JP7452374B2 - Object detection device and object detection program - Google Patents

Description

The present invention relates to an object detection device configured to be mounted on a vehicle to detect objects existing around the vehicle, and an object detection program executed by the object detection device .

The device described in Patent Document 1 includes an ultrasonic sensor and an electronic camera. The ultrasonic sensors are installed at the rear of both sides of the vehicle, facing laterally. The ultrasonic sensor detects the distance from the sensor to an object within a detection area. The electronic camera is installed at the rear of the vehicle facing sideways so that its optical axis coincides with the central axis of the ultrasonic sensor so that it overlaps the detection area of the ultrasonic sensor.

The device described in Patent Document 1 uses an ultrasonic sensor to measure the distance to an object and determines whether a parking space exists. Furthermore, this device uses an electronic camera to photograph the parking space and obtain an image in order to detect the depth position of the parking space.

Japanese Patent Application Publication No. 2002-170103

In this type of device, among a plurality of distance measurement points by a distance measurement sensor such as an ultrasonic sensor that corresponds to the surface of an obstacle such as a parked vehicle, an "outlier point" with a large positional error may occur. When such an "outlier point" occurs, for example, even though a parking space actually exists, an erroneous detection of an obstacle caused by such an "outlier point" may result in a false determination that a parking space does not exist. may occur.

The present invention has been made in view of the circumstances illustrated above. That is, the present invention provides, for example, an object detection device and an object detection program that can achieve better object detection accuracy than conventional methods.

The object detection device (8) according to claim 1 is configured to be mounted on a host vehicle (V) to detect an object (Vt) existing around the host vehicle.
This object detection device is
a ranging point acquisition unit (803) that obtains a detection result of the ranging point using a ranging sensor (2) for detecting ranging points (Ps) around the own vehicle;
a sensing surface acquisition unit (805) that acquires a sensing surface (Ft) that is a detection result of the object surface using a surface detection sensor (3) for detecting an object surface that is the outer surface of the object;
an object detection unit (808) that detects the object based on the distance measurement point selected according to the positional relationship with the detection surface in the space around the host vehicle;
Equipped with
The object detection unit selects the ranging point used for detecting the object based on a high-precision surface (FtH), which is the detection surface with a predetermined high detection accuracy.
The object detection program according to claim 6 is executed by an object detection device (8) configured to be mounted on the own vehicle (V) to detect objects (Vt) existing around the own vehicle. A program that
The process executed by the object detection device includes:
A process of acquiring a detection result of the distance measurement point using a distance measurement sensor (2) for detecting the distance measurement point (Ps) around the own vehicle;
A process of acquiring a detection surface (Ft) that is a detection result of the object surface using a surface detection sensor (3) for detecting an object surface that is the outer surface of the object;
a process of detecting the object based on the distance measurement point selected according to the positional relationship with the detection surface in the space surrounding the host vehicle;
including,
The process of detecting the object includes a process of selecting the ranging point used for detecting the object based on a high-precision surface (FtH), which is the detection surface with a predetermined high detection accuracy.

Note that in the application documents, each element may be given a reference sign in parentheses. However, even in this case, such reference numerals merely indicate one example of the correspondence between each element and specific means described in the embodiments described later. Therefore, the present invention is not limited in any way by the description of the above reference numerals.

FIG. 3 is a schematic plan view showing an outline of a parking assistance operation in the host vehicle. 2 is a block diagram showing a schematic functional configuration of the in-vehicle system shown in FIG. 1. FIG. 3 is a schematic plan view for explaining an example of the operation of the in-vehicle system shown in FIG. 2. FIG. 3 is a schematic plan view for explaining an example of the operation of the in-vehicle system shown in FIG. 2. FIG. 3 is a schematic plan view for explaining an example of the operation of the in-vehicle system shown in FIG. 2. FIG. 3 is a flowchart showing an example of the operation of the parking assist ECU shown in FIG. 2. FIG.

(Embodiment)
Embodiments of the present invention will be described below based on the drawings. Note that if various modifications applicable to one embodiment are inserted in the middle of a series of explanations regarding the embodiment, understanding of the embodiment may be hindered. For this reason, the modified examples will be explained together after the series of explanations regarding the embodiment, rather than during the series of explanations.

(composition)
Referring to FIG. 1, the host vehicle V is a so-called four-wheeled vehicle and has a generally rectangular body in plan view. "Planar view" refers to the appearance of an object when viewed from above with a line of sight that coincides with the direction of gravity. The own vehicle V is equipped with an in-vehicle system 1. The in-vehicle system 1 is configured to be able to execute various controls (for example, parking assistance control) on the host vehicle V. Specifically, the in-vehicle system 1 is configured to detect objects existing around the own vehicle V, including other vehicles Vt, based on the outputs of the distance measurement sensor 2 and the surface detection sensor 3.

The distance measurement sensor 2 is a sensor (for example, an ultrasonic sensor or a millimeter wave radar sensor) for detecting a distance measurement point Ps around the own vehicle V, and is a sensor that transmits a search wave and detects the detection of the search wave by an object. The device is configured to receive reflected waves. That is, the distance measurement sensor 2 has a configuration as a so-called TOF type sensor that measures the distance to the object corresponding to the distance measurement point Ps based on the propagation time of the exploration wave and the reflected wave. TOF is an abbreviation for Time of Flight. The “distance measurement point” is a point on the object plane Fb, which is the outer surface of the object, that is estimated to have reflected the exploration wave emitted from the distance measurement sensor 2. This is the point on the half line passing through the axis center of the detection range, where the distance is the distance measurement distance. The ranging distance is the distance between the ranging sensor 2 and the ranging point Ps, which is calculated based on the propagation time.

A distance measuring sensor 2 for detecting a parking space SP laterally while the own vehicle V is traveling forward is mounted on the side surface of the own vehicle V so as to transmit a search wave laterally. Specifically, the distance measuring sensor 2 is provided at each of the right front part, right rear part, left front part, and left rear part of the vehicle body of the own vehicle V. Note that the own vehicle V may also be equipped with a distance measuring sensor 2 for detecting obstacles in front of the own vehicle V, on the front side, on the rear side, and on the rear side. However, unless otherwise specified, the ranging sensor 2 in the following description is assumed to be attached to the side of the vehicle body in order to detect the parking space SP on the side while the own vehicle V is traveling forward. .

The surface detection sensor 3 is a sensor for detecting the object surface Fb. That is, the in-vehicle system 1 is configured to use the surface detection sensor 3 to detect the detection surface Ft corresponding to the object surface Fb. In this embodiment, the surface detection sensor 3 has a configuration as an optical sensor (for example, a camera or a LIDAR sensor) that optically detects an object within a predetermined field of view. LIDAR is an abbreviation for Light Detection and Ranging or Laser Imaging Detection and Ranging. That is, the surface detection sensor 3 is provided so as to be able to detect an object as a surface within a wider field of view than the detection range of the ranging sensor 2 indicated by the dashed line in the figure. Therefore, the surface detection sensor 3 may also be referred to as a "wide-angle sensor." A "field of view" may also be referred to as a sensing range or a scanning range.

A surface detection sensor 3 for detecting a parking space SP on the side while the host vehicle V is moving forward is arranged so that an object present on the side of the host vehicle V is included in the field of view. Specifically, one surface detection sensor 3 is provided at a predetermined location (for example, in or near a door mirror) on both sides of the vehicle body. The surface detection sensor 3 is installed so that its center line Lc is substantially parallel to the central axis of the detection range of the surface detection sensor 3 in plan view. The center line Lc is a half straight line passing through the axial center of the detection range by the surface detection sensor 3, and corresponds to the optical axis when the surface detection sensor 3 is a camera.

The object plane Fb includes a facing surface Fbf and a non-facing surface Fbs. As shown in FIG. 1, when the own vehicle V is searching for a parking space SP among a plurality of other vehicles Vt parked in parallel, the facing surface Fbf is the side surface of the body of the other vehicle Vt. The non-opposed surface Fbs is the front or rear surface of the vehicle body of the other vehicle Vt.

The opposing surface Fbf is an object surface Fb that faces the host vehicle V or the path in which the host vehicle V is traveling. Specifically, the facing surface Fbf has an outward normal line that is the same as the above-mentioned outward normal line and the center line Lc in a direct view state where the own vehicle V passes the surface detection sensor 3 while the host vehicle V is traveling forward in a plan view. This is an object plane Fb whose azimuth angle is within the viewing angle of the surface detection sensor 3. Typically, the opposing surface Fbf is an object surface Fb whose azimuth angle is 0 to α degrees. α is, for example, 45-60. The non-opposing surface Fbs is an object surface Fb on which the above-mentioned direct viewing condition does not hold, or even if it does, the above-mentioned azimuth angle exceeds the viewing angle of the surface detection sensor 3. Typically, the non-opposing surface Fbs is an object surface Fb that is substantially orthogonal to the opposing surface Fbf.

Referring to FIG. 2, the in-vehicle system 1 includes, in addition to the distance measurement sensor 2 and surface detection sensor 3 described above, a behavior sensor 4, a locator 5, a DCM 6, a user interface device 7, a parking assist ECU 8, and a vehicle It includes a control system 9 and an in-vehicle network 10. DCM is an abbreviation for Data Communication Module. ECU is an abbreviation for Electronic Control Unit.

The behavior sensor 4 is provided to output behavior information, that is, information or signals corresponding to the driving state or behavior of the own vehicle V to the parking assist ECU 8 and the vehicle control system 9. That is, the behavior sensor 4 is a general term for various sensors such as a shift position sensor, an accelerator opening sensor, a vehicle speed sensor, a steering angle sensor, an acceleration sensor, and a yaw rate sensor.

The locator 5 is configured to obtain highly accurate position information of the own vehicle V through so-called composite positioning. "Highly accurate location information" is, for example, location accuracy that can be used for advanced driving automation levels of levels 2 to 5 in "SAE J3016", specifically, location accuracy that has an error of less than 10 cm. This is location information with . SAE stands for Society of Automotive Engineers. Levels 2, 3, 4, and 5 in "SAE J3016" are referred to as "highly automated driving," "conditionally automated driving," "highly automated driving," and "fully automated driving," respectively.

The DCM 6 is an in-vehicle communication module, and is installed to be able to communicate information with base stations around the own vehicle V by wireless communication compliant with communication standards such as LTE or 5G. LTE is an abbreviation for Long Term Evolution. 5G is an abbreviation for 5th Generation. Specifically, for example, the DCM 6 is configured to acquire the latest high-precision map information from a probe server provided on a cloud (not shown).

The high-precision map information includes map information that is more accurate than the map information used in conventional car navigation systems, which can accommodate positional errors of about several meters. Specifically, high-precision map information includes three-dimensional road shape information, lane number information, regulation information, etc., in accordance with predetermined standards such as the ADASIS standard, and is required for driving at level 2 or higher in "SAE J3016". Information available at the automation level is stored. ADASIS stands for Advanced Driver Assistance Systems Interface Specification.

The user interface device 7 is provided to be capable of receiving various inputs from occupants of the own vehicle V, including the driver, and outputting information for presenting various information to the occupants. Specifically, the user interface device 7 includes, for example, an input device (such as a touch panel), a display device, a speaker, a microphone, and the like.

The parking assist ECU 8 is configured as a so-called in-vehicle microcomputer that includes a CPU, ROM, RAM, nonvolatile rewritable memory, input/output interface, etc. (not shown). CPU is an abbreviation for Central Processing Unit. ROM is an abbreviation for Read Only Memory. RAM is an abbreviation for Random Access Memory. Examples of the nonvolatile rewritable memory include a hard disk, EEPROM, and flash ROM. EEPROM is an abbreviation for Electronically Erasable and Programmable ROM. ROM and non-volatile rewritable memory represent computer-readable non-transitory tangible storage media.

The parking assist ECU 8 is configured such that the CPU can perform various control operations by reading and executing programs from the ROM or nonvolatile rewritable memory. This program includes those corresponding to the flowcharts or routines described below. Further, the RAM and the nonvolatile rewritable memory are configured to be able to temporarily store processing data when the CPU executes a program. Further, the ROM and/or nonvolatile rewritable memory prestores various data used when executing the program. Such various data include, for example, initial values, lookup tables, maps, and the like.

The parking assistance ECU 8 as an object detection device of the present invention is installed in the own vehicle V to detect objects such as other vehicles Vt existing around the own vehicle V, and also provides parking assistance according to the object detection result. Configured to perform control. Specifically, the parking assist ECU 8 has the following functional configuration, which is realized on an on-vehicle microcomputer.

That is, the parking assistance ECU 8 includes a behavior acquisition section 801 and a position estimation section 802. Furthermore, the parking assist ECU 8 includes a distance measurement point acquisition section 803, a detection point acquisition section 804, a detection surface acquisition section 805, and a high precision surface acquisition section 806. Furthermore, the parking assistance ECU 8 includes a map generation section 807, an object detection section 808, a parking space extraction section 809, and a parking route calculation section 810.

3 to 5 show an overview of the parking assistance operation by the parking assistance ECU 8, particularly the object detection operation for detecting the parking space SP. Each of the above functional configurations of the parking assist ECU 8 will be described below with reference to FIGS. 1 to 5.

The behavior acquisition unit 801 is configured to acquire behavior information necessary for estimating the position of the own vehicle V from the start of the parking assistance operation. That is, the behavior acquisition unit 801 receives behavior information from the behavior sensor 4 and temporarily stores it in the RAM and/or nonvolatile rewritable memory. The position estimating unit 802 estimates, that is, calculates, the position of the own vehicle V in a predetermined XY coordinate system based on the behavior information acquired by the behavior acquiring unit 801. The "predetermined XY coordinate system" is a two-dimensional coordinate system within a plane orthogonal to the vehicle height direction of the host vehicle V, which is set at the time of starting the parking assistance operation. Such a predetermined XY coordinate system will be simply referred to as a "translational coordinate system" hereinafter.

The distance measurement point acquisition unit 803 is configured to acquire the detection result of the distance measurement point Ps using the distance measurement sensor 2. Specifically, the distance measurement point acquisition unit 803 calculates the translation based on the position of the own vehicle V estimated by the position estimation unit 802 and the distance measurement information included in the output signal received from the distance measurement sensor 2. The position of the ranging point Ps in the coordinate system is calculated. Further, the ranging point acquisition unit 803 retains a predetermined number or a predetermined period of the calculated locations of the ranging points Ps in time series, that is, temporarily stores them in the RAM and/or non-volatile rewritable memory. .

The detection point acquisition unit 804 acquires the detection result of the detection point Pd using the surface detection sensor 3, as shown in FIG. The detection point Pd is a point detected on the object plane Fb, which is the outer surface of an object existing within the field of view of the surface detection sensor 3. Specifically, for example, when the surface detection sensor 3 is a camera, the detection point Pd is a so-called feature point. Alternatively, for example, when the surface detection sensor 3 is a LIDAR sensor, the detection point Pd is a data point included in point cloud data constituted by three-dimensional coordinates, and may also be referred to as a "ranging point." . Specifically, in this embodiment, the detection point acquisition unit 804 receives from the surface detection sensor 3 the extraction result of the detection point Pd by the sensor ECU built in the surface detection sensor 3, and stores it in the RAM and the /or temporarily stored in non-volatile rewritable memory.

The detection surface acquisition unit 805 is configured to acquire the detection result of the detection surface Ft using the surface detection sensor 3, as shown in FIG. The detection surface Ft is a plane or a curved surface formed by the detection points Pd, and corresponds to the detection result of the object surface Fb. Specifically, in the present embodiment, the sensing surface acquisition unit 805 acquires, that is, calculates or estimates the sensing surface Ft, based on the collective state of the plurality of sensing points Pd acquired using the surface detection sensor 3. It looks like this.

The high-precision surface acquisition unit 806 is configured to obtain a high-precision surface FtH, which is a detection surface Ft with high detection accuracy, that is, high positional accuracy in the translational coordinate system. The high-precision surface FtH is a detection surface Ft whose detection accuracy is assumed to be a predetermined high precision. Specifically, in the present embodiment, the high-precision surface acquisition unit 806 determines the arrangement relationship between the surface detection sensor 3 and the detection surface Ft in the space around the own vehicle V from at least one detected detection surface Ft. Based on this, a high-precision surface FtH is obtained. Details of obtaining the high-precision surface FtH will be described later.

The map generation unit 807 is configured to generate a map M indicating the probability of existence of an object in the space around the own vehicle V. Map M may also be referred to as an "existence probability map." Specifically, the map generation unit 807 generates a map M based on predetermined allocation conditions for each grid G in a map M formed by dividing the translational coordinate system into a plurality of grids G arranged two-dimensionally. The probability of the object's existence is then assigned. The "predetermined allocation condition" includes, for example, at least one of the acquisition results of the distance measurement point Ps, the detection point Pd, the detection surface Ft, and the high-precision surface FtH. In this embodiment, the "predetermined allocation condition" includes at least the acquisition result of the high-precision surface FtH. More specifically, in this embodiment, the map generation unit 807 sets the existence probability in the detection point grid Gn to be high and the existence probability in the out-of-plane grid Gt to be low, as shown in FIG. ing.

The detection point grid Gn is a grid G that includes distance measurement points Ps or detection points Pd. In FIGS. 3 to 5, a detection point grid Gn formed by detection points Pd is indicated by diagonal hatching in the figures. The out-of-plane grid Gt is a grid G that is determined to be located "outside" of the on-plane grid Gp based on the arrangement state of the detection point Pd and the high-precision surface FtH. The "outside" refers to the side that is separated from the object corresponding to the in-plane grid Gp. In FIG. 5, the out-of-plane grid Gt is shown by cross hatching in the figure. The surface grid Gp is a grid G through which the high-precision surface FtH passes, and is typically a detection point grid Gn through which the high-precision surface FtH passes. That is, the map generation unit 807 sets, in the map M, an out-of-plane grid Gt corresponding to a low-probability region whose existence probability is lower than other regions, based on the high-precision surface FtH.

The object detection unit 808 is configured to detect objects around the host vehicle V based on the distance measurement point Ps. Specifically, the object detection unit 808 selects a ranging point Ps used for detecting objects around the own vehicle V based on the map M, and detects the object in the translational coordinate system based on the selected ranging point Ps. The position or external shape of the object is estimated. "Distance measuring point Ps used for detecting objects around host vehicle V" is hereinafter referred to as "distance measuring point Ps for object detection." That is, the object detection unit 808 selects the distance measurement point Ps corresponding to the grid G whose existence probability is higher than a predetermined degree as the distance measurement point Ps for object detection.

In this embodiment, the object detection unit 808 selects the distance measurement point Ps for detecting an object according to the positional relationship with the detection surface Ft in the space around the own vehicle V. Specifically, the object detection unit 808 selects the distance measurement point Ps for object detection based on the high-precision surface FtH. More specifically, the object detection unit 808 excludes the distance measurement point Ps corresponding to the out-of-plane grid Gt from the distance measurement points Ps for object detection. Further, the object detection unit 808 selects a distance measurement point Ps for object detection based on the detection surface Ft detected by the surface detection sensor 3 during normal operation.

The parking space extraction unit 809 extracts a parking space SP based on the object detection result by the object detection unit 808. The parking route calculation unit 810 calculates a parking route, which is a travel route for parking the host vehicle V in the parking space SP, based on the parking space SP extracted by the parking space extraction unit 809.

The vehicle control system 9 controls the operation of the host vehicle V, such as a drive system including a drive mechanism and a drive control ECU, a brake system including a brake mechanism and a brake control ECU, and a steering system including a steering mechanism and a steering control ECU. It has a configuration for executing control. That is, the vehicle control system 9 performs driving control such as acceleration/deceleration, steering, and braking necessary for the own vehicle V to park in the parking space SP based on the parking space SP and parking route acquired by the parking support ECU 8. is configured to run.

The distance sensor 2, the surface detection sensor 3, the behavior sensor 4, the locator 5, the DCM 6, the user interface device 7, the parking assist ECU 8, and the vehicle control system 9 are connected to each other via the in-vehicle network 10 so that they can send and receive signals or information. has been done. The in-vehicle network 10 is configured to comply with predetermined standards such as CAN (international registered trademark: international registration number 1048262A), FlexRay (international registered trademark), and LIN. CAN (International Registered Trademark) is an abbreviation for Controller Area Network. LIN is an abbreviation for Local Interconnect Network.

(Operation overview)
An outline of the operation of the parking assist ECU 8 will be described below. The parking assistance ECU 8 starts the parking assistance operation based on a predetermined command input for starting the parking assistance operation. Such command input is, for example, an input operation of the user interface device 7 by the driver of the own vehicle V. Alternatively, such command input is, for example, input from the locator 5 of information indicating that the own vehicle V has entered the parking lot.

In response to a predetermined command input, the CPU in the parking assistance ECU 8 reads out and executes a program for parking assistance operation from the ROM or nonvolatile storage medium. Such a program corresponds to an object detection program according to the present invention. Further, by executing this program, the object detection method according to the present invention is implemented. The parking assist operation start point is the start point of execution of the program, specifically, the point of input of the above-mentioned command. When such a program is executed, each functional block configuration in the parking assist ECU 8 operates as follows.

The map generation unit 807 sets the predetermined position at the start of the parking assistance operation (for example, the center position of the own vehicle V in a plan view) as the origin, sets the front of the own vehicle V in the positive direction of the X-axis, and moves to the right or left of the own vehicle V. Set up a translational coordinate system with the positive Y-axis direction. Furthermore, the map generation unit 807 sets a plurality of grids G arranged two-dimensionally in a lattice shape in a translational coordinate system. Furthermore, the map generation unit 807 sets the existence probability assigned to each grid G to an initial value. The existence probability increases as the value increases, and decreases as the value decreases. In this embodiment, the initial value is "0", which indicates the presence or absence of the object is unknown, and the maximum value is +Ω. The minimum value shall be -Ω. Ω is a natural number (for example, 255).

The position estimation unit 802 estimates the position of the host vehicle V in the translational coordinate system based on the behavior information acquired by the behavior acquisition unit 801. The distance measurement point acquisition unit 803 acquires the detection result of the distance measurement point Ps using the distance measurement sensor 2. Specifically, the distance measurement point acquisition unit 803 performs distance measurement in the translational coordinate system based on the position of the host vehicle V estimated by the position estimation unit 802 and the distance measurement information received from the distance measurement sensor 2. Calculate the position of point Ps. Further, the ranging point acquisition unit 803 holds the calculated locations of the ranging points Ps for a predetermined number or a predetermined period in time series.

The detection point acquisition unit 804 acquires the detection result of the detection point Pd using the surface detection sensor 3. When the surface detection sensor 3 is a camera, the detection point Pd is a feature point. A feature point is a point that characterizes the shape of an object in a captured image. Specifically, a feature point is a characteristic point, ie, a pixel, within the angle of view of a photographed image, ie, within an image frame. For example, a feature point is a pixel that has a large luminance change between adjacent pixels.

The sensing surface acquisition unit 805 acquires the sensing surface Ft based on the gathering state of the sensing points Pd, that is, the number and arrangement state (for example, length). Specifically, the detection surface acquisition unit 805 recognizes a point group consisting of a plurality of detection points Pd, which are N or more and have a length D or more, as the detection surface Ft. For example, the detection surface acquisition unit 805 extracts a point group consisting of i detection points Pd in which the distance between adjacent points in the translational coordinate system is within a predetermined distance. If i≧N, the detection surface acquisition unit 805 calculates an approximate curve that fits the i points using the method of least squares or the like. Then, when the length of the approximate curve is greater than or equal to D, the detection surface acquisition unit 805 sets the approximate curve as the detection surface Ft.

The high-precision surface acquisition unit 806 obtains a high-precision surface FtH, which is a detection surface Ft with high detection accuracy, that is, high positional accuracy in the translational coordinate system. Specifically, in this embodiment, the high-precision surface acquisition unit 806 obtains the high-precision surface FtH based on the arrangement relationship between the surface detection sensor 3 and the detection surface Ft in the translational coordinate system. Hereinafter, details of acquiring the high-precision surface FtH when the surface detection sensor 3 is a camera will be described.

The azimuth accuracy of the detection point Pd by the surface detection sensor 3, which is a camera, depends on the resolution of an image sensor such as a CMOS. CMOS is an abbreviation for complementary MOS. Recent image sensors have extremely high resolution (for example, about 25 pixels/deg). Therefore, the azimuth accuracy of the detection point Pd is high.

On the other hand, the distance from the surface detection sensor 3 to the detection point Pd is calculated based on the principle of moving stereo using a plurality of image data taken at different times. At this time, a distance error of the detection point Pd is generated by amplifying the position error of the detection point Pd on the image. Therefore, the distance accuracy of the detection point Pd is lower than the azimuth accuracy.

3 to 5 show specific examples of parking support operations when the own vehicle V parallel parks. At this time, referring to FIG. 4, the first detection surface Ft1 corresponding to the non-facing surface Fbs, which is the front surface of the vehicle body of the other vehicle Vt, which is the parked vehicle, extends in the depth direction when viewed from the surface detection sensor 3. Therefore, while distance accuracy is not an issue, as mentioned above, azimuth accuracy is high. On the other hand, since the second detection surface Ft2 corresponding to the facing surface Fbf, which is the side surface of the vehicle body of the other vehicle Vt, extends along the traveling direction of the own vehicle V, the azimuth accuracy is not a problem. ,As mentioned above, the distance accuracy is low.

Therefore, in the examples of FIGS. 3 to 5, the high-precision surface acquisition unit 806 selects the first detection surface Ft1 with high azimuth accuracy as the high-precision surface FtH. Specifically, the high-precision surface acquisition unit 806 selects, as the high-precision surface FtH, a detection surface Ft whose angle between the detection surface Ft and the center line Lc in the translational coordinate system is less than or equal to a predetermined threshold angle. do.

Alternatively, the high-precision surface acquisition unit 806 may detect a detection surface where the shortest distance between the extension line of the detection surface Ft and the surface detection sensor 3 at the time of acquisition of the detection surface Ft in the translational coordinate system is equal to or less than a predetermined threshold distance. Ft is selected as the high precision surface FtH. In the example shown in FIG. 4, the first extension line L1 corresponding to the first detection surface Ft1 passes near the surface detection sensor 3, and the shortest distance is equal to or less than the threshold distance. On the other hand, the second extension line L2 corresponding to the second detection surface Ft2 passes far from the surface detection sensor 3, and the shortest distance exceeds the threshold distance. Therefore, the high-precision surface acquisition unit 806 selects the first detection surface Ft1 as the high-precision surface FtH.

The map generation unit 807 assigns an object existence probability to each grid G in the map M based on predetermined assignment conditions. Specifically, the map generation unit 807 adds a predetermined positive value to the existence probability in the detection point grid Gn where the ranging point Ps and the detection point Pd exist, for example.

Furthermore, the map generation unit 807 sets an out-of-plane grid Gt corresponding to a low-probability region whose existence probability is lower than other regions in the map M based on the acquired or selected high-precision surface FtH. Specifically, the map generation unit 807 sets the existence probability in the out-of-plane grid Gt to a predetermined negative value corresponding to the absence of the object, that is, a negative value with a large absolute value. In this embodiment, if the detection point grid Gn in which the distance measurement point Ps exists also corresponds to the out-of-plane grid Gt, priority is given to processing for the out-of-plane grid Gt.

By the way, the detection accuracy of the surface detection sensor 3, which is a camera or a LIDAR sensor, decreases when adverse conditions such as rain or fog occur. Alternatively, some abnormality may occur in the surface detection sensor 3. In these cases, it is difficult to satisfactorily obtain a high-precision surface FtH. Therefore, the map generation unit 807 obtains a high-precision surface FtH based on the detection surface Ft detected by the surface detection sensor 3 during normal operation.

The object detection unit 808 selects a distance measurement point Ps for object detection based on the map M. That is, the object detection unit 808 selects the ranging point Ps included in the grid G with a high existence probability as the ranging point Ps for object detection. Then, the object detection unit 808 determines the position of an object (for example, another vehicle Vt shown in FIGS. 3 to 5) around the own vehicle V in the translational coordinate system based on the selected ranging point Ps. Alternatively, the object is detected by estimating its external shape. The parking space extraction unit 809 extracts a parking space SP based on the object detection result by the object detection unit 808. The parking route calculation unit 810 calculates a parking route corresponding to the parking space SP extracted by the parking space extraction unit 809.

(Operation example)
Hereinafter, a specific example of operation according to the configuration of this embodiment will be described using the flowchart shown in FIG. Note that in FIG. 6, "step" is simply abbreviated as "S".

The parking assist ECU 8 repeatedly activates the routine shown in FIG. 6 at predetermined time intervals while a predetermined activation condition is satisfied. When this routine is activated, the parking assist ECU 8 first sequentially executes steps 601 to 605.

In step 601, parking assistance ECU 8 acquires behavior information of own vehicle V. At step 602, parking assist ECU 8 estimates the position of host vehicle V in the translational coordinate system. In step 603, the parking support ECU 8 acquires the distance measurement point Ps. At step 604, the parking assistance ECU 8 revises the map M based on the distance measurement point Ps acquired at step 603. Specifically, the parking assistance ECU 8 adds a predetermined positive value to the existence probability in the detection point grid Gn corresponding to the distance measurement point Ps.

At step 605, the parking assist ECU 8 determines whether the surface detection sensor 3 is operating normally. Specifically, for example, the parking assist ECU 8 determines whether the surface detection sensor 3 is undergoing self-diagnosis. Alternatively, the parking assist ECU 8 receives the self-diagnosis result of the surface detection sensor 3 from the surface detection sensor 3 . Alternatively, for example, the parking assist ECU 8 receives the raindrop detection result and the fog detection result from the behavior sensor 4 or an external device. Such an external device is, for example, the vehicle control system 9 or a raindrop detection sensor (not shown). Alternatively, for example, the parking assist ECU 8 receives the operating states of wipers and fog lamps. Then, the parking assist ECU 8 determines whether the surface detection sensor 3 is operating normally or not based on the above-mentioned received information or signal.

If the surface detection sensor 3 is operating normally (ie, step 605=YES), the parking assist ECU 8 advances the process to steps 606 to 609. In step 606, the parking assist ECU 8 obtains the detection result of the detection point Pd using the surface detection sensor 3. In step 607, parking assist ECU 8 obtains a detection surface Ft based on the detection point Pd obtained in step 606. In step 608, the parking assist ECU 8 obtains a high-precision surface FtH based on the detection surface Ft obtained in step 607.

In step 609, the parking assist ECU 8 determines whether or not the high-precision surface FtH has been acquired through the process in step 608. If the high-precision surface FtH has been acquired (that is, step 609=YES), the parking assist ECU 8 advances the process to step 610. On the other hand, if the high-precision surface FtH cannot be acquired (that is, step 609=NO), the parking assist ECU 8 advances the process to step 611.

At step 610, parking assist ECU 8 revise map M based on detection point Pd and high precision surface FtH. Specifically, the parking assistance ECU 8 adds a predetermined positive value to the existence probability in the detection point grid Gn corresponding to the detection point Pd. Furthermore, the parking assist ECU 8 sets the existence probability in the out-of-plane grid Gt, which is specified or set based on the high-precision surface FtH, to a predetermined negative value corresponding to the absence of the object.

In step 611, parking assist ECU 8 revise map M based on detection point Pd. Specifically, the parking assistance ECU 8 adds a predetermined positive value to the existence probability in the detection point grid Gn corresponding to the detection point Pd.

After the process in step 610 or step 611, the parking assist ECU 8 advances the process to step 612. Further, if the surface detection sensor 3 is not operating normally (ie, step 605=NO), the parking assist ECU 8 skips the processing from step 606 to step 611 and advances the processing from step 605 to step 612.

At step 612, the parking support ECU 8 selects a distance measurement point Ps for detecting an object based on the map M. In step 613, the parking assistance ECU 8 estimates the position of objects existing around the host vehicle V based on the selected ranging point Ps. That is, the parking assistance ECU 8 detects or recognizes objects existing around the host vehicle V based on the selected ranging point Ps. After that, the parking assist ECU 8 temporarily ends this routine.

(effect)
According to the parking assist ECU 8 according to the present embodiment, and the object detection method and object detection program executed thereby, the following effects can be achieved. Hereinafter, the parking assist ECU 8 according to the present embodiment, as well as the object detection method and object detection program executed thereby, will be collectively referred to as simply "the present embodiment."

(1) In the present embodiment, the distance measurement point acquisition unit 803 acquires the detection result of the distance measurement point Ps using the distance measurement sensor 2 for detecting the distance measurement point Ps around the host vehicle V. . The detection plane acquisition unit 805 acquires a detection plane Ft that is a detection result of the object plane Fb using the surface detection sensor 3 for detecting the object plane Fb. The object detection unit 808 detects an object based on the ranging point Ps selected according to the positional relationship with the detection surface Ft in the space around the own vehicle V.

According to the present embodiment, even if an "outlier point" with a large positional error occurs among a plurality of distance measurement points Ps by the distance measurement sensor 2, the distance measurement point Ps that does not correspond to such an "outlier point" is selected from the detection surface. It can be selected depending on the positional relationship with Ft and used for object detection. Specifically, for example, by identifying an "outlier point" according to the positional relationship with the detection surface Ft in the space around the own vehicle V, the identified "outlier point" is used as the distance measurement point Ps for object detection. can be excluded from.

In this way, according to the present embodiment, it is possible to satisfactorily suppress the erroneous detection of an obstacle due to an "outlier point" and the occurrence of erroneous determinations associated with such erroneous detection. Therefore, according to the present embodiment, it is possible to provide an object detection device, an object detection method, and an object detection program that can achieve better object detection accuracy than conventional ones.

(2) In the present embodiment, the object detection unit 808 selects the distance measurement point Ps for object detection based on the detection surface Ft detected by the surface detection sensor 3 during normal operation. According to the present embodiment, by using the acquisition result of the detection surface Ft in a situation where the detection surface Ft can be detected satisfactorily, it is possible to select the distance measurement point Ps for detecting an object favorably.

(3) In the present embodiment, the sensing surface acquisition unit 805 acquires the sensing surface Ft based on the collective state of the plurality of sensing points Pd on the object surface Fb acquired using the surface detection sensor 3. do. According to the surface detection sensor 3, a large number of detection points Pd can be acquired almost simultaneously or in a short time. If a predetermined number of point groups exist within a certain width and depth range among the large number of detection points Pd acquired in this way, it can be said that there is a high probability that the point group corresponds to the object plane Fb. Therefore, according to the present embodiment, it is possible to obtain the detection surface Ft with good accuracy through simple processing.

(4) In the present embodiment, the object detection unit 808 selects the distance measurement point Ps for object detection based on the high-precision surface FtH, which is the detection surface Ft with a predetermined high detection accuracy. According to the present embodiment, even if an "outlier point" with a large positional error occurs among the plurality of distance measurement points Ps by the distance measurement sensor 2, the distance measurement point Ps that does not correspond to the "outlier point" is selected with high precision. It can be selected based on the plane FtH and used for object detection. Therefore, according to the present embodiment, it is possible to select distance measurement points Ps that do not correspond to "outlier points" with high accuracy.

(5) The present embodiment includes a map generation unit 807 that generates a map M indicating the probability of existence of an object in the space around the own vehicle V. The map generation unit 807 sets, in the map M, a low probability region, that is, an out-of-plane grid Gt, whose existence probability is lower than other regions, based on the high-precision surface FtH. The object detection unit 808 excludes the ranging point Ps corresponding to this low probability area from the ranging points Ps for object detection.

In this embodiment, by excluding the ranging points Ps corresponding to the low probability area set in the map M, that is, the out-of-plane grid Gt, as "outlying points", it is possible to select the ranging points Ps for object detection. can. Therefore, according to the present embodiment, it is possible to select the ranging point Ps for detecting an object with high accuracy.

(6) In the present embodiment, the high-precision surface acquisition unit 806 obtains the high-precision surface FtH based on the arrangement relationship between the surface detection sensor 3 and the detection surface Ft in the space around the own vehicle V. Therefore, according to this embodiment, it is possible to obtain the high-precision surface FtH with good accuracy.

(Modified example)
The present invention is not limited to the above embodiments. Therefore, the above embodiment can be modified as appropriate. Typical modified examples will be described below. In the following description of the modified example, differences from the above embodiment will be mainly described. Further, in the above embodiment and the modified example, parts that are the same or equivalent to each other are given the same reference numerals. Therefore, in the following description of the modification, the description in the above embodiment may be used as appropriate for components having the same reference numerals as those in the above embodiment, unless there is a technical contradiction or special additional explanation.

The present invention is not limited to the specific device configuration shown in the above embodiments. That is, for example, the subject vehicle V to which the present invention is applied is not limited to a four-wheel vehicle. Specifically, the own vehicle V may be a three-wheeled vehicle, or a six-wheeled or eight-wheeled vehicle such as a cargo truck. The type of vehicle V may be a car equipped only with an internal combustion engine, an electric car or a fuel cell car without an internal combustion engine, or a so-called hybrid car. The shape and structure of the vehicle body are also not limited to a box shape, that is, a substantially rectangular shape in plan view.

In the embodiment described above, the parking assist ECU 8 has a configuration as a so-called in-vehicle microcomputer in which the CPU reads a program from a ROM or the like and starts the program. However, the present invention is not limited to such a configuration. Specifically, for example, all or part of the parking assist ECU 8 may be configured to include a digital circuit configured to enable the above operations, such as an ASIC or FPGA. ASIC is an abbreviation for Application Specific Integrated Circuit. FPGA is an abbreviation for Field Programmable Gate Array. Therefore, in the parking assist ECU 8, the in-vehicle microcomputer portion and the digital circuit portion may coexist.

The program according to the present invention, which enables execution of the various operations, procedures, or processes described in the above embodiments, can be downloaded or upgraded via V2X communication using the DCM 6 or the like. V2X is an abbreviation for Vehicle to X. Alternatively, such a program may be downloaded or upgraded via a terminal device provided at a manufacturing factory, maintenance shop, dealership, etc. of the own vehicle V. Such a program may be stored in a memory card, an optical disk, a magnetic disk, or the like.

As such, each of the functional configurations and methods described above may be implemented by a dedicated computer including a processor and memory programmed to perform one or more of the functions embodied by the computer program. Alternatively, each of the functional configurations and methods described above may be implemented by dedicated hardware provided by one or more dedicated logic circuits.

Alternatively, each of the above functional configurations and methods is configured by a combination of a processor and memory programmed to perform one or more functions, and hardware configured by one or more logic circuits. It may be implemented by one or more dedicated processing devices.

A computer program may be stored as instructions executed by a computer on a computer-readable non-transitory tangible storage medium. That is, each of the functional configurations and methods described above can be expressed as a computer program including processes or procedures for realizing the same, or as a non-transitional physical storage medium storing the program.

In the embodiment described above, the object detection device according to the present invention is realized as a parking assistance ECU 8 that is a parking assistance control device. However, the present invention is not limited to such embodiments. That is, for example, the behavior acquisition unit 801 to the object detection unit 808 may be realized by a driving support ECU or an automatic driving ECU for realizing driving automation levels of levels 1 to 5 in “SAE J3016”. Level 1 "driving support" includes collision mitigation braking, adaptive cruise control, lane keep assist, etc.

The present invention is not limited to the specific operational aspects or processing aspects shown in the above embodiments. That is, for example, without using the map M, the distance measurement point Ps for object detection can be selected based on the relative positional relationship in the translational coordinate system between the detection point Pd, the high-precision surface FtH, and the distance measurement point Ps. is possible. Specifically, based on this relative positional relationship, it is possible to specify the distance measuring point Ps that is determined to be located on the "outside" of the high-precision surface FtH and exclude it from the distance measuring points Ps for object detection. It is. In this case, the map generation unit 807 is omitted.

The distance measurement point Ps is determined by the position and distance measurement of the distance measurement sensor 2 at a certain time t(n), and the position and measurement distance of the distance measurement sensor 2 at a time t(n-1) immediately before the time t(n). The point may be calculated based on the measured distance using the principle of triangulation. That is, the "distance measuring point Ps" may also be referred to as the "reflection point Ps", the "distance measuring sensor detection point Ps", or the "first detection point Ps". In this case, the "detection point Pd" may also be referred to as the "surface detection sensor detection point Pd" or the "second detection point Pd."

Machine learning may be used to obtain the detection surface Ft. Specifically, pattern matching, semantic segmentation, etc. can be used. Thereby, it becomes possible to obtain the detection surface Ft with higher precision or higher speed.

The flowchart shown in FIG. 6 can be modified as appropriate. Specifically, for example, step 604 may be omitted. That is, the ranging point Ps does not need to be reflected in the existence probability in the map M.

The processing for the existence probability in the out-of-plane grid Gt is not limited to setting it to a predetermined negative value corresponding to the absence of an object. That is, for example, such processing may be addition of a predetermined negative value.

The out-of-plane grid Gt is not limited to the grid G outside the high-precision surface FtH. That is, for example, depending on the type or structure of the surface detection sensor 3, all the detection surfaces Ft may be acquired with high accuracy. Specifically, for example, when the surface detection sensor 3 is a compound eye stereo camera device or a fusion sensor including a camera and other sensors, all the detection surfaces Ft can be acquired with high precision. In such a case, the out-of-plane grid Gt becomes a grid G outside the detection surface Ft, and the high-precision surface acquisition unit 806 is omitted.

When K is a threshold value or a predetermined value, "more than K" and "more than K" are interchangeable as long as there is no technical contradiction or inconvenience. The same applies to "less than K" and "less than K."

Similar expressions such as "obtain", "estimate", "detect", "detect", "calculate", "extract", "generate", etc. can be replaced as appropriate within the scope of technical contradiction. It is.

It goes without saying that the elements constituting the above-described embodiments are not necessarily essential, except in cases where they are specifically specified as essential or where they are clearly considered essential in principle. In addition, when numerical values such as the number, amount, range, etc. of a component are mentioned, unless it is clearly stated that it is essential or it is clearly limited to a specific numerical value in principle, the specific numerical value The present invention is not limited thereto. Similarly, when the shape, direction, positional relationship, etc. of components, etc. is mentioned, unless it is clearly stated that it is essential, or when it is limited in principle to a specific shape, direction, positional relationship, etc. , the present invention is not limited to its shape, direction, positional relationship, etc.

Modifications are also not limited to the above examples. Also, multiple variants may be combined with each other. Furthermore, all or part of the above embodiments and all or part of the modifications may be combined with each other.

2 Distance sensor 3 Surface detection sensor 8 Parking support ECU (object detection device)
803 Range-finding point acquisition unit 805 Detection plane acquisition unit 808 Object detection unit Ft Detection plane Ps Range-finding point V Own vehicle Vt Other vehicle

Claims (10)

An object detection device (8) configured to be mounted on the own vehicle (V) to detect objects (Vt) existing around the own vehicle,
a ranging point acquisition unit (803) that obtains a detection result of the ranging point using a ranging sensor (2) for detecting ranging points (Ps) around the own vehicle;
a sensing surface acquisition unit (805) that acquires a sensing surface (Ft) that is a detection result of the object surface using a surface detection sensor (3) for detecting an object surface that is the outer surface of the object;
an object detection unit (808) that detects the object based on the distance measurement point selected according to the positional relationship with the detection surface in the space around the host vehicle;
Equipped with
The object detection unit selects the ranging point to be used for detecting the object based on a high-precision surface (FtH) that is the detection surface with a predetermined high detection accuracy.
Object detection device. further comprising a map generation unit (807) that generates a map (M) indicating the existence probability of the object in the space around the own vehicle;
The map generation unit sets, in the map, a low probability region (Gt) in which the existence probability is lower than other regions, based on the high-precision surface;
The object detection unit excludes the ranging point corresponding to the low probability area from the ranging points used for detecting the object.
The object detection device according to claim 1 . further comprising a high-precision surface acquisition unit (806) that obtains the high-precision surface based on the arrangement relationship between the surface detection sensor and the detection surface in the space around the own vehicle;
The object detection device according to claim 1 or 2 . The object detection unit selects the ranging point to be used for detecting the object based on the detection surface detected by the surface detection sensor during normal operation.
The object detection device according to any one of claims 1 to 3. The sensing surface acquisition unit acquires the sensing surface based on a collective state of a plurality of sensing points (Pd) on the object surface acquired using the surface detection sensor.
The object detection device according to any one of claims 1 to 4 . An object detection program executed by an object detection device (8) configured to be mounted on the own vehicle (V) to detect objects (Vt) existing around the own vehicle,
The process executed by the object detection device includes:
A process of acquiring a detection result of the distance measurement point using a distance measurement sensor (2) for detecting the distance measurement point (Ps) around the own vehicle;
A process of acquiring a detection surface (Ft) that is a detection result of the object surface using a surface detection sensor (3) for detecting an object surface that is the outer surface of the object;
a process of detecting the object based on the distance measurement point selected according to the positional relationship with the detection surface in the space surrounding the host vehicle;
including;
The process of detecting the object includes a process of selecting the ranging point used for detecting the object based on a high-precision surface (FtH) that is the detection surface with a predetermined high detection accuracy.
Object detection program. A process of setting a low probability region (Gt) in which the existence probability of the object in the space around the host vehicle is lower than other regions in a map (M) indicating the existence probability, based on the high-precision surface. further including;
The process of selecting the ranging point used for detecting the object includes a process of excluding the ranging point corresponding to the low probability area from the ranging points used for detecting the object.
The object detection program according to claim 6 . further comprising a process of acquiring the high-precision surface based on the arrangement relationship of the surface detection sensor and the detection surface in the space around the own vehicle;
The object detection program according to claim 6 or 7 . The process of detecting the object includes a process of selecting the ranging point used for detecting the object based on the detection surface detected by the surface detection sensor during normal operation.
The object detection program according to any one of claims 6 to 8 . The process of acquiring the detection plane includes a process of acquiring the detection plane based on a collective state of a plurality of detection points (Pd) on the object plane acquired using the surface detection sensor. ,
The object detection program according to any one of claims 6 to 9 .

Images