Description

  The present invention relates to an object detection device that detects an object by estimating a position of a reflection point when the object reflects a detection wave from a distance sensor, a parking assist device including the object detection device, and an object detection method. It is.

  Conventionally, parking for notifying the driver of the own vehicle who wants to park in the parking space existing between the first parked vehicle and the second parked vehicle whether or not the own vehicle can be parked in this parking space. Support devices have been proposed (see, for example, Patent Document 1).

  The parking assistance apparatus described in Patent Document 1 obtains distance data corresponding to the corner portion of each parked vehicle from the time series of distance data from the vehicle to the parked vehicle measured using a distance sensor mounted on the vehicle. The extracted distance data is subjected to noise processing and data interpolation by curve approximation as data processing.

  Moreover, this parking assistance apparatus estimates the reflection point position of each parked vehicle from the distance data after data processing and the sensor position data indicating the movement locus of the sensor position of the distance sensor. Further, the parking assist device determines whether or not the own vehicle can be parked in the parking space by estimating the length of the parking space based on the position of the corner portion of each parked vehicle obtained from the estimated reflection point position. To do.

Japanese Patent No. 5506803

However, the prior art has the following problems.
In the prior art described in Patent Document 1, when the own vehicle passes the side of the parking space, distance data detected from the first parked vehicle, and distance data detected from the second parked vehicle, The distance data corresponding to the corner portion of each parked vehicle is extracted by temporarily interrupting the vehicle.

  However, in a situation where the distance between the first parked vehicle and the second parked vehicle is not sufficiently large, even if there is actually a parking space where the own vehicle can be parked, it is detected from each parked vehicle. The distance data is not interrupted. Therefore, since the distance data corresponding to the corner portion of each parked vehicle cannot be extracted, the reflection point position of each parked vehicle cannot be estimated, and as a result, whether or not the own vehicle can be parked cannot be determined. There is a problem.

  Moreover, in the prior art described in Patent Document 1, each parked vehicle is calculated from the distance data of each parked vehicle and the sensor position of the distance sensor at the current time, and the distance data of the parked vehicle and the sensor position of the distance sensor at the past time. The position of the reflection point is estimated.

  Here, in estimating the reflection point position of each parked vehicle, the position of the reflection point of the parked vehicle (in other words, the shortest distance point where the distance from the distance sensor to the parked vehicle is the shortest) is the current time and the past time. Is almost the same. However, when the sensor position of the distance sensor moves between the current time and the past time, the position of the reflection point also changes. Therefore, if the vehicle speed of the vehicle on which the distance sensor is mounted is high, this assumption is not valid, and therefore the error in the estimated reflection point position becomes large. There is a problem that it cannot be determined.

  The present invention has been made to solve the above-described problems, and an object detection device capable of accurately estimating the reflection point position of an object that is a detection target of a distance sensor with a simple configuration, It is an object of the present invention to obtain a parking assistance device and an object detection method provided with an object detection device.

The object detection device according to the present invention is mounted on the host vehicle, irradiates the detection target object with a detection wave, and acquires the detection wave reflected at the reflection point position on the object corresponding to the shortest distance to the object. The vehicle is equipped with a distance sensor that detects distance data to the object, and a vehicle information sensor that is mounted on the vehicle and that detects a state related to the speed and traveling direction of the vehicle as the vehicle data. Detecting the parking space information of the vehicle by acquiring the time-series detection results by the distance sensor and the vehicle information sensor and specifying the reflection point position of the object that changes in time series from the detection results It is predicted that the distance sensor will detect at the time tn from the distance track generated by following the time transition of the distance data detected by the distance sensor before the time tn. If the distance data Dn detected by the distance sensor at the time tn is included in the first gate set around the generated distance track prediction value Dn ′ The distance track update value Dn ″ is generated using the data Dn and the distance track predicted value Dn ′ to update the distance track, and if the distance data Dn is not included in the first gate, the distance track predicted value Dn ′ is calculated. The distance track update value Dn "and the new distance track starting from the distance data Dn are started to follow the time transition of the distance data, and the distance data is discriminated for each different object to generate the distance track. An object discriminating unit that detects the position of the host vehicle and the direction of the host vehicle as the host vehicle position information VPn from the host vehicle data Vn detected by the vehicle information sensor at time tn, From the position information VPn, the known mounting position of the distance sensor on the vehicle, and the sensor direction information of the known distance sensor, the position and direction of the distance sensor at time tn are calculated as sensor position information SPn, For each different object, an arc An having an azimuth specified by the sensor position information SPn is calculated on the circumference with the radius specified by the distance track update value Dn ″ centered on the position specified by the sensor position information SPn. The position of the contact point on the arc An that is in contact with the common tangent line of the calculated arc An and one of the arcs calculated at a time before the arc An is estimated as the reflection point position Rn of the object at the time tn. The reflection point positioning unit, the reflection point position Rn estimated for each different object by the reflection point positioning unit, and each of the reflection point position estimation units estimated at a time before the reflection point position Rn And a launch pad position, by specifying the position of each object, instead and a object detecting unit that estimates a parking space information, reflection point positioning unit for estimating a reflection point position Rn using common tangent In addition, the position of the contact point on the arc An that is in contact with the common tangent line between the arc An and each arc calculated at a plurality of times before the arc An is a reflection point temporary position RTn and is included in the sensor position information SPn. A point on the arc An located in the azimuth corresponding to the average value of the respective azimuths of the reflection point temporary position RTn as viewed from the position of the distance sensor or the azimuth corresponding to the median value of each azimuth is defined as the reflection point position Rn. To be estimated .

  In addition, a parking assist device according to the present invention includes the object detection device described above and a vehicle control device that performs parking support for parking the host vehicle in the parking space from the parking space information estimated by the object detection device. It is a thing.

Furthermore, the object detection method according to the present invention is mounted on the own vehicle, irradiates a detection wave on an object to be detected, and detects a detection wave reflected at a reflection point position on the object corresponding to the shortest distance to the object. A vehicle having a distance sensor that detects distance data to the object by acquiring the vehicle and a vehicle information sensor that is mounted on the vehicle and detects a state related to the speed and traveling direction of the vehicle as the vehicle data. The object that estimates the parking space information of the vehicle by acquiring the time-series detection results by the distance sensor and the vehicle information sensor while moving the vehicle, and identifying the reflection point position of the object that changes in time series from the detection results This is a detection method, and the distance detected by the distance sensor at the time tn from the distance track generated by following the time transition of the distance data detected by the distance sensor before the time tn. Based on the data Dn, the step of generating the distance track update value Dn ″ for each different object, and the own vehicle position and the direction of the own vehicle from the own vehicle data Vn detected by the vehicle information sensor at time tn From the step of estimating the position information VPn, the own vehicle position information VPn, the mounting position of the known distance sensor on the own vehicle, and the sensor direction information of the known distance sensor, the position of the distance sensor at time tn and The azimuth is calculated as the sensor position information SPn, and for each different object, the sensor position information SPn out of the circumference with the radius as the distance track update value Dn ″ with the position specified by the sensor position information SPn as the center. An arc An having a specified orientation is calculated, and the calculated arc An is tangent to a common tangent line between the calculated arc An and one of the arcs calculated at a time before the arc An. The position of the contact point on the arc An, a reflection point positioning step of estimating a reflection point position Rn of the object at time tn, the reflecting point position Rn, each reflection point position estimated in the previous time than the reflection point position Rn And estimating the parking space information by specifying the position of each object, and in the reflection point positioning step, instead of estimating the reflection point position Rn using the common tangent, the arc An and The position of the contact point on the arc An that is in contact with the common tangent to each arc calculated at a plurality of times before the arc An is a reflection point temporary position RTn, and the position of the distance sensor included in the sensor position information SPn. An azimuth corresponding to the average value of the respective azimuths of the reflected reflection point temporary position RTn or a point on the arc An located in the azimuth corresponding to the median value of each azimuth is estimated as the reflection point position Rn. To do .

  According to the present invention, for each different object, the position specified by the sensor position information SPn out of the circumference with the radius as the distance track update value Dn ″ centered on the position specified by the sensor position information SPn at time tn. The position of the contact point on the arc An that is in contact with the common tangent of the calculated arc An and one of the arcs calculated at a time before the arc An is calculated at time tn. An object detection device capable of accurately estimating the reflection point position of an object, which is a detection target of a distance sensor, with a simple configuration, is provided. A parking assist device and an object detection method provided can be obtained.

It is a block diagram which shows the structure of the parking assistance system containing the object detection apparatus in Embodiment 1 of this invention. It is a flowchart which shows a series of operation | movement of the parking assistance system in Embodiment 1 of this invention. It is a flowchart which shows a series of operation | movement of the object discrimination process by the object discrimination part in Embodiment 1 of this invention. It is a flowchart which shows a series of operation | movement of the object discrimination process by the object discrimination part in Embodiment 1 of this invention. It is explanatory drawing which shows an example of the distance track | truck for every object obtained when the object discrimination part in Embodiment 1 of this invention performs an object discrimination process at each time. It is explanatory drawing for demonstrating the effect acquired when the object discrimination part in Embodiment 1 of this invention performs an object discrimination process. It is a flowchart which shows a series of operation | movement of the reflection point positioning process by the reflection point positioning part in Embodiment 1 of this invention. It is a flowchart which shows a series of operation | movement of the reflection point positioning performed by step S305 of FIG. 6A. It is explanatory drawing for demonstrating the calculation method of the reflective point position of the object by the reflective point positioning part in Embodiment 1 of this invention. It is a flowchart which shows a series of operation | movement of the parking space information estimation process by the object detection part in Embodiment 1 of this invention. It is a flowchart which shows a series of operation | movement of the reflective point tracking performed by step S505 of FIG. 8A. It is a flowchart which shows a series of operation | movement of the reflection point wake process before establishment performed by step S610 of FIG. 8B. It is a flowchart which shows a series of operation | movement of the reflection point wake process after establishment performed by step S611 of FIG. 8B. It is explanatory drawing for demonstrating the estimation method of the reflective point position of the object by the object detection part in Embodiment 1 of this invention. It is explanatory drawing for demonstrating another example of the estimation method of the reflective point position of the object by the object detection part in Embodiment 1 of this invention. It is explanatory drawing for demonstrating the estimation method of the parking space information by the object detection part in Embodiment 1 of this invention. It is a flowchart which shows a series of operation | movement of the reflection point positioning process by the reflection point positioning part in Embodiment 2 of invention. It is a flowchart which shows a series of operation | movement of the reflection point positioning performed by step S905 of FIG. 11A. It is explanatory drawing for demonstrating the calculation method of the reflective point position of the object by the reflective point positioning part in Embodiment 2 of this invention. It is a flowchart which shows a series of operation | movement of the reflection point positioning process by the reflection point positioning part in Embodiment 3 of this invention. It is a flowchart which shows a series of operations of reflection point positioning performed by step S1105 of FIG. 13A. It is explanatory drawing for demonstrating the calculation method of the reflective point position of the object by the reflective point positioning part in Embodiment 3 of this invention.

  Hereinafter, an object detection apparatus, a parking assistance apparatus including the object detection apparatus, and an object detection method according to the present invention will be described with reference to the drawings according to a preferred embodiment. In the description of the drawings, the same reference numerals are assigned to the same elements, and duplicate descriptions are omitted. Moreover, in the following embodiment, the case where the object detection apparatus and object detection method of this invention are applied to a parking assistance apparatus is illustrated.

Embodiment 1 FIG.
FIG. 1 is a block diagram showing a configuration of a parking assistance system including an object detection device 310 according to Embodiment 1 of the present invention. The parking support system in FIG. 1 includes a distance sensor 100, a vehicle information sensor 200, and a parking support device 300.

  The distance sensor 100 is mounted on the vehicle, irradiates the detection target object with a detection wave, and acquires the detection wave reflected at the reflection point position on the object corresponding to the shortest distance to the object. Distance data up to is detected. Specifically, the distance sensor 100 emits light to an object to be detected, receives light reflected from the object, performs signal processing or image processing based on the received light, Is detected as distance data. In addition, the distance sensor 100 detects distance data for each preset setting timing.

  The distance sensor 100 outputs the detected distance data to the object detection device 310. The distance data detected by the distance sensor 100 is stored in a storage unit (not shown) in association with the time when the distance data is detected.

  Here, it is assumed that the mounting position of the distance sensor 100 on the vehicle and the sensor orientation information of the distance sensor 100 are already known. This sensor azimuth information includes the sensor azimuth (that is, the line-of-sight direction) of the distance sensor 100 and the sensor viewing angle (that is, a detectable azimuth width). As the distance sensor 100, a sensor of a type that detects a distance to an object as distance data using a detection wave such as an electromagnetic wave instead of light may be used. Furthermore, as the distance sensor 100, specifically, a millimeter wave radar, a laser radar, an ultrasonic sensor, an infrared sensor, an optical camera, or the like can be used.

  The vehicle information sensor 200 is mounted on a vehicle and detects a state relating to the speed and the traveling direction of the own vehicle as own vehicle data. Specifically, the vehicle information sensor 200 detects a state relating to the vehicle speed and the traveling direction such as the vehicle speed, wheel speed, steering angle, and yaw rate as the vehicle data. Moreover, the vehicle information sensor 200 detects own vehicle data for every set timing mentioned above.

  The vehicle information sensor 200 outputs the detected own vehicle data to the object detection device 310. In addition, the host vehicle data detected by the vehicle information sensor 200 is stored in the storage unit in association with the time when the host vehicle data is detected.

  In addition, you may comprise the vehicle information sensor 200 so that the latitude, the longitude, and the advancing direction of the own vehicle may be detected as own vehicle data using GPS (Global Positioning System).

  The parking assist device 300 includes an object detection device 310 and a vehicle control device 320. The object detection device 310 detects an object (for example, a parked vehicle) located next to a parking space where the host vehicle is to park. That is, the time-series detection results by the distance sensor 100 and the vehicle information sensor 200 are acquired, and the reflection space position of the object that changes in time series is identified from the detection results, thereby estimating the parking space information of the own vehicle. In addition, as a specific example of parking space information, the position of the side surface or corner of an object located next to a parking space, etc. are mentioned, for example.

  The object detection device 310 includes an object discriminating unit 311, a vehicle position estimating unit 312, a reflection point positioning unit 313, and an object detection unit 314.

  The object discriminating unit 311 uses the distance data input from the distance sensor 100 to generate a distance track for each object. In addition, the object discriminating unit 311 compares the distance track for each object held with the distance data input from the distance sensor 100 to determine which object the distance data is detected from. Identify. Further, the object discriminating unit 311 outputs the distance track for each held object to the reflection point positioning unit 313.

  The own vehicle position estimation unit 312 estimates the position of the own vehicle and the direction of the own vehicle (that is, the traveling direction) from the own vehicle data input from the vehicle information sensor 200 as the own vehicle position information. Is output to the reflection point positioning unit 313.

  The reflection point positioning unit 313 generates reflection point position information for each object based on the distance track for each object input from the object discrimination unit 311 and the own vehicle position information input from the own vehicle position estimation unit 312. And output to the object detection unit 314.

  The object detection unit 314 generates a reflection point track for each object based on the reflection point position information for each object input from the reflection point positioning unit 313. Further, the object detection unit 314 estimates parking space information based on the reflection point track for each object, and outputs the parking space information to the vehicle control device 320.

  Based on the parking space information input from the object detection device 310, the vehicle control device 320 provides parking assistance for the own vehicle to park in the parking space. Specifically, the vehicle control device 320 estimates the width of the parking space and compares the width of the parking space with the width of the own vehicle to determine whether the vehicle can be parked in the parking space. . In addition, the vehicle control device 320 informs the driver who drives the vehicle of the determination result.

  If it is determined that the host vehicle can be parked in the parking space, the vehicle control device 320 calculates a route for the host vehicle to move to the parking space based on the positional relationship between the parking space and the host vehicle. In addition, the vehicle control device 320 guides the host vehicle to the parking space according to the calculated route.

  Next, the operation of the parking support system according to the first embodiment will be described with reference to FIG. FIG. 2 is a flowchart showing a series of operations of the parking assistance system according to Embodiment 1 of the present invention. Hereinafter, the time at which the distance sensor 100 detects the distance data and the time at which the vehicle information sensor 200 detects the own vehicle data are time t1, time t2,..., Time tn-1, time tn. , Time tn + 1,... (N is an integer equal to or greater than 1) and detected at the same time.

  In step S101, the distance sensor 100 detects the distance data Dn at time tn, and outputs the distance data Dn at time tn to the object discriminating unit 311.

  In step S102, the object discriminating unit 311 performs object discrimination processing on the distance data Dn input from the distance sensor 100, thereby identifying from which object the distance data Dn is detected. . The object discriminating unit 311 outputs the distance track for each object at the time tn to the reflection point positioning unit 313 by reflecting the distance data Dn.

  In step S103, the vehicle information sensor 200 detects the host vehicle data Vn at time tn and outputs the host vehicle data Vn at time tn to the host vehicle position estimation unit 312.

  In step S <b> 104, the own vehicle position estimation unit 312 estimates the own vehicle position information VPn at time tn from the own vehicle data Vn input from the vehicle information sensor 200, and outputs it to the reflection point positioning unit 313. The reflection point positioning unit 313 also obtains reflection point position information for each object from the distance track for each object input from the object discrimination unit 311 and the vehicle position information VPn input from the vehicle position estimation unit 312. And output to the object detection unit 314.

  In step S105, the object detection unit 314 updates the reflection point track for each object based on the reflection point position information for each object input from the reflection point positioning unit 313. Further, the object detection unit 314 estimates parking space information based on the reflection point track for each object, and outputs the parking space information to the vehicle control device 320. The parking space may not be estimated depending on the number of reflection point position data included in the reflection point track.

  In step S <b> 106, the vehicle control device 320 performs parking support for parking the own vehicle in the parking space based on the parking space information input from the object detection unit 314, and ends a series of processes. If the parking space is not estimated in step S105, the vehicle control device 320 ends the series of processes without performing parking assistance.

  Next, the object discrimination processing by the object discrimination unit 311 executed in step S102 of FIG. 2 will be described with reference to FIGS. 3A and 3B. 3A and 3B are flowcharts showing a series of operations of object discrimination processing by the object discriminating unit 311 according to Embodiment 1 of the present invention. 3A and 3B are each a flowchart divided into two drawings, and the part shown in FIG. 3A and the part shown in FIG. 3B are continued.

  Here, a time series of a series of distance data detected by the distance sensor 100 is referred to as a distance wake. Each time the object discriminating unit 311 newly acquires the distance data detected by the distance sensor 100, the object discriminating section 311 takes in the distance data into the distance track of the detected object and updates the distance track. In updating the distance track, for example, an update value can be generated using a least square method or a Kalman filter, and the update value thus generated is called a distance track update value.

  First, in step S201, the object discriminating unit 311 generates distance data predicted to be detected by the distance sensor 100 at time tn as a distance wake prediction value Dn ′ for each of the currently held distance wakes. Proceed to step S202.

  In step S202, the object discriminating unit 311 determines whether or not the distance data Dn at the time tn has been acquired from the distance sensor 100. If the object discriminating unit 311 determines that the distance data Dn has been acquired (that is, YES), the process proceeds to step S204, and if it is determined that the distance data Dn has not been acquired (that is, NO), Proceed to step S203.

  When the process proceeds to step S203, the object discriminating unit 311 sets the distance track predicted value Dn ′ generated in step S201 for each distance track currently held as the distance track update value Dn ″ and extrapolation. The number of times is incremented, and the process proceeds to step S215.

  On the other hand, when the process proceeds to step S204, the object discriminating unit 311 performs a gate inside / outside determination for associating the distance data Dn with the currently held distance track, and the process proceeds to step S205.

  Here, the gate is a fixed distance range centered on the distance track prediction value Dn ′. As for the gate width for determining the distance range, for example, a constant multiple of the distance error standard deviation of the distance sensor 100 may be determined as the gate width. When the Kalman filter is used for updating the distance track, a constant multiple of the distance residual variance determined by the sum of the distance component of the prediction error covariance matrix and the distance observation error variance may be determined as the gate width.

  In step S204, the object discriminating unit 311 sets a gate for each of the currently held distance tracks using the distance track predicted value Dn ′ generated in step S201. Further, when the distance data Dn enters the set gate, the object discriminating unit 311 determines that there is a possibility that the distance track in which such a gate is set and the distance data Dn are related. On the other hand, when the distance data Dn at the time tn does not enter the set gate, the object discriminating unit 311 determines that there is no possibility that the distance track in which such a gate is set and the distance data Dn are related.

  In step S205, the object discriminating unit 311 performs distance linking that associates with the distance data Dn from the distance wakes determined to be related to the distance data Dn in step S204 among the currently held distance wakes. Is uniquely determined, and the process proceeds to step S206. In step S205, the object discriminating unit 311 does not determine the distance track to be associated with the distance data Dn when there is no distance track determined to be related to the distance data Dn in step S204.

  Here, the distance track associated with the distance data Dn is referred to as an updated distance track. Further, in uniquely determining the update distance track, for example, the distance track with the largest number of updates may be set as the update distance track. It should be noted that the distance track that can take the smallest absolute value of the difference between the distance track predicted value Dn ′ and the distance data Dn may be used as the updated distance track.

  In step S206, the object discriminating unit 311 determines whether or not the updated distance track is determined in step S205. If the object discriminating unit 311 determines that the update distance track has been determined (that is, YES), the process proceeds to step S209, and if it is determined that the update distance track has not been determined (that is, NO), Proceed to step S207.

  When the process proceeds to step S207, the object discriminating unit 311 sets the distance track predicted value Dn ′ generated in step S201 for each distance track currently held as the distance track update value Dn ″ and extrapolation. The number of times is incremented, and the process proceeds to step S208.

  In step S208, the object discriminating unit 311 generates a new distance wake starting from the distance data Dn. Further, the object discriminating unit 311 sets the update count to “1”, the extrapolation count to “0”, and the state to “unestablished” for the generated new distance wake, and further holds this new distance wake. To step S215.

  On the other hand, when the process proceeds to step S209, the object discriminating unit 311 selects one unselected distance track that has not yet been selected from the currently stored distance tracks, and selects the selected distance track. The selected mark is attached to the wake as selected, and the process proceeds to step S210.

  In step S210, the object discriminating unit 311 determines whether or not the distance track selected in step S209 is the updated distance track determined in step S205. When the object discriminating unit 311 determines that this distance track is the updated distance track (that is, YES), the object discrimination unit 311 proceeds to step S211 and determines that this distance track is not the updated distance track (that is, NO). Then, the process proceeds to step S212.

  When the process proceeds to step S211, the object discriminating unit 311 generates and updates the distance track update value Dn ″ using the distance data Dn and the distance track predicted value Dn ′ for the distance track selected in step S209. The number of times is incremented, the number of extrapolations is reset to “0”, and the process proceeds to step S213.

  On the other hand, when the process proceeds to step S212, the object discriminating unit 311 sets the distance track predicted value Dn ′ generated in step S201 for the distance track selected in step S209 as the distance track update value Dn ″ and extrapolated. The number of times is incremented, and the process proceeds to step S213.

  In step S213, the object discriminating unit 311 determines whether there is an unselected distance track that has not yet been marked in step S209 among the distance tracks currently held. If the object discriminating unit 311 determines that this distance track exists (that is, YES), the object discrimination unit 311 returns to step S209, executes the processing after step S209 again, and does not have this distance track (that is, NO). ), The process proceeds to step S214.

  In step S214, the object discriminating unit 311 cancels the selected mark attached to each of the distance tracks currently held, and proceeds to step S215.

  In step S215, the object discriminating unit 311 selects one unselected distance track that has not yet been selected from the distance tracks that are currently held, and has already been selected as the selected distance track. A mark is attached, and the process proceeds to step S216.

  In step S216, the object discriminating unit 311 determines whether or not the state of the distance track selected in step S215 is “established”. When the object discriminating unit 311 determines that the state of the distance wake is “established” (ie, YES), the process proceeds to step S217, and the state of the distance wake is “not established” (ie, If NO is determined, the process proceeds to step S219. In the following, a distance wake having a state of “established” is referred to as an established distance wake, and a distance wake having a state of “not established” is referred to as an unestablished distance wake.

  When the process proceeds to step S217, the object discriminating unit 311 determines whether or not the number of extrapolation of the distance wake selected in step S215 is a constant M or more. If the object discriminating unit 311 determines that the extrapolation count of the distance wake is equal to or greater than the constant M (that is, YES), the process proceeds to step S218, and the extrapolation count of the distance wake is less than the constant M. If it is determined (that is, NO), the process proceeds to step S223. The constant M is a positive integer determined as the upper limit value of the extrapolation count of the established distance wake, and can be designed as appropriate.

  In step S218, the object discriminating unit 311 determines that distance data is not detected after the time tn + 1 by the distance sensor 100 from the object corresponding to the distance track selected in step S215, and the track processing for the distance track of this object. Exit. For the distance track for which the wake process has been completed, the time when the extrapolation count was 0 last from the time tn and the distance track update value at that time are associated with each other and stored in the storage unit. In addition, the object discriminating unit 311 removes the distance track for which the track processing has been completed from the holding target, and proceeds to step S223.

  On the other hand, when the process proceeds to step S219, the object discriminating unit 311 determines whether or not the number of extrapolation of the distance wake selected in step S215 is equal to or greater than a constant N. If the object discriminating unit 311 determines that the extrapolation count of the distance wake is equal to or greater than a constant N (that is, YES), the process proceeds to step S220, and the extrapolation count of the distance wake is less than the constant N. If it is determined (that is, NO), the process proceeds to step S221. The constant N is a positive integer determined as the upper limit value of the extrapolation count of the unestablished distance track, and can be designed as appropriate.

  When the process proceeds to step S220, the object discriminating unit 311 determines that the distance track selected in step S215 does not become the established distance track even after time tn + 1, deletes this distance track, and removes the distance track from the holding target. The process proceeds to step S223.

  On the other hand, when the process proceeds to step S221, the object discriminating unit 311 determines whether or not the number of updates of the distance track selected in step S215 is equal to or greater than a constant K. If the object discriminating unit 311 determines that the number of updates of the distance track is equal to or greater than the constant K (that is, YES), the process proceeds to step S222, and the number of updates of the distance track is less than the constant K (that is, , NO), the process proceeds to step S223. The constant K is a positive integer determined as the number of updates required for the distance wake selected in step S215 to change from the unestablished distance wake to the established distance wake, and can be designed as appropriate. And

  In step S222, the object discriminating unit 311 changes the state of the distance track selected in step S215 from “not established” to “established”, and proceeds to step S223. By such a change, the distance track selected in step S215 is changed from the unestablished distance track to the established distance track.

  In step S223, the object discriminating unit 311 determines whether there is an unselected distance track that is not yet marked in step S215 among the distance tracks currently held. If the object discriminating unit 311 determines that the distance wake exists (that is, YES), the process returns to step S215, and the processing after step S215 is executed again, and the distance wake does not exist (that is, NO). ), The process proceeds to step S224.

  In step S224, the object discriminating unit 311 cancels the selected mark attached to each of the currently held distance tracks, ends the series of processes, and reflects the distance tracks for each object at time tn as reflection points. The data is output to the positioning unit 313.

  Next, an example of the distance track for each object obtained when the object discriminating unit 311 executes the object discrimination process at each of the times t1 to t7 will be described with reference to FIG. FIG. 4 is an explanatory diagram illustrating an example of a distance track for each object obtained when the object discriminating unit 311 according to Embodiment 1 of the present invention executes the object discrimination processing at each time t1 to t7. FIG. 4 illustrates the case where constant M = 3, constant N = 1, and constant K = 3.

  As shown in FIG. 4, at time t1, the object discriminating unit 311 acquires distance data D1 from the distance sensor 100 for the first time, and therefore generates a new distance wake starting from the acquired distance data D1 as the distance wake A. . In this case, the update count of the distance track A is “1”, the extrapolation count is “0”, and the state is “unestablished”. Further, this distance track A is added to the holding target. The object discriminating unit 311 outputs the distance track A at the time t1 to the reflection point positioning unit 313 as the distance track for each object at the time t1.

  At time t2, the object discriminating unit 311 generates a distance track prediction value D2 'at time t2 for the distance track A currently held (that is, the new distance track generated at time t1). Further, since the distance data D2 enters the gate set with the distance track predicted value D2 ′ as the center, the distance track update value D2 ″ is generated using the distance data D2 and the distance track predicted value D2 ′, and the distance The wake A is updated, in this case, the number of updates of the distance wake A is “2”, the number of extrapolation is “0”, and the state is “unestablished”. The object discriminating unit 311 outputs the distance track A at the time t2 to the reflection point positioning unit 313 as the distance track for each object at the time t2.

  Specifically, at time t2, in the flowchart of FIGS. 3A and 3B, the processing is executed in the order of steps S201 → S202 → S204 to S206 → S209 to S211 → S213 to S216 → S219 → S221 → S223 → S224. It will be.

  At time t3, the object discriminating unit 311 generates a distance track prediction value D3 'at time t3 for the distance track A currently held. Further, since the distance data D3 enters the gate set with the distance track predicted value D3 ′ as the center, the distance track update value D3 ″ is generated using the distance data D3 and the distance track predicted value D3 ′, and the distance The wake A is updated, in this case, the number of updates of the distance wake A is “3”, the number of extrapolation is “0”, and the number of updates of the distance wake A is equal to the constant K. Established. The object discriminating unit 311 outputs the distance track A at the time t3 to the reflection point positioning unit 313 as the distance track for each object at the time t3.

  Specifically, at time t3, in the flowchart of FIGS. 3A and 3B, the processing is executed in the order of steps S201 → S202 → S204 to S206 → S209 to S211 → S213 to S216 → S219 → S221 to S224. Become.

  At time t4, the object discriminating unit 311 generates a distance track prediction value D4 'at time t4 for the distance track A currently held. Further, since the distance data D4 enters the gate set with the distance track predicted value D4 ′ as the center, the distance track update value D4 ″ is generated using the distance data D4 and the distance track predicted value D4 ′, and the distance The wake A is updated, in this case, the number of updates of the distance wake A is “4”, the number of extrapolation is “0”, and the state is “established”. The object discriminating unit 311 outputs the distance track A at the time t4 to the reflection point positioning unit 313 as the distance track for each object at the time t4.

  Specifically, at time t4, in the flowchart of FIGS. 3A and 3B, the processing is executed in the order of steps S201 → S202 → S204 to S206 → S209 to S211 → S213 to S217 → S223 → S224.

  At time t5, the object discriminating unit 311 generates a distance track prediction value D5 'at time t5 for the distance track A currently held. Further, since the distance data D5 does not enter the gate set with the distance track predicted value D5 ′ as the center, the distance track predicted value D5 ′ is set as the distance track updated value D5 ″. The distance data D5 is the starting point. Is generated as the distance wake B. Thus, if the distance wake B is generated, the distance data D5 is the distance data detected from an object different from the object corresponding to the distance wake A. It turns out that it is.

  In this case, the update count of the distance track A is “4”, the extrapolation count is “1”, and the state is “established”. In addition, the update count of the distance track B is “1”, the extrapolation count is “0”, and the state is “unestablished”. Further, this distance track B is added to the holding target. The object discriminating unit 311 outputs the distance track A and the distance track B at the time t5 to the reflection point positioning unit 313 as the distance track for each object at the time t5.

  Specifically, at time t5, in the flowchart of FIGS. 3A and 3B, the processing is executed in the order of steps S201 → S202 → S204 to S208 → S215 to S217 → S223 → S215 → S216 → S219 → S221 → S223 → S224. Will be.

  At time t6, the object discriminating unit 311 performs the distance track predicted value D6 ′ at time t6 corresponding to the distance track A and the time corresponding to the distance track B for each of the distance track A and the distance track B currently held. A distance track predicted value D6 ′ at t6 is generated.

  For distance track A, distance data D6 does not enter the gate set with the corresponding distance track predicted value D6 'as the center, and therefore distance track predicted value D6' is set as distance track updated value D6 ".

  On the other hand, for the distance track B, the distance data D6 enters the gate set with the corresponding distance track predicted value D6 ′ as the center. Therefore, the distance track updated value is calculated using the distance data D6 and the distance track predicted value D6 ′. D6 ″ is generated and the distance track B is updated.

  In this case, the update count of the distance track A is “4”, the extrapolation count is “2”, and the state is “established”. Further, the update count of the distance track B is “2”, the extrapolation count is “0”, and the state is “unestablished”. The object discriminating unit 311 outputs the distance track A and the distance track B at the time t6 to the reflection point positioning unit 313 as the distance track for each object at the time t6.

  Specifically, at time t6, in the flowcharts of FIGS. 3A and 3B, steps S201 → S202 → S204 to S206 → S209 to S211 → S213 → S209 → S210 → S212 to S217 → S223 → S215 → S216 → S219 Processing is executed in the order of S221 → S223 → S224.

  At time t7, the object discriminating unit 311 has the distance track predicted value D7 ′ corresponding to the distance track A at time t7 and the time corresponding to the distance track B for each of the distance track A and the distance track B currently held. A distance track prediction value D7 ′ for t7 is generated.

  For distance track A, distance data D7 does not enter the gate set around the corresponding distance track predicted value D7 ′, and therefore distance track predicted value D7 ′ is set as distance track updated value D7 ″. For the distance wake A, the wake processing ends.

  On the other hand, for the distance track B, the distance data D7 enters the gate set with the corresponding distance track predicted value D7 ′ as the center. Therefore, the distance track updated value using the distance data D7 and the distance track predicted value D7 ′. D7 ″ is generated and the distance track B is updated.

  In this case, the update count of the distance track A is “4”, the extrapolation count is “3”, and the state is “established”. This distance track A is excluded from the holding target because it is determined that the track processing has been completed by the processing of steps S216 to S218. Then, the last time t4 when the extrapolation count of the distance track A is “0” and the distance track update value D4 ″ at the time t4 are associated with each other and stored in the storage unit. The update count is “3”, the extrapolation count is “0”, and the state is “established”. The object discriminating unit 311 outputs the distance track A and the distance track B at the time t7 to the reflection point positioning unit 313 as the distance track for each object at the time t7.

  Specifically, at time t7, in the flowcharts of FIGS. 3A and 3B, steps S201 → S202 → S204 to S206 → S209 to S211 → S213 → S209 → S210 → S212 to S218 → S223 → S215 → S216 → S219. → Processing is executed in the order of S221 to S224.

  Next, the effect obtained by the object discrimination unit 311 executing the object discrimination process will be described with reference to FIG. FIG. 5 is an explanatory diagram for explaining an effect obtained when the object discriminating unit 311 according to Embodiment 1 of the present invention executes the object discrimination process. More specifically, FIG. 5A shows a state in which the distance sensor 100 moving in the direction of the arrow together with the own vehicle detects distance data from the parked vehicle A and the parked vehicle B, and FIG. A change in distance data detected at each time by the sensor 100 is shown.

  As shown in FIG. 5A, the distance data detected by the distance sensor 100 changes as shown in FIG. 5B as the own vehicle moves in front of the parked vehicle A and the parked vehicle B. . That is, when the distance sensor 100 continuously detects the distance data, the distance data corresponding to the parked vehicle A and the distance data corresponding to the parked vehicle B may actually be connected. In such a case, in the prior art, in the portion where the distance data is connected, it is not known whether the distance data of this portion is detected from the parked vehicle A or the parked vehicle B. As a result, the parked vehicle The distance data corresponding to A and the distance data corresponding to parked vehicle B could not be distinguished.

  On the other hand, in the present invention, the object discriminating unit 311 executes the object discriminating process to obtain the distance wake as illustrated in FIG. 4, so that the distance data corresponding to the parked vehicle A and the parking data The distance data corresponding to the vehicle B can be discriminated.

  Here, in FIG. 5A, for example, a case where the time when the distance data corresponding to the parked vehicle B is detected for the first time is tn is considered. In such a case, the distance track predicted value Dn ′ at time tn is generated using the distance track generated from the distance data corresponding to the parked vehicle A detected at the time before time tn. Further, the distance data corresponding to the parked vehicle B detected at time tn does not enter the gate set around the distance track predicted value Dn ′.

  Therefore, a new wake having a starting point of distance data corresponding to the parked vehicle B detected at time tn is generated as a distance wake corresponding to the parked vehicle B. Further, for the distance track corresponding to the parked vehicle A, the distance track predicted value Dn ′ is used as the distance track update value Dn ″. Further, the object discriminating unit 311 uses the time track as the distance track for each object at the time tn. The distance track corresponding to the parked vehicle A and the distance track corresponding to the parked vehicle B at time tn are output to the reflection point positioning unit 313.

  From the above, the object discrimination unit 311 can generate a distance wake corresponding to the parked vehicle A and a distance wake corresponding to the parked vehicle B by executing the object discrimination process. The distance data of the parked vehicle A and the distance data of the parked vehicle B can be discriminated. Further, by generating a distance track predicted value and setting a gate around the distance track predicted value, a noise component of the distance data detected by the distance sensor 100 can be removed.

  Next, the reflection point positioning process performed by the reflection point positioning unit 313 executed in step S104 of FIG. 2 will be described with reference to FIGS. 6A and 6B. FIG. 6A is a flowchart showing a series of operations of the reflection point positioning process by the reflection point positioning unit 313 according to Embodiment 1 of the present invention. FIG. 6B is a flowchart showing a series of operations of reflection point positioning executed in step S305 of FIG. 6A.

  First, in step S <b> 301, the vehicle position estimation unit 312 estimates the vehicle position information VPn based on the vehicle data Vn at the time tn input from the vehicle information sensor 200, and outputs the vehicle position information VPn to the reflection point positioning unit 313. . The reflection point positioning unit 313 acquires the vehicle position information VPn from the vehicle position estimation unit 312 and proceeds to step S302.

  Here, the coordinate system of the position may be arbitrarily set. For example, an orthogonal coordinate system in which the center of the rear wheel axle of the vehicle at the time when the parking assist device 300 is activated is the origin and the coordinate axis is the front and the left of the vehicle at the same time is set as the coordinate system of the position. Can do. Hereinafter, the coordinate system of the set position is referred to as an absolute coordinate system.

  In step S302, the reflection point positioning unit 313 determines from the own vehicle position information VPn acquired in step S301, the known mounting position of the distance sensor 100 on the own vehicle, and the known sensor orientation information of the distance sensor 100. The position and orientation of the distance sensor 100 at time tn in the absolute coordinate system are calculated as sensor position information SPn. The reflection point positioning unit 313 obtains the sensor position information SPn at time tn by performing the calculation in this way, and proceeds to step S303.

  In step S303, the reflection point positioning unit 313 determines whether there is an established distance track among the distance tracks for each object at time tn input from the object discrimination unit 311. If the reflection point positioning unit 313 determines that the established distance track exists (that is, YES), the reflection point positioning unit 313 proceeds to step S304, and determines that the established distance track does not exist (that is, NO), Terminate the process.

  In step S <b> 304, the reflection point positioning unit 313 selects a distance track that is an established distance track and has not yet been selected from among the distance tracks for each object at time tn input from the object discriminating unit 311. One is selected, a selected mark is added as selected to the selected established distance track, and the process proceeds to step S305.

  In step S305, the reflection point positioning unit 313 estimates the reflection point position Rn of the object corresponding to the established distance track using the established distance track selected in step S304. The reflection point positioning executed in step S305 will be described later.

  In step S306, the reflection point positioning unit 313 includes a distance track that is an established distance track and has not yet been assigned a selected mark among the distance tracks for each object at time tn input from the object discriminating unit 311. Determine if it exists. When the reflection point positioning unit 313 determines that there is an established distance track without a selected mark (that is, YES), the reflection point positioning unit 313 returns to step S304, executes the processing after step S304 again, and selects it. If it is determined that there is no established distance track without a completed mark (that is, NO), the process proceeds to step S307.

  In step S307, the reflection point positioning unit 313 releases the selected mark attached to each of the established distance tracks, and ends the series of processes.

  Next, the reflection point positioning operation executed in step S305 will be described with reference to FIG. 6B.

  First, in step S401, the reflection point positioning unit 313 acquires the distance track update value Dn ″ at time tn included in the established distance track selected in step S304, and proceeds to step S402.

  In step S402, the reflection point positioning unit 313 substitutes 1 for the variable k, and proceeds to step S403. This variable k is used to represent the number of samples that go back for reflection point positioning.

  In step S403, the reflection point positioning unit 313 determines whether the detection time difference between the time tn and the time tn−k is equal to or greater than T, or whether the variable k is greater than the constant kmax. For example, when k = 1 at the time of execution of step S403, the reflection point positioning unit determines whether the time difference between the time tn and the time tn−1 is equal to or greater than T, or 1 is greater than a constant kmax. Will be.

  When the reflection point positioning unit 313 determines that the detection time difference is equal to or greater than the constant T, or the variable k is larger than the constant kmax (that is, YES), the process proceeds to step S404 and the detection time difference is less than the constant T. If it is determined that the variable k is equal to or less than the constant kmax (that is, NO), the process proceeds to step S405.

  The constant T is a positive number determined as an upper limit value of the time that goes back from the time tn for reflection point positioning, and can be designed as appropriate. The constant kmax is a positive integer determined as the upper limit value of the number of samples that go back for reflection point positioning, and can be designed as appropriate. For example, the number of distance data samples detected from the time corresponding to the starting point of the established distance track selected in step S304 to time tn can be determined as a constant kmax.

  When the process proceeds to step S404, the reflection point positioning unit 313 assumes that the reflection point position Rn at time tn cannot be uniquely determined for the object corresponding to the established distance wake selected in step S304. It is determined that it is indefinite, and a series of processing ends. In step S404, the reflection point positioning unit 313 sets the state of the reflection point position information RPn to “positioning indefinite” for the object corresponding to the established distance track selected in step S304. Further, the reflection point positioning unit 313 includes information that associates the time tn, the sensor position information SPn in the absolute coordinate system, and the established distance track selected in step S304 in the reflection point position information RPn. The position information RPn is output to the object detection unit 314.

  On the other hand, when the process proceeds to step S405, the vehicle position estimation unit 312 determines the time tn-k before k samples based on the vehicle data Vn-k at the time tn-k input from the vehicle information sensor 200. Vehicle position information VPn-k is estimated and output to the reflection point positioning unit 313. The reflection point positioning unit 313 acquires the vehicle position information VPn-k from the vehicle position estimation unit 312 and proceeds to step S406. The reflection point positioning unit 313 stores the vehicle position information VPn-k at the time tn-k input from the vehicle position estimation unit 312 in the storage unit, and the vehicle position information from the storage unit in step S405. The configuration may be such that VPn-k is read.

  In step S406, the reflection point positioning unit 313 acquires the own vehicle position information VPn-k acquired in step S405, the known mounting position of the distance sensor 100 on the own vehicle, and the known sensor orientation information of the distance sensor 100. Thus, the position and orientation of the distance sensor 100 at time tn-k in the absolute coordinate system are calculated as sensor position information SPn-k. The reflection point positioning unit 313 obtains the sensor position information SPn-k at the time tn-k by performing the calculation in this way, and proceeds to step S407.

  In step S407, the reflection point positioning unit 313 calculates the difference between the position of the distance sensor 100 at time tn included in the sensor position information SPn and the position of the distance sensor 100 at time tn-k included in the sensor position information SPn-k. Calculate a vector. If the reflection point positioning unit 313 determines that the sensor displacement amount, which is the magnitude of the calculated difference vector, is larger than the constant d (that is, YES), the process proceeds to step S408, where the sensor displacement amount is a constant. If it is determined that it is less than or equal to d (ie, NO), the process proceeds to step S411. The constant d is a positive number determined as the lower limit value of the sensor displacement amount, and can be designed as appropriate.

  In step S408, the reflection point positioning unit 313 acquires the distance track update value Dn-k ″ at the time tn-k included in the established distance track selected in step S304, and proceeds to step S409.

  In step S409, the reflection point positioning unit 313 detects the sensor position information SPn acquired in step S302, the distance track update value Dn ″ acquired in step S401, the sensor position information SPn−k acquired in step S406, and step S408. Using the distance track update value Dn−k ″ acquired in step S2, the reflection point position Rn at time tn is calculated for the object corresponding to the established distance track selected in step S304.

  Here, the calculation of the reflection point position Rn executed in step S409 will be described with reference to FIG. FIG. 7 is an explanatory diagram for explaining a method of calculating the reflection point position Rn of the object by the reflection point positioning unit 313 according to Embodiment 1 of the present invention. More specifically, FIG. 7A shows a case where the calculation of the reflection point position Rn is possible, and FIG. 7B shows a case where the calculation of the reflection point position Rn is impossible.

  As shown in FIG. 7A, the reflection point position Rn of the object is a circle whose center is the sensor position that is the position of the distance sensor 100 included in the sensor position information SPn and whose radius is the distance track update value Dn ″. Further, on such a circumference, the existence range of the reflection point position Rn of the object is represented by an arc An from the azimuth range determined from the sensor azimuth and sensor viewing angle included in the sensor position information SPn. That is, the reflection point position Rn of the object exists somewhere on the arc, and the existence range of the reflection point position Rn-k of the object at the time tn-k is also the arc An−. k can be represented.

  The reflection point position of the object at each time is the shortest distance point at which the distance from the distance sensor 100 to the object is the shortest, and is moving as the distance sensor 100 moves. Therefore, a common tangent line between the arc An representing the existence range of the reflection point position Rn at time tn and the arc An-k representing the existence range of the reflection point position Rn-k at time tn-k is drawn and is in contact with the common tangent line. A contact point on the arc An is set as a reflection point position Rn at time tn.

  On the other hand, depending on the positional relationship between the circular arc An and the circular arc An-k, the reflection point position at time tn may not be estimated. For example, as shown in FIG. 7B, when the movement directions of the position and orientation of the distance sensor 100 are the same between the time tn-k and the time tn, the arc An and the arc An-k Except for the line connecting the end points, there is no common tangent line between the arc An and the arc An-k. In this case, the reflection point positioning unit 313 cannot calculate the reflection point position Rn at time tn. For example, when the sensor displacement amount is not sufficiently large, the reflection point position does not move between the time tn-k and the time tn, and therefore there is no common tangent line between the arc An and the arc An-k. Also in this case, the reflection point positioning unit 313 cannot calculate the reflection point position Rn at time tn.

  Returning to the description of FIG. 6B, in step S410, the reflection point positioning unit 313 determines whether or not the reflection point position Rn at time tn has been calculated in step S409. When the reflection point positioning unit 313 determines that the reflection point position Rn can be calculated (that is, YES), the process proceeds to step S412 and the reflection point position Rn cannot be calculated. If (ie, NO) is determined, the process proceeds to step S411.

  When the process proceeds to step S411, the reflection point positioning unit 313 increments the variable k, returns to step S403, and executes the processes after step S403 again.

  On the other hand, when the process proceeds to step S412, the reflection point positioning unit 313 determines that the reflection point position Rn at time tn can be uniquely determined for the object corresponding to the established distance track selected in step S304. It is determined that it is possible, and a series of processing ends. In step S412, the reflection point positioning unit 313 sets the state of the reflection point position information RPn to “positioning possible” for the object corresponding to the established distance track selected in step S304. Further, the reflection point positioning unit 313 includes, in the reflection point position information RPn, information that associates the time tn, the reflection point position Rn at the time tn in the absolute coordinate system, and the established distance track selected in Step S304. The reflection point position information RPn is output to the object detection unit 314.

  Next, the parking space information estimation process performed by the object detection unit 314 executed in step S105 of FIG. 2 will be described with reference to FIGS. FIG. 8A is a flowchart showing a series of operations of parking space information estimation processing by the object detection unit according to Embodiment 1 of the present invention. FIG. 8B is a flowchart showing a series of reflection point tracking operations executed in step S505 of FIG. 8A. FIG. 8C is a flowchart showing a series of operations of the pre-established reflection point track process executed in step S610 of FIG. 8B. Further, FIG. 8D is a flowchart showing a series of operations of the post-establishment reflection point wake process executed in step S611 of FIG. 8B.

  Note that the processing in steps S501 to S504, S507, and S508 in FIG. 8A is the same as the processing in steps S301 to S304, S306, and S307 in FIG. 6A. Therefore, in the following, description of these steps that are the same in FIG. 8A and the previous FIG. 6A will be omitted, the reflection point tracking operation executed in step S505, and the parking space information estimation executed in step S506. The operation will be described.

  First, the reflection point tracking operation executed in step S505 will be described. Here, a time series of a series of reflection point positions input from the reflection point positioning unit 313 is referred to as a reflection point track. The object detection unit 314 updates the reflection point track based on the reflection point position information input from the reflection point positioning unit 313. The reflection point track includes at least a position vector component and a velocity vector component. In updating the reflection point track, for example, an update value can be generated using a least square method or a Kalman filter, and the update value thus generated is called a reflection point track update value.

  First, in step S601, if the object detection unit 314 has generated the reflection point track of the object corresponding to the established distance track selected in step S504, the reflection point position of the predicted time tn is predicted for this reflection point track. Is generated as the reflection point track prediction value Rn ′, and the process proceeds to step S602.

  In step S602, the object detection unit 314 acquires the reflection point position information RPn of the object corresponding to the established distance wake selected in step S504 from the reflection point positioning unit 313, and proceeds to step S603.

  In step S603, the object detection unit 314 determines whether or not the state of the reflection point position information RPn acquired in step S602 is “positioning is possible”. When the object detection unit 314 determines that the state of the reflection point position information RPn is “positioning is possible” (that is, YES), the process proceeds to step S604, and the state of the reflection point position information RPn is “positioning indefinite”. If it is determined that NO (ie, NO), the process proceeds to step S605.

  When the process proceeds to step S604, the object detection unit 314 acquires the reflection point position Rn at time tn included in the reflection point position information RPn acquired in step S602, and the process proceeds to step S609. Note that the reflection point at the reflection point position Rn acquired in this way by the object detection unit 314 is referred to as a positionable reflection point.

  On the other hand, when the process proceeds to step S605, the object detection unit 314 determines whether the state of the reflection point track of the object corresponding to the established distance track selected in step S504 is “established”. If the object detection unit 314 determines that the state of the reflection point track is “established” (that is, YES), the process proceeds to step S606, and the state of the reflection point track is “unestablished” ( That is, if it is determined as NO), the process proceeds to step S609. Note that the object detection unit 314 proceeds to step S609 even when the reflection point track has not been generated at time tn.

  In step S606, the object detection unit 314 determines whether or not the distance sensor 100 is moving at time tn. For example, if it is determined in step S407 in FIG. 6B that the sensor displacement amount is greater than d, the object detection unit 314 determines that the distance sensor 100 is moving at time tn. If the object detection unit 314 determines that the distance sensor 100 is moving at time tn (ie, YES), the object detection unit 314 proceeds to step S608, and the distance sensor 100 is not moving at time tn (ie, NO). ), The process proceeds to step S607.

  If the process proceeds to step S607, the object detection unit 314 corrects the reflection point track predicted value Rn ′ generated in step S601, and then proceeds to step S608. Specifically, the object detection unit 314 replaces the position vector component included in the reflection point track predicted value Rn ′ with the position vector component included in the reflection point track update value Rn−1 ″ at time tn−1. In addition, the velocity vector component included in the reflection point wake prediction value Rn ′ is a zero vector, and if it is a higher-order vector component, all the zero vectors are used, and thus the reflection point wake generated in step S601. The reason for correcting the predicted value Rn ′ is to consider that the reflection point position Rn has not moved when it is determined in step S606 that the distance sensor 100 has not moved at time tn.

  In step S608, the object detection unit 314 estimates the reflection point position Rn at time tn from the reflection point track predicted value Rn ′ and the reflection point position information RPn. The reflection point at the reflection point position Rn estimated in this way by the object detection unit 314 is referred to as a positioning indefinite reflection point.

  Here, estimation of the reflection point position Rn executed in step S608 will be described with reference to FIGS. 9A and 9B. FIG. 9A is an explanatory diagram for explaining a method of estimating the reflection point position Rn of the object by the object detection unit 314 according to Embodiment 1 of the present invention. FIG. 9B is an explanatory diagram for explaining another example of the method of estimating the reflection point position Rn of the object by the object detection unit 314 according to Embodiment 1 of the present invention.

  When step S608 is executed, the state of the reflection point position information RPn acquired in step S602 is “positioning indefinite”. The reflection point position information RPn corresponds to the established distance track selected in step S504.

As shown in FIG. 9A, the reflection point position Rn of the object is centered on the sensor position that is the position of the distance sensor 100 included in the sensor position information SPn, and is anywhere on the circumference whose radius is the distance track update value Dn ″. Further, in such a circumference, the existence range of the reflection point position Rn of the object can be represented by an arc An from the azimuth range determined from the sensor azimuth and sensor viewing angle included in the sensor position information SPn. This concept is the same as in Fig. 7 (a) Therefore, the object detection unit 314 reflects a point on the arc An that minimizes the distance from the reflection point track predicted value Rn 'at time tn to the arc An. Let it be point position Rn.
Note that the object detection unit 314 may be configured to estimate the reflection point position Rn by an estimation method as shown in FIG. 9B instead of the estimation method as shown in FIG. 9A. In this case, the object detection unit 314 sets, as the reflection point position Rn, an end point having a shorter distance from the reflection point track predicted value Rn ′ among the two end points of the arc An corresponding to the time tn.

  Returning to the description of FIG. 8B, in step S609, the object detection unit 314 determines whether or not the state of the reflection point track of the object corresponding to the established distance track selected in step S504 is “established”. If the object detection unit 314 determines that the state of the reflection point track is “established” (that is, YES), the process proceeds to step S611, and the state of the reflection point track is “unestablished” ( That is, if it is determined as NO), the process proceeds to step S610. Note that the object detection unit 314 proceeds to step S610 even when the reflection point track has not been generated at time tn.

  Next, the operation of the pre-established reflection point track process executed in step S610 will be described with reference to FIG. 8C.

  First, in step S701, the object detection unit 314 determines whether or not the state of the reflection point position information RPn acquired in step S602 is “positioning is possible”. If the object detection unit 314 determines that the state of the reflection point position information RPn is “positioning is possible” (that is, YES), the process proceeds to step S703, and the state of the reflection point position information RPn is “positioning indefinite”. If it is determined that NO (that is, NO), the process proceeds to step S702.

  When the process proceeds to step S702, the object detection unit 314 deletes the reflection point track in an unestablished state from the reflection point track of the object corresponding to the established distance track selected in step S504, and performs a series of processing. finish.

  On the other hand, when the process proceeds to step S703, the object detection unit 314 determines whether the reflection point track of the object corresponding to the established distance track selected in step S504 has been generated. If the object detection unit 314 determines that this reflection point track has been generated (ie, YES), the object detection unit 314 proceeds to step S705, and determines that this reflection point track has not been generated (ie, NO). In step S704, the process proceeds to step S704.

  When the process proceeds to step S704, the object detection unit 314 generates a reflection point track starting from the reflection point position Rn acquired in step S604, and sets the state of the reflection point track to “unestablished” and “ Set to “Uninitialized” to end the series of processing.

  On the other hand, when the process proceeds to step S705, the object detection unit 314 sets a gate based on the reflection point wake prediction value Rn ′ at time tn. The gate here is a fixed region centered on the reflection point track predicted value Rn ′, similarly to the gate set in step S204 of FIG. 3A. For example, an ellipse determined from a constant multiple of a predetermined reflection point position error covariance matrix is set as the gate region. When the Kalman filter is used to update the reflection point track, an ellipse determined from a constant multiple of the position residual covariance matrix determined by the sum of the position component of the prediction error covariance matrix and the reflection point position error covariance matrix is used as the gate region. It may be set.

  In step S706, the object detection unit 314 determines whether or not the reflection point position Rn acquired in step S604 is within the gate set in step S705. When the object detection unit 314 determines that the reflection point position Rn enters the gate (that is, YES), the object detection unit 314 proceeds to step S707 and does not enter the gate (that is, the reflection point position Rn (that is, YES). If NO is determined, the process proceeds to step S702.

  When the process proceeds to step S707, the object detection unit 314 determines whether the state of the reflection point wake of the object corresponding to the established distance wake selected in step S504 has been initialized. If the object detection unit 314 determines that the state of the reflection point wake has been initialized (ie, YES), the object detection unit 314 proceeds to step S709, and the state of the reflection point wake has not been initialized (ie, has been completed) If NO is determined, the process proceeds to step S708.

  When the process proceeds to step S708, the object detection unit 314 uses the reflection point position Rn acquired in step S604 for the reflection point track of the object corresponding to the established distance track selected in step S504, and updates the reflection point track update value. Rn ″ is generated, and the state of the reflection point track is changed from “uninitialized” to “initialized”, and a series of processing ends.

  On the other hand, when the process proceeds to step S709, the object detection unit 314 uses the reflection point position Rn acquired in step S604 for the reflection point track of the object corresponding to the established distance track selected in step S504. The update value Rn ″ is generated, the state of the reflection point track is changed from “not established” to “established”, and the series of processes is ended.

  In the above description, the reflection point track is established if the reflection point position Rn is a positioning reflection point for three consecutive times. However, the positioning reflection and reflection necessary for establishing the reflection point track is described. The number of consecutive points may be four or more. Also, for example, after a positioning reflex point is obtained twice in succession, when a positioning reflex point is obtained within X times of sampling (X is an integer of 1 or more), a reflection point track is established. It is good also as composition to do.

  Next, the operation of the post-establishment reflection point wake process executed in step S611 will be described with reference to FIG. 8D.

  First, in step S801, the object detection unit 314 sets a gate based on the reflection point wake prediction value Rn ′ at time tn, similarly to step S705 in FIG. 8C, and proceeds to step S802.

  In step S802, the object detection unit 314 determines whether or not the reflection point position Rn at time tn falls within the gate set in step S801. If the object detection unit 314 determines that the reflection point position Rn enters the gate (that is, YES), the object detection unit 314 proceeds to step S803 and does not enter the gate (that is, the reflection point position Rn (that is, YES). If NO is determined, the process proceeds to step S804.

  In step S803, the object detection unit 314 calculates the reflection point position Rn and the reflection point track prediction value Rn ′ at the time tn for the reflection point track of the object corresponding to the established distance track selected in step S504. The reflection point wake update value Rn ″ is generated using this, and a series of processing ends.

  On the other hand, when the process proceeds to step S804, the object detection unit 314 uses the reflection point track predicted value Rn ′ as the reflection point track update value Rn ″ for the reflection point track of the object corresponding to the established distance track selected in step S504. Then, a series of processing ends.

  Next, the parking space information estimation operation executed in step S506 will be described. In step S506, the object detection unit 314 estimates parking space information from the reflection point track obtained in step S505.

  Here, when estimating the side surface of the parked vehicle as the parking space information, the object detection unit 314 estimates, for example, a broken line connecting the reflection point track update values at each time included in the reflection point track as the side surface. Can do.

  Alternatively, the side surface of the object may be estimated by approximating the parked vehicle with a rectangle and switching the estimation method according to the driving situation of the host vehicle. For example, in the driving situation where the vehicle is searching for a parking space, the reflection point wake becomes one side of the approximate rectangle when passing the side of the parked vehicle, so the reflection point wake at each time included in the reflection point wake The side of the parked vehicle is estimated by approximating the updated value with a line segment. On the other hand, in the driving situation where the vehicle is guided to the parking space, the side of the parked vehicle is estimated by determining the side of the approximate rectangle corresponding to the reflection point track from the positional relationship between the parked vehicle and the vehicle.

  When estimating the corner of a parked vehicle as parking space information, for example, the corner can be estimated by applying the conventional technique described in Patent Document 1.

  Here, the parking space information estimation executed in step S506 will be described with reference to FIG. FIG. 10 is an explanatory diagram for describing a method for estimating parking space information by the object detection unit 314 according to Embodiment 1 of the present invention.

  As shown in FIG. 10, the reflection point track generated when the vehicle moves in the direction of the arrow includes the reflection point track update value at each time when the distance sensor 100 detects the distance data. The reflection point track update value at each time is the reflection point track update value generated using the reflection point position of the positioning reflection point or the reflection point track update generated using the reflection point position of the positioning indefinite reflection point. Value. Thus, the side surface of the parked vehicle can be estimated from the reflection point track update value at each time included in the reflection point track.

  In addition, although the method mentioned above was illustrated as a method of estimating parking space information based on the reflective point position of an object, it is not limited to this. That is, the present invention has a technical feature of accurately estimating the position of the reflection point of the object, and can be applied to the conventional technique for estimating the parking space information based on the position of the reflection point of the object.

  In the first embodiment, the reflection point track is established based on the positioning reflection point, and the reflected point track is used to follow the change in the reflection point position accompanying the movement of the sensor position. It can be assumed that the change amount of the reflection point position is substantially the same as the change amount of the sensor position.

  Therefore, in step S706 of FIG. 8C, the object detection unit 314 has a change amount of the reflection point position larger than a constant multiple even if the reflection point position Rn at time tn is in the gate. If larger, step S702 may be executed instead of step S707. Similarly, in step S802 in FIG. 8D, the object detection unit 314 changes the reflection point position change amount by a constant multiple even if the reflection point position Rn at time tn is within the gate. If it is larger, step S804 may be executed instead of step S803.

  As described above, when the change amount of the reflection point position is larger than the constant multiple of the change amount of the sensor position, each reflection is performed even when the estimated reflection point position Rn is included in the second gate. You may comprise so that tracking of the time transition of a point position may be stopped.

  The change amount of the reflection point position means the magnitude of a difference vector between the reflection point position Rn at time tn and the reflection point track update value Rn-1 ″ at time tn−1. The amount of change means the magnitude of a difference vector between the position of the distance sensor 100 at time tn included in the sensor position information SPn and the position of the distance sensor 100 at time tn-1 included in the sensor position information SPn-1. To do.

  As described above, according to the first embodiment, it is predicted that the distance sensor is detected at time tn from the distance track generated by following the time transition of the distance data detected by the distance sensor before time tn. If the distance data Dn detected by the distance sensor is included at the time tn in the first gate that generates the distance track predicted value Dn ′ and is set around the generated distance track predicted value Dn ′, the distance data The distance track update value Dn ″ is generated by using Dn and the distance track predicted value Dn ′ to update the distance track. If the distance data Dn is not included in the first gate, the distance track predicted value Dn ′ is determined as the distance. The track update value Dn ″ and the new distance track starting from the distance data Dn are started to follow the time transition of the distance data, and the distance data is set for each different object. Betsushite comprising an object discrimination unit for generating a distance wake.

  As a result, the noise component of the distance data detected by the distance sensor can be removed, and the object corresponding to the distance data can be discriminated with high accuracy.

  Further, the sensor position information SPn at time tn is calculated, and the sensor position information SPn out of the circumference with the radius as the distance track update value Dn ″ with the position specified by the sensor position information SPn for each different object as the center. The position of the contact point on the arc An that is in contact with the common tangent line of the calculated arc An and one of the arcs calculated at a time before the arc An is calculated. A reflection point positioning unit that estimates the reflection point position Rn of the object at time tn is further provided.

  Furthermore, by specifying the position of each object from the reflection point position Rn estimated for each different object and each reflection point position estimated at a time before the reflection point position Rn by the reflection point positioning unit, And an object detection unit for estimating parking space information.

  Thereby, it is possible to accurately estimate the reflection point position of the object that is the detection target of the distance sensor with a simple configuration. Further, since the reflection point position can be estimated with high accuracy in this way, the parking space information of the object can also be estimated with high accuracy from the reflection point position.

  Further, the object detection unit generates a reflection point track predicted value Rn ′ at time tn by following the time transition of each reflection point position estimated at the time before the reflection point position Rn by the reflection point positioning unit. When the reflection point positioning unit cannot estimate the reflection point position Rn, the position on the circular arc An having the shortest distance from the reflection point track predicted value Rn ′ is estimated as the reflection point position Rn.

  As a result, even when the reflection point position Rn at time tn cannot be calculated due to the positional relationship between the arc An calculated by the reflection point positioning unit and each arc calculated at a time prior to the arc An. The object detection unit can estimate the reflection point position Rn.

  Further, if the estimated reflection point position Rn is included in the second gate set around the generated reflection track predicted value Rn ′, the object detection unit uses the reflection point position Rn to reflect the reflection point track. When the update value Rn ″ is generated and the tracking of the time transition of each reflection point position is continued, and the position of the distance sensor does not move at the time tn, the time tn which is the time before the time tn. The reflection point track predicted value Rn ′ is corrected so that the amount of change between the reflection point track update value Rn−1 ″ of −1 and the reflection point position Rn becomes zero. Thereby, when the distance sensor has not moved at time tn, it can be considered that the reflection point position Rn has not moved.

Embodiment 2. FIG.
In the first embodiment, the case where the reflection point position Rn is estimated from the common tangent line between the arc An corresponding to the time tn and the arc corresponding to one time before the time tn has been described. On the other hand, in Embodiment 2 of the present invention, the reflection point position Rn is estimated from the envelope of the arc An corresponding to the time tn and the arcs corresponding to two or more times before the time tn. Will be described.

  FIG. 11A is a flowchart showing a series of operations of reflection point positioning processing by the reflection point positioning unit 313 according to Embodiment 2 of the present invention. FIG. 11B is a flowchart showing a series of operations of reflection point positioning executed in step S905 of FIG. 11A.

  In the second embodiment, the description of the same points as in the first embodiment is omitted, and the operation of the reflection point positioning process by the reflection point positioning unit 313 is different from that in the first embodiment. The explanation will be focused on. Further, when FIG. 11A is compared with FIG. 6A, the reflection point positioning operations in steps S905 and S305 are different, and the operations in the other steps are the same. Therefore, the description will focus on the flowchart of FIG. 11B, which is a series of processing of the reflection point positioning operation in step S905.

  First, the reflection point positioning unit 313 sequentially executes steps S1001 to S1003 similar to steps S401 to S403 in FIG. 6B.

  In step S1003, when the reflection point positioning unit 313 determines that the detection time difference is equal to or greater than the constant T, or the variable k is larger than the constant kmax (that is, YES), the process proceeds to step S1009, and the detection time difference Is determined to be less than the constant T and the variable k is equal to or less than the constant kmax (that is, NO), the process proceeds to step S1004.

  When the process proceeds to step S1004, the reflection point positioning unit 313 sequentially executes steps S1004 to S1006 similar to steps S405 to 407 in FIG. 6B.

  In step S1007, the reflection point positioning unit 313 acquires the distance track update value Dn-k ″ at time tn-k included in the established distance track selected in step S904, and proceeds to step S1008.

  In step S1008, the reflection point positioning unit 313 increments the variable k, returns to step S1003, and executes the processing after step S1003 again.

  On the other hand, when the process proceeds to step S1009, the reflection point positioning unit 313, the sensor position information SPn acquired in step S902, the distance track update value Dn ″ acquired in step S1001, and the plurality of sensors acquired in step S1005. Using the position information and the plurality of distance track update values acquired in step S1007, the reflection point position Rn at time tn is calculated for the object corresponding to the established distance track selected in step S904.

  Here, the calculation of the reflection point position Rn executed in step S1009 will be described with reference to FIG. FIG. 12 is an explanatory diagram for describing a method of calculating the reflection point position Rn of the object by the reflection point positioning unit 313 according to Embodiment 2 of the present invention.

  As shown in FIG. 12, the existence range of the reflection point position Rn of the object can be represented by an arc An. Similar to the time tn, the existence range of the reflection point position Rn-k of the object at the time tn-k can also be represented by the arc An-k. This way of thinking is the same as in FIG.

  Further, in the second embodiment, unlike the first embodiment, the sensor position information and the distance track update value corresponding to one time before the time tn are not two or more times before the time tn. Corresponding sensor position information and distance track update values are acquired. Therefore, as shown in FIG. 12, the reflection point position Rn is calculated using an arc representing the existence range of the reflection point position at each time before the time tn and the arc An at the time tn.

  The reflection point position of the object at each time is the shortest distance point at which the distance from the distance sensor 100 to the object is the shortest, and moves with the movement of the sensor. Therefore, an envelope of an arc indicating the existence range of the reflection point at the time tn and an arc indicating the existence range of the reflection point position at each time before the time tn is drawn, and a contact point on the arc An in contact with the envelope is set to the time. The reflection point position Rn is tn.

  Returning to the description of FIG. 11B, in step S1010, the reflection point positioning unit 313 determines whether or not the reflection point position Rn at time tn has been calculated in step S1009. When the reflection point positioning unit 313 determines that the reflection point position Rn can be calculated (that is, YES), the process proceeds to step S1012 and the reflection point position Rn cannot be calculated. If (ie, NO) is determined, the process proceeds to step S1011.

  When the process proceeds to step S1011, the reflection point positioning unit 313 executes the same process as in step S404 of FIG. 6B, and ends the series of processes. On the other hand, when the process proceeds to step S1012, the reflection point positioning unit 313 performs the same process as step S412 of FIG. 6B and ends the series of processes.

  As described above, according to the second embodiment, instead of estimating the reflection point position Rn using the common tangent line with respect to the first embodiment, the arc An and a plurality of times before the arc An are obtained. A reflection point positioning unit that estimates the position of the contact point on the arc An that is in contact with the calculated envelope with each arc as the reflection point position Rn is provided. Thereby, the same effect as in the first embodiment can be obtained.

Embodiment 3 FIG.
In the first embodiment, the case where the reflection point position Rn is estimated from the common tangent line between the arc An corresponding to the time tn and the arc corresponding to one time before the time tn has been described. On the other hand, in Embodiment 3 of the present invention, the reflection point position Rn is estimated from the common tangent line of the arc An corresponding to the time tn and each arc corresponding to one or more times before the time tn. Will be described.

  FIG. 13A is a flowchart showing a series of operations of reflection point positioning processing by the reflection point positioning unit 313 according to Embodiment 3 of the present invention. FIG. 13B is a flowchart showing a series of operations of reflection point positioning executed in step S1105 of FIG. 13A.

  In the third embodiment, description of points that are the same as in the second embodiment is omitted, and the operation of the reflection point positioning process by the reflection point positioning unit 313 is different from that in the second embodiment. The explanation will be focused on. Further, when FIG. 13A is compared with FIG. 11A, the reflection point positioning operations in steps S1105 and S905 are different, and the operations in other steps are the same. Therefore, the description will focus on the flowchart of FIG. 13B, which is a series of processing of the reflection point positioning operation in step S1105.

  First, the reflection point positioning unit 313 sequentially executes steps S1201 to S1203 similar to steps S1001 to S1003 in FIG. 11B.

  In step S1203, if the reflection point positioning unit 313 determines that the detection time difference is equal to or greater than the constant T, or the variable k is greater than the constant kmax (ie, YES), the process proceeds to step S1210, and the detection time difference is determined. Is determined to be less than the constant T and the variable k is equal to or less than the constant kmax (ie, NO), the process proceeds to step S1204.

  When the process proceeds to step S1204, the reflection point positioning unit 313 sequentially executes steps S1204 to S1207 similar to steps S1004 to S1007 in FIG. 11B, and then proceeds to step S1208.

  In step S1208, the reflection point positioning unit 313 acquires the sensor position information SPn acquired in step S1102, the distance track update value Dn ″ acquired in step S1201, the sensor position information SPn-k acquired in step S1205, and step S1207. Using the distance track update value Dn−k ″ acquired in step S1, the reflection point temporary position RTn at time tn is calculated for the object corresponding to the established distance track selected in step S1104, and the process proceeds to step S1209. That is, the reflection point positioning unit 313 calculates a common tangent line between the arc An representing the existence range of the reflection point position Rn at time tn and the arc An-k representing the existence range of the reflection point position Rn-k at time tn-k. Then, a contact point on the arc An that is in contact with the common tangent line is set as a reflection point temporary position RTn at time tn.

  In step S1209, the reflection point positioning unit 313 increments the variable k, returns to step S1203, and again executes the processing from step S1203.

  On the other hand, when the process proceeds to step S1210, the reflection point positioning unit 313 determines whether or not the number of calculated reflection point temporary positions RTn at time tn is one or more. When the reflection point positioning unit 313 determines that this number is one or more (that is, YES), the process proceeds to step S1212 and determines that this number is less than one (that is, NO). Then, the process proceeds to step S1211.

  In step S1212, the reflection point positioning unit 313 uses the sensor position information SPn acquired in step S1102, the distance wake update value Dn ″ acquired in step S1201, and the reflection point temporary position RTn acquired in step S1208. For the object corresponding to the established distance track selected in step S1104, the reflection point position Rn at time tn is calculated, and the process proceeds to step S1213.

  Here, the calculation of the reflection point position Rn executed in step S1212 will be described with reference to FIG. FIG. 14 is an explanatory diagram for explaining a method of calculating the reflection point position Rn of the object by the reflection point positioning unit 313 according to Embodiment 3 of the present invention.

  As shown in FIG. 14, the existence range of the reflection point position Rn of the object can be represented by an arc An. Similar to the time tn, the existence range of the reflection point position Rn-k of the object at the time tn-k can also be represented by the arc An-k. This way of thinking is the same as in FIG.

  In the third embodiment, unlike the second embodiment, while the variable k is incremented, a common tangent line between the arc An and the arc An-k is drawn, and the arc tangent to the common tangent line is drawn. The process of setting the contact point as the reflection point temporary position RTn at time tn (ie, step S1208) is repeatedly executed.

  Therefore, as shown in FIG. 14, the reflection point position Rn is calculated using a plurality of reflection point temporary positions RTn. FIG. 14 illustrates a case where there are three temporary reflection point positions RTn. More specifically, the reflection point temporary position 1 corresponding to the contact point on the arc An that is in contact with the common tangent line 1 of the arc An and the arc An-1, and the arc that is in contact with the common tangent line 2 of the arc An and the arc An-2. The case where there is a reflection point temporary position 2 corresponding to the contact point on An and a reflection point temporary position 3 corresponding to the contact point on the arc An that is in contact with the common tangent line 3 of the arc An and the arc An-3 is illustrated.

  Thus, each of the plurality of reflection point temporary positions RTn is a point on the arc An. Each of the plurality of temporary reflection point positions RTn has a different orientation as viewed from the sensor position at time tn included in the sensor position information SPn, while the distance from the sensor position at time tn is all the same. Therefore, the distance from the sensor position at time tn is the distance track update value Dn ″, and is located in an azimuth corresponding to the average value of the respective azimuths of the reflection point temporary position RTn viewed from the sensor position at time tn. The point is defined as a reflection point position Rn at time tn, that is, a point on the arc An located at an azimuth corresponding to the average value of the respective azimuths of the reflection point temporary position RTn as viewed from the sensor position at time tn. The reflection point position Rn is tn.

  Note that a point on the arc An located in an azimuth corresponding to the median value of each azimuth of the reflection point temporary position RTn viewed from the sensor position at the time tn may be set as the reflection point position Rn at the time tn.

  Returning to the description of FIG. 13B, when the process proceeds to step S <b> 1211, the reflection point positioning unit 313 executes the same process as step S <b> 1011 of FIG. 11B, and ends a series of processes. On the other hand, when the process proceeds to step S1213, the reflection point positioning unit 313 executes the same process as step S1012 of FIG. 11B and ends the series of processes.

  As described above, according to the third embodiment, instead of estimating the reflection point position Rn using the common tangent line with respect to the first embodiment, the arc An and a plurality of times before the arc An are obtained. The position of the contact point on the arc An that is in contact with the calculated common tangent to each arc is the reflection point temporary position RTn, and the respective orientations of the reflection point temporary position RTn viewed from the position of the distance sensor included in the sensor position information SPn. A reflection point positioning unit is provided that estimates a point on the arc An located in the azimuth corresponding to the average value of the above or the azimuth corresponding to the median value of each azimuth as the reflection point position Rn. Thereby, the same effect as in the first embodiment can be obtained.

  DESCRIPTION OF SYMBOLS 100 Distance sensor, 200 Vehicle information sensor, 300 Parking assistance apparatus, 310 Object detection apparatus, 311 Object discrimination part, 312 Own vehicle position estimation part, 313 Reflection point positioning part, 314 Object detection part, 320 Vehicle control apparatus.

Claims (9)

It is mounted on the host vehicle, irradiates the detection target object with a detection wave, and obtains the detection wave reflected at the reflection point position on the object corresponding to the shortest distance to the object. While moving the own vehicle provided with a distance sensor that detects distance data and a vehicle information sensor that is mounted on the own vehicle and detects a state related to the speed and traveling direction of the own vehicle as the own vehicle data, An object for estimating parking space information of the own vehicle by acquiring a time-series detection result by the distance sensor and the vehicle information sensor and specifying the reflection point position of the object that changes in time series from the detection result A detection device,
A distance track prediction value Dn ′ predicted to be detected by the distance sensor at time tn is generated from a distance track generated by following the time transition of the distance data detected by the distance sensor before time tn. If the distance data Dn detected by the distance sensor at the time tn is included in the first gate set around the generated distance track predicted value Dn ′, the distance data Dn and the distance track The distance track update value Dn ″ is generated using the predicted value Dn ′ to update the distance track, and if the distance data Dn is not included in the first gate, the distance track predicted value Dn ′ is The distance track update value Dn ″ is set, and a new distance track starting from the distance data Dn is newly started to follow the time transition of the distance data. An object discriminating unit that discriminates distance data and generates a distance track;
A vehicle position estimation unit that estimates the position of the vehicle and the direction of the vehicle as the vehicle position information VPn from the vehicle data Vn detected by the vehicle information sensor at the time tn;
Based on the own vehicle position information VPn, the known mounting position of the distance sensor on the own vehicle, and the known sensor direction information of the distance sensor, the position and direction of the distance sensor at the time tn are: It is calculated as sensor position information SPn, and for each different object, the sensor position information SPn out of the circumference with the radius specified as the distance track update value Dn ″ centered on the position specified by the sensor position information SPn. The arc An of the specified orientation is calculated, and the position of the contact point on the arc An that is in contact with the common tangent line of the calculated arc An and one of the arcs calculated at a time before the arc An is calculated. , A reflection point positioning unit that estimates the reflection point position Rn of the object at time tn;
From the reflection point position Rn estimated for each different object by the reflection point positioning unit and each reflection point position estimated at a time prior to the reflection point position Rn by the reflection point positioning unit, each object An object detection unit that estimates the parking space information by specifying the position of
With
The reflection point positioning unit is
Instead of estimating the reflection point position Rn using the common tangent, the contact point on the arc An that touches the common tangent of the arc An and each arc calculated at a plurality of times before the arc An is calculated. The position is a reflection point temporary position RTn,
An arc positioned in an azimuth corresponding to an average value of the respective azimuths of the reflection point temporary position RTn viewed from the position of the distance sensor included in the sensor position information SPn, or in an azimuth corresponding to a median value of the respective azimuths. A point on An is estimated as the reflection point position Rn.
Object detection apparatus. It is mounted on the host vehicle, irradiates the detection target object with a detection wave, and obtains the detection wave reflected at the reflection point position on the object corresponding to the shortest distance to the object. While moving the own vehicle provided with a distance sensor that detects distance data and a vehicle information sensor that is mounted on the own vehicle and detects a state related to the speed and traveling direction of the own vehicle as the own vehicle data, An object for estimating parking space information of the own vehicle by acquiring a time-series detection result by the distance sensor and the vehicle information sensor and specifying the reflection point position of the object that changes in time series from the detection result A detection device,
A distance track prediction value Dn ′ predicted to be detected by the distance sensor at time tn is generated from a distance track generated by following the time transition of the distance data detected by the distance sensor before time tn. If the distance data Dn detected by the distance sensor at the time tn is included in the first gate set around the generated distance track predicted value Dn ′, the distance data Dn and the distance track The distance track update value Dn ″ is generated using the predicted value Dn ′ to update the distance track, and if the distance data Dn is not included in the first gate, the distance track predicted value Dn ′ is The distance track update value Dn ″ is set, and a new distance track starting from the distance data Dn is newly started to follow the time transition of the distance data. An object discriminating unit that discriminates distance data and generates a distance track;
A vehicle position estimation unit that estimates the position of the vehicle and the direction of the vehicle as the vehicle position information VPn from the vehicle data Vn detected by the vehicle information sensor at the time tn;
Based on the own vehicle position information VPn, the known mounting position of the distance sensor on the own vehicle, and the known sensor direction information of the distance sensor, the position and direction of the distance sensor at the time tn are: It is calculated as sensor position information SPn, and for each different object, the sensor position information SPn out of the circumference with the radius specified as the distance track update value Dn ″ centered on the position specified by the sensor position information SPn. The arc An of the specified orientation is calculated, and the position of the contact point on the arc An that is in contact with the common tangent line of the calculated arc An and one of the arcs calculated at a time before the arc An is calculated. , A reflection point positioning unit that estimates the reflection point position Rn of the object at time tn;
From the reflection point position Rn estimated for each different object by the reflection point positioning unit and each reflection point position estimated at a time prior to the reflection point position Rn by the reflection point positioning unit, each object An object detection unit that estimates the parking space information by specifying the position of
With
The object detection unit is
By following the time transition of each reflection point position estimated at the time before the reflection point position Rn by the reflection point positioning unit, the reflection point track predicted value Rn ′ at the time tn is generated,
When the reflection point positioning unit cannot estimate the reflection point position Rn, the position on the arc An that has the shortest distance from the reflection point track predicted value Rn ′ is estimated as the reflection point position Rn.
Object detection apparatus. It is mounted on the host vehicle, irradiates the detection target object with a detection wave, and obtains the detection wave reflected at the reflection point position on the object corresponding to the shortest distance to the object. While moving the own vehicle provided with a distance sensor that detects distance data and a vehicle information sensor that is mounted on the own vehicle and detects a state related to the speed and traveling direction of the own vehicle as the own vehicle data, An object for estimating parking space information of the own vehicle by acquiring a time-series detection result by the distance sensor and the vehicle information sensor and specifying the reflection point position of the object that changes in time series from the detection result A detection device,
A distance track prediction value Dn ′ predicted to be detected by the distance sensor at time tn is generated from a distance track generated by following the time transition of the distance data detected by the distance sensor before time tn. If the distance data Dn detected by the distance sensor at the time tn is included in the first gate set around the generated distance track predicted value Dn ′, the distance data Dn and the distance track The distance track update value Dn ″ is generated using the predicted value Dn ′ to update the distance track, and if the distance data Dn is not included in the first gate, the distance track predicted value Dn ′ is The distance track update value Dn ″ is set, and a new distance track starting from the distance data Dn is newly started to follow the time transition of the distance data. An object discriminating unit that discriminates distance data and generates a distance track;
A vehicle position estimation unit that estimates the position of the vehicle and the direction of the vehicle as the vehicle position information VPn from the vehicle data Vn detected by the vehicle information sensor at the time tn;
Based on the own vehicle position information VPn, the known mounting position of the distance sensor on the own vehicle, and the known sensor direction information of the distance sensor, the position and direction of the distance sensor at the time tn are: It is calculated as sensor position information SPn, and for each different object, the sensor position information SPn out of the circumference with the radius specified as the distance track update value Dn ″ centered on the position specified by the sensor position information SPn. The arc An of the specified orientation is calculated, and the position of the contact point on the arc An that is in contact with the common tangent line of the calculated arc An and one of the arcs calculated at a time before the arc An is calculated. , A reflection point positioning unit that estimates the reflection point position Rn of the object at time tn;
From the reflection point position Rn estimated for each different object by the reflection point positioning unit and each reflection point position estimated at a time prior to the reflection point position Rn by the reflection point positioning unit, each object An object detection unit that estimates the parking space information by specifying the position of
With
The object detection unit is
By following the time transition of each reflection point position estimated at the time before the reflection point position Rn by the reflection point positioning unit, the reflection point track predicted value Rn ′ at the time tn is generated,
When the reflection point positioning unit cannot estimate the reflection point position Rn, the position of the end point having the shorter distance from the reflection point wake prediction value Rn ′ among the two end points of the arc An is Estimated as reflection point position Rn
Object detection apparatus. The reflection point positioning unit is
Instead of estimating the reflection point position Rn using the common tangent, the contact point on the arc An that touches the envelope of the arc An and each arc calculated at a plurality of times before the arc An is calculated. The position is estimated as the reflection point position Rn.
The object detection apparatus according to claim 2 or 3. The reflection point positioning unit is
Instead of estimating the reflection point position Rn using the common tangent, the contact point on the arc An that touches the common tangent of the arc An and each arc calculated at a plurality of times before the arc An is calculated. The position is a reflection point temporary position RTn,
An arc positioned in an azimuth corresponding to an average value of the respective azimuths of the reflection point temporary position RTn viewed from the position of the distance sensor included in the sensor position information SPn, or in an azimuth corresponding to a median value of the respective azimuths. A point on An is estimated as the reflection point position Rn.
The object detection apparatus according to claim 2 or 3. The object detection unit is
If the estimated reflection point position Rn is included in the second gate set around the generated reflection point track predicted value Rn ′, the reflection point track update value Rn is used using the reflection point position Rn. ”To continue to follow the time transition of each reflection point position,
When the position of the distance sensor does not move at the time tn, the reflection point track update value Rn-1 ″ at the time tn−1 that is one time before the time tn, and the reflection point position Rn The object detection apparatus according to claim 2 , wherein the reflection point wake prediction value Rn ′ is corrected so that a change amount of the reflection point becomes zero. The object detection unit is
The amount of change in the reflection point position, which is the magnitude of the difference vector between the reflection point position Rn and the reflection point track update value Rn−1 ″, is the position of the distance sensor at the time tn and the time tn−1. Is larger than a constant multiple of the change amount of the sensor position, which is the magnitude of the difference vector with respect to the position of the distance sensor, when the estimated reflection point position Rn is included in the second gate. The object detection apparatus according to claim 6, wherein the tracking of the time transition of each reflection point position is stopped even if it exists. The object detection device according to any one of claims 1 to 7,
From the parking space information estimated by the object detection device, a vehicle control device that performs parking assistance for the host vehicle to park in the parking space;
Parking assistance device with It is mounted on the host vehicle, irradiates the detection target object with a detection wave, and obtains the detection wave reflected at the reflection point position on the object corresponding to the shortest distance to the object. While moving the own vehicle provided with a distance sensor that detects distance data and a vehicle information sensor that is mounted on the own vehicle and detects a state related to the speed and traveling direction of the own vehicle as the own vehicle data, An object for estimating parking space information of the own vehicle by acquiring a time-series detection result by the distance sensor and the vehicle information sensor and specifying the reflection point position of the object that changes in time series from the detection result A detection method,
Distance track update for each different object based on the distance data Dn detected by the distance sensor at the time tn from the distance track generated by following the time transition of the distance data detected by the distance sensor before the time tn Generating a value Dn ″;
Estimating the position of the host vehicle and the direction of the host vehicle as the host vehicle position information VPn from the host vehicle data Vn detected by the vehicle information sensor at the time tn;
Based on the own vehicle position information VPn, the known mounting position of the distance sensor on the own vehicle, and the known sensor direction information of the distance sensor, the position and direction of the distance sensor at the time tn are: It is calculated as sensor position information SPn, and for each different object, the sensor position information SPn out of the circumference with the radius specified as the distance track update value Dn ″ centered on the position specified by the sensor position information SPn. The arc An of the specified orientation is calculated, and the position of the contact point on the arc An that is in contact with the common tangent line of the calculated arc An and one of the arcs calculated at a time before the arc An is calculated. , A reflection point positioning step for estimating the reflection point position Rn of the object at time tn;
Estimating the parking space information by identifying the position of each object from the reflection point position Rn and each reflection point position estimated at a time prior to the reflection point position Rn;
Equipped with a,
In the reflection point positioning step,
Instead of estimating the reflection point position Rn using the common tangent, the contact point on the arc An that touches the common tangent of the arc An and each arc calculated at a plurality of times before the arc An is calculated. The position is a reflection point temporary position RTn,
An arc positioned in an azimuth corresponding to an average value of the respective azimuths of the reflection point temporary position RTn viewed from the position of the distance sensor included in the sensor position information SPn, or in an azimuth corresponding to a median value of the respective azimuths. An object detection method for estimating a point on An as the reflection point position Rn .

Images