JP2010164349A - Control unit - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a control unit for reducing processing loads in calculating the shape of a parking space. <P>SOLUTION: The control unit is positioned on a line extended radially at each prescribed angle (for each five degree) in reference to a vehicle 1, detects distance to respective walls 200b-200d for dividing the parking space 200 by a distance sensor 26a, and calculates the shape of the parking space 200 according to the detection result, thus reducing the processing loads of a CPU of the control unit as compared with a case for calculating the shape of the parking space 200 by analyzing images input by a camera one pixel at a time. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a control device, and more particularly to a control device that can reduce the processing burden for calculating the shape of a parking space.

  Japanese Patent Application Laid-Open No. 2008-44529 discloses that a camera is attached to each of the front, rear, left and right sides of the vehicle, images acquired by the four cameras are analyzed, and an overhead view image of the parking space is created from the analysis result. An automobile that displays a created overhead image on a touch panel is disclosed. This automobile is configured to display an optimal route for parking in a parking space on a touch panel and provide parking assistance to the driver.

JP 2008-44529 A

  2. Description of the Related Art In recent years, some automobiles have an automatic parking function as one of parking assistances, in which an accelerator operation, a brake operation, and wheel steering are automatically performed and moved to a parking space without a driver's operation. When the technology described in Patent Document 1 is applied to an automobile having this automatic parking function, an image acquired by a camera attached to each side of the vehicle is analyzed and the shape of the parking space is calculated (overhead image is created) Then, automatic parking is performed by selecting an optimum route for the vehicle to move from the calculated shape of the parking space.

  However, since the analysis of the image input by the camera is performed on all pixels one by one, the processing load on the arithmetic processing unit such as a CPU increases. In particular, if a plurality of cameras are attached or a camera with a high resolution (ie, a large number of pixels) is attached in order to accurately calculate the shape of the parking space, the processing load on the CPU further increases. was there.

  The present invention has been made to solve the above-described problems, and an object of the present invention is to provide a control device that can reduce the processing burden for calculating the shape of a parking space.

  In order to achieve this object, the control device according to claim 1 is mounted on a vehicle that is movable in a predetermined parking space by automatic driving, is attached to the vehicle, and radiates at predetermined angles with respect to the vehicle. Distance detecting means for detecting the distance to the object located on the line extending to the point, and point extracting means for extracting the distance to the object detected for each predetermined angle by the distance detecting means as a point in the coordinate system; A generating means for generating a virtual line segment that divides the parking space based on the points of the coordinate system extracted by the point extracting means; and a point of the coordinate system extracted by the point extracting means. And a position calculating means for calculating an estimated position where the vehicle is estimated to be positioned with respect to the parking space.

  The control device according to claim 2 is the control device according to claim 1, wherein the generation means is one of the first points when the first point extracted by the point extraction means is used as a reference. A point-to-point distance with the second point extracted before or after one is calculated, and when the calculated point-to-point distance is equal to or greater than a predetermined value, at least one of the first and second points is invalidated And a virtual line segment that partitions the parking space is generated by virtually connecting points other than the points invalidated by the point invalidating means.

  According to a third aspect of the present invention, in the control device according to the second aspect, the line segment generated by virtually connecting by the generating unit is divided into portions that are substantially straight lines to generate a divided line segment. A dividing unit that extracts a dividing line segment having a point with the shortest distance from the vehicle among the end points of the dividing line segment divided by the dividing unit, and the line A dividing line segment having an inclination of a certain angle or more with respect to the reference line segment extracted by the segment extracting means, and a dividing line segment having an inclination of a certain angle or more with respect to the orthogonal line segment orthogonal to the reference line segment. First line segment invalidating means for invalidation.

  The control device according to claim 4 is the control device according to claim 2 or 3, wherein the dividing line segment generated by virtually connecting by the generating means is divided into portions that are substantially straight lines. A dividing means for generating a line segment, a line segment length calculating means for calculating the length of the divided line segment divided by the dividing means, and a length of the divided line segment calculated by the line segment length calculating means is constant. Second line segment invalidating means for invalidating the divided line segment when the length is equal to or shorter than the length.

  According to the control device of the first aspect, the distance to the object located on the line extending radially at every predetermined angle with respect to the vehicle is detected by the distance detecting means, and the distance to the detected object is determined. A point is extracted by the point extraction means as a point of the coordinate system, and based on the point of the coordinate system, a virtual line segment that divides the parking space is generated by the generation means, and it is estimated that the vehicle is positioned with respect to the parking space. The estimated position is calculated by the position calculating means. And since the shape of the parking space can be recognized from the virtual line segment that divides the parking space and the position of the vehicle relative to the parking space can be recognized from the estimated position of the vehicle, the vehicle can be moved into the parking space by automatic driving. . In this case, according to the present invention, since the shape of the parking space can be calculated as a virtual line segment from the detection result of the distance to the object detected by the distance detection means, the image input by the camera is analyzed. As compared with the case where the shape of the parking space is calculated, there is an effect that the processing load of the control device can be reduced.

  According to the control device of the second aspect, in addition to the effect produced by the control device according to the first aspect, the following effect is obtained. That is, the inter-point distance between the first point extracted by the point extracting means and the second point extracted immediately before or after the first point is calculated, and the calculated distance between the points is calculated. If the distance is greater than or equal to a predetermined value, at least one of the first and second points is invalidated by the point invalidating means. And the virtual line segment which divides a parking space is produced | generated by connecting points other than the point invalidated by the point invalidation means virtually. Therefore, since a virtual line segment can be generated by invalidating a point that is extremely far away from a plurality of points extracted by the point extraction means, for example, the distance from the vehicle to the object is affected by noise or the like. Anomalous points detected to be zero or the maximum value and line segments of shadows of objects can be invalidated. Therefore, since it is possible to suppress the virtual line segment that partitions the parking space from being extremely different from the actual shape of the parking space, it is possible to accurately recognize the shape of the parking space.

  According to the control device of the third aspect, in addition to the effect produced by the control device according to the second aspect, the distance from the vehicle among the end points of the dividing line segment divided by the dividing means into portions that are substantially straight lines. The reference line segment is the dividing line segment having the shortest point, and the fixed line segment has an inclination of a certain angle or more with respect to the reference line segment, and a constant angle with respect to the orthogonal line segment orthogonal to the reference line segment. Since the dividing line segment having the above inclination is invalidated by the first line segment invalidating means, it is possible to remove the dividing line segment whose inclination is extremely deviated from the reference line segment and the orthogonal line segment. Therefore, for example, the dividing line segment generated when the distance detection means detects a human figure or object in the parking space, or the dividing line segment generated due to the influence of noise or the like can be removed. Thus, it is possible to prevent the actual line segment from being extremely different from the actual shape of the parking space and to accurately recognize the shape of the parking space.

  According to the control device of the fourth aspect, in addition to the effect exerted by the control device according to the second or third aspect, the length of the divided line segment divided by the dividing means for each portion that is substantially a straight line is a line segment. When the length of the dividing line segment calculated by the length calculating unit is equal to or less than a predetermined length, the dividing line segment can be invalidated by the second line segment invalidating unit. Can be removed. Therefore, for example, the dividing line segment generated when the distance detection means detects a human figure or object in the parking space, or the dividing line segment generated due to the influence of noise or the like can be removed. Thus, it is possible to prevent the actual line segment from being extremely different from the actual shape of the parking space and to accurately recognize the shape of the parking space.

It is the schematic diagram which showed typically the top view of the vehicle in one Embodiment of this invention. It is the block diagram which showed the electric constitution of the control apparatus. It is the flowchart which showed the automatic parking process performed with CPU of a vehicle. It is the flowchart which showed the filtering process performed within an automatic parking process. (A) is the perspective view which showed the parking space where a vehicle is parked, (b) is the schematic diagram which showed typically the positional relationship of a parking space and a recognition position. (A) is a figure which shows the detection range which a distance sensor can detect the distance to a target object, (b) is the figure which showed the detection result by a distance sensor. (A) is the figure which showed the state which extracted the line segment from the detection result by a distance sensor, (b) is the figure which showed the state which divided | segmented the line segment extracted from the detection result by a distance sensor. (A) is the figure which showed the search method of the reference line segment, (b) is the figure which showed the calculation direction of the length filter and the inclination filter. It is the figure which showed the conditions of the inclination filter. It is the figure which showed the method of extracting the orthogonal line segment orthogonal to a reference line segment. It is the figure which showed the method of extracting the intersection of a reference line segment and an orthogonal line segment. It is the figure which showed the calculation method of the shape of the parking space divided according to the approach direction of the vehicle with respect to a parking space, and the position of the vehicle with respect to a parking space. It is the figure which showed the calculation method of the shape of the parking space divided according to the approach direction of the vehicle with respect to a parking space, and the position of the vehicle with respect to a parking space. It is the figure which showed the calculation method of the shape of the parking space divided according to the approach direction of the vehicle with respect to a parking space, and the position of the vehicle with respect to a parking space. It is the figure which showed the calculation method of the shape of the parking space divided according to the approach direction of the vehicle with respect to a parking space, and the position of the vehicle with respect to a parking space. It is the figure which showed the calculation method of the shape of the parking space divided according to the approach direction of the vehicle with respect to a parking space, and the position of the vehicle with respect to a parking space.

  Hereinafter, preferred embodiments of the present invention will be described with reference to the accompanying drawings. FIG. 1 is a schematic diagram schematically showing a top view of a vehicle 1 according to an embodiment of the present invention. Note that arrows UD, LR, and FB in FIG. 1 indicate the up-down direction, the left-right direction, and the front-rear direction of the vehicle 1, respectively.

  First, a schematic configuration of the vehicle 1 will be described with reference to FIG. As shown in FIG. 1, a vehicle 1 includes a body frame BF, a plurality of (four wheels in this embodiment) wheels 2 supported by the body frame BF, and a part of the plurality of wheels 2 ( In the present embodiment, the wheel drive device 3 that rotationally drives the left and right front wheels 2FL, 2FR, the suspension device 4 that suspends each wheel 2 on the vehicle body frame BF, and a part of the plurality of wheels 2 (this embodiment) The embodiment mainly includes a steering device 6 that steers the left and right front wheels 2FL, 2FR) and a steering drive device 5 that steers the wheels 2 (left and right front wheels 2FL, 2FR) in the same manner as the steering device 6. The shape of the parking space 200 (refer to FIG. 5A) is calculated based on the detection result of the distance sensor 26a of the monitoring device 26 (refer to FIG. 2), and automatic parking is possible in the calculated parking space 200. ing

  Next, the detailed configuration of each part will be described. The vehicle body frame BF forms a skeleton of the vehicle 1, supports the suspension device 4, and supports the wheels 2 via the suspension device 4. The suspension device 4 is a device that functions as a so-called suspension, and is provided on each wheel 2 independently as shown in FIG.

  As shown in FIG. 1, the wheel 2 includes left and right front wheels 2FL, 2FR disposed on the front side (arrow F side) of the vehicle body frame BF and left and right wheels disposed on the rear side (arrow B side) of the vehicle body frame BF. Four rear wheels 2RL and 2RR are provided. The left and right front wheels 2FL and 2FR are configured as driving wheels that are rotationally driven by the wheel driving device 3, while the left and right rear wheels 2RL and 2RR are configured as driven wheels that are driven as the vehicle 1 travels. ing.

  The wheel drive device 3 is a device for rotationally driving the left and right front wheels 2FL, 2FR, and includes a motor 3a that applies a rotational driving force to the left and right front wheels 2FL, 2FR. As shown in FIG. 1, the motor 3a is connected to the left and right front wheels 2FL, 2FR via a differential gear (not shown) and a pair of drive shafts 31.

  For example, when the driver operates the accelerator pedal 11, a rotational driving force is applied from the motor 3a to the left and right front wheels 2FL, 2FR, and the left and right front wheels 2FL, 2FR move at a speed corresponding to the depression state of the accelerator pedal 11. Is driven to rotate. The difference in rotation between the left and right front wheels 2FL and 2FR is absorbed by the differential gear.

  As described above, the steering device 6 is a device for steering the left and right front wheels 2FL, 2FR. As shown in FIG. 1, the steering shaft 61, the hook joint 62, the steering gear 63, the tie rod 64, The knuckle arm 65 is mainly provided. In the steering device 6, the steering gear 63 is configured by a rack and pinion mechanism including a pinion 63a and a rack 63b.

  For example, when the driver operates the steering wheel 13, the operation of the steering wheel 13 is transmitted to the hook joint 62 via the steering shaft 61, and the angle is changed by the hook joint 62, and the steering gear 63 rotates to the pinion 63 a. It is transmitted as movement. Then, the rotational motion transmitted to the pinion 63a is converted into the linear motion of the rack 63b, and the rack 63b moves linearly, whereby the tie rods 64 connected to both ends of the rack 63b move, and the wheels are connected via the knuckle arm 65. 2 is steered.

  Similar to the steering device 6, the steering drive device 5 is a device for steering the left and right front wheels 2 FL and 2 FR and includes a motor 5 a that applies a rotational driving force to the steering shaft 61. That is, when the motor 5a is driven and the steering shaft 61 rotates, the wheels 2 are steered in the same manner as when the driver operates the steering wheel 13.

  The accelerator pedal 11, the brake pedal 12, the steering wheel 13, and the automatic parking switch 27 are all operation members operated by the driver, and the vehicle is operated according to the depression state (depression amount, depression speed, etc.) of each pedal 11, 12. 1 acceleration force and braking force are determined, and the turning radius and turning direction of the vehicle 1 are determined according to the operation state (operation amount, operation direction) of the steering wheel 13. Furthermore, when the automatic parking switch 27 is operated by the driver, an automatic parking mode is set in which the vehicle 1 is parked by autonomous traveling.

  The control device 100 is a device for controlling each part of the vehicle 1. For example, the control device 100 detects the depression state of each pedal 11, 12 and controls the wheel drive device 3 according to the detection result, The depression state of the pedal 12 is detected, and a brake device (not shown) is controlled according to the detection result. Moreover, the control apparatus 100 is based on the input from the distance sensor 26a attached to the front right (the edge part of the arrow F direction, and the edge part of the arrow R direction) of the vehicle 1, The parking space 200 (refer Fig.5 (a)). ) To calculate the shape and the like.

  Here, with reference to FIG. 2, the detailed structure of the control apparatus 100 is demonstrated. FIG. 2 is a block diagram showing an electrical configuration of the control device 100. As shown in FIG. 2, the control device 100 includes a CPU 91, a ROM 92, a RAM 93, and a hard disk (hereinafter referred to as “HDD”) 96, which are connected to an input / output port 95 via a bus line 94.

  The input / output port 95 includes a wheel drive device 3, a steering drive device 5, an accelerator pedal sensor device 21, a brake pedal sensor device 22, a steering sensor device 23, a vehicle speed information acquisition device 24, a vehicle body posture sensor device 25, and a surrounding environment. A monitoring device 26, an automatic parking switch 27, a liquid crystal display (hereinafter referred to as "LCD") 97, a GPS receiver 98, and other input / output devices 99 are connected.

  The CPU 91 is an arithmetic device that controls each unit connected by the bus line 94, and the ROM 92 is a non-rewritable nonvolatile memory for storing a control program executed by the CPU 91, fixed value data, and the like. A program for executing an automatic parking process shown in FIG. 3 and a filtering process shown in FIG.

  The RAM 93 is a memory that stores various data in a rewritable manner when the control program is executed, and includes an end point filter repeat flag 93a. The end point filter repeat flag 93a is a flag for determining whether the end point filter processing (S204 in FIG. 4) in the filtering processing described later is the first time or the second time, and is turned on in the step S207 in FIG. 3 is turned off (not shown) when the automatic parking process shown in FIG. 3 is executed.

  The HDD 96 is a rewritable nonvolatile large-capacity memory, and includes a map database (hereinafter referred to as “map DB”) 96a, a parking lot database (hereinafter referred to as “parking place DB”) 96b, and a parking position storage area 96c. And a path storage area 96d. In the present embodiment, the parking lot DB 96b, the parking position storage area 96c, and the route storage area 96d store various data related to a parking lot (not shown) with which the driver of the vehicle 1 is contracted.

  The map DB 96a is a database that stores map data. For example, the map DB 96a is read from a medium (for example, a DVD) on which map data is recorded by a data reading device (for example, a DVD device) (not shown), an external information center, or the like. The map data received from the communication device (not shown) is stored.

  The parking lot DB 96b is a database that stores parking lot data. The parking lot data stored in the parking lot DB 96b includes, for example, the overall shape of the parking lot contracted by the driver of the vehicle 1, the position and size of the parking space, the position of the wall, the width of the road, etc. Is included.

  In the parking position storage area 96c, in the parking lot indicated by the parking lot data of the parking lot DB 96b, parking position data (for example, referring to FIG. 5A) indicating a specific parking space 200 contracted by the driver of the vehicle 1 is provided. The coordinate data of the center position of the parking space 200 is stored. In the parking position storage area 96c, coordinate data of a stop position 201 (see FIG. 5B) that is located in the vicinity of the parking space 200 and stops when the vehicle 1 calculates the shape of the parking space 200, and Data indicating the area of the recognition range 201a (see FIG. 5B) with the stop position 201 as a reference is also stored.

  In the route storage area 96d, a route indicating a movement route until the vehicle 1 automatically parks in a specific parking space 200 (see FIG. 5A) by autonomous traveling in the parking lot indicated by the parking lot data of the parking lot DB 96b. Data is stored. That is, when the vehicle 1 performs autonomous traveling, the vehicle 1 automatically moves along the movement route indicated by the route data.

  As described above, the wheel driving device 3 is a device for rotationally driving the left and right front wheels 2FL, 2FR (see FIG. 1), and an electric motor 3a for applying a rotational driving force to the left and right front wheels 2FL, 2FR; A control circuit (not shown) for controlling the electric motor 3a based on a command from the CPU 91 is mainly provided.

  As described above, the steering drive device 5 is a device for steering the left and right front wheels 2FL and 2FR. The motor 5a for applying a rotational driving force to the steering shaft 61 and the electric motor 5a according to a command from the CPU 91. It mainly comprises a control circuit (not shown) for controlling based on it.

  The accelerator pedal sensor device 21 is a device for detecting the depression state of the accelerator pedal 11 and outputting the detection result to the CPU 91. An angle sensor (not shown) for detecting the depression amount of the accelerator pedal 11; And a processing circuit (not shown) that processes the detection result of the angle sensor and outputs the result to the CPU 91.

  The brake pedal sensor device 22 is a device for detecting the depression state of the brake pedal 12 and outputting the detection result to the CPU 91. An angle sensor (not shown) for detecting the depression amount of the brake pedal 12; And a processing circuit (not shown) that processes the detection result of the angle sensor and outputs the result to the CPU 91.

  The steering sensor device 23 is a device for detecting the operation state of the steering wheel 13 and outputting the detection result to the CPU 91, and an angle sensor (not shown) for detecting the operation amount of the steering wheel 13 in association with the operation direction. And a processing circuit (not shown) for processing the detection result of the angle sensor and outputting the result to the CPU 91.

  In the present embodiment, each angle sensor is configured as a contact-type potentiometer using electric resistance. The CPU 91 can obtain the amount of depression of the pedals 11 and 12 and the amount of operation of the steering wheel 13 based on the detection results of the angle sensors input from the sensor devices 21, 22 and 23.

  The vehicle speed information acquisition device 24 is a device for outputting the vehicle speed of the vehicle 1 to the CPU 91, an angle sensor (not shown) that outputs a pulse each time the drive shaft 31 rotates by a predetermined angle, and detection of the angle sensor. A processing circuit (not shown) that calculates the vehicle speed of the vehicle 1 based on the number of pulses within a predetermined time from the result and outputs it to the CPU 91 is provided.

  The vehicle body posture sensor device 25 is a device for outputting the posture (traveling direction) of the vehicle 1 to the CPU 91, a gyro sensor (not shown) for detecting the posture of the vehicle 1 using geomagnetism, and the gyro sensor of the gyro sensor. A processing circuit (not shown) that calculates the attitude of the vehicle 1 from the detection result and outputs the calculated attitude to the CPU 91 is provided.

  The surrounding environment monitoring device 26 is a device for outputting distance data to an object such as a wall existing around the vehicle 1 to the CPU 91, and a distance for detecting a distance to the object existing around the vehicle 1. A sensor 26a is connected, and a processing circuit (not shown) for calculating the distance to the object from the input result input by the distance sensor 26a and outputting it to the CPU 91 is provided. In the present embodiment, the distance sensor 26a is attached to the front right of the vehicle 1, and is configured by a laser range finder that irradiates the target with an infrared laser and measures the distance to the target with the degree of reflection. .

  The automatic parking switch 27 is a switch operated when the vehicle 1 is automatically parked in a specific parking space 204 (see FIG. 3) by autonomous traveling. When the automatic parking switch 27 is operated, the automatic parking switch 27 is switched to the automatic parking mode. It shifts and it is constituted so that automatic parking processing shown in Drawing 3 may be performed.

  The LCD 97 is a liquid crystal display that displays a map based on the map data, a map of the parking lot based on the parking lot data, the shape of the parking space 200 (see FIG. 5A) calculated by the CPU 91, and various other information. .

  The GPS receiver 98 is a device that receives position information (for example, latitude information and longitude information) from a GPS satellite (not shown) via an antenna 98a. When the position information is received by the GPS receiver 98, Based on the position information, the CPU 91 calculates the actual position of the vehicle 1.

  Further, when reception of the position information by the GPS receiver 98 is interrupted, information on the vehicle speed of the vehicle 1 input from the vehicle speed information acquisition device 24 and the posture of the vehicle 1 input from the vehicle body posture sensor device 25 are included. Based on this, the CPU 91 calculates a virtual position of the vehicle 1. The calculation of the virtual position of the vehicle 1 is hereinafter referred to as dead reckoning.

  Examples of other input / output devices 97 include input from a key cylinder of the vehicle 1 and input / output devices such as a window opening / closing switch, a light, and a lamp.

  Next, with reference to FIGS. 3-16, the automatic parking process performed when the vehicle 1 is automatically parked in the parking space 200 is demonstrated. FIG. 3 is a flowchart showing an automatic parking process executed by the CPU 91 of the vehicle 1, and FIG. 4 is a flowchart showing a filtering process executed within the automatic parking process. The automatic parking process is performed within the range in which the vehicle 1 can perform automatic parking in the parking space 200 (for example, in the parking lot indicated by the parking data stored in the parking lot DB 96b). It is executed when an operation is performed.

  If the automatic parking process shown in FIG. 3 is performed, the movement process to the recognition position 201 located in front of the parking space 200 mentioned later is performed (S101). This movement process is a process of outputting an instruction to the wheel drive device 3 and the steering drive device 5 to move the vehicle 1 by automatic driving. If the GPS receiver 98 can receive position information, the GPS reception is performed. If the vehicle 1 is automatically operated based on the position information received by the machine 98 and the position information cannot be received by the GPS receiver 98, the virtual current position of the vehicle 1 calculated by dead reckoning is set. Based on this, the vehicle 1 is automatically driven.

  Here, a parking space 200 where the vehicle 1 is parked by automatic driving and a recognition position 201 where the vehicle 1 moves by automatic driving will be described with reference to FIG. FIG. 5A is a perspective view showing a parking space 200 where the vehicle 1 is parked, and FIG. 5B is a schematic diagram schematically showing the positional relationship between the parking space 200 and the recognition position 201. It is. In FIG. 5B, the center point of the vehicle 1 is represented by Po, and the distance sensor 26a attached to the front right of the vehicle 1 is shown in an emphasized manner.

  As shown in FIG. 5A, the parking space 200 in which the vehicle 1 is parked has an entrance 200a that serves as an entrance when the vehicle 1 moves into the parking space 200, and faces the entrance 200a. Three sides are surrounded by the back wall 200b located in the back and the side walls 200c located on both sides, and extending walls 200d are erected in directions away from the ends of the side walls 200c on the inlet 200a side. That is, the parking space 200 is partitioned by the walls 200b to 200d.

  As shown in FIG. 5B, when the vehicle 1 is automatically parked in the parking space 200 in front of the entrance 200a of the parking space 200 (the left side in FIG. A recognition position 201 to stop is determined in advance. The range up to the radius X [m] with the recognition position 201 as the center is the recognition range 201a. In this embodiment, the recognition range 201a has a radius of 1 [m] with the recognition position 201 as the center, but this numerical value can be changed as appropriate according to the performance of the distance sensor 26a.

  Returning to FIG. When the movement process to the recognition position 201 is executed in the process of S101, it is determined whether or not the center point Po of the vehicle 1 is within the recognition range 201a (S102), and the center point Po of the vehicle 1 is the recognition position 201. If it is within the recognition range 201a (S102: Yes), the process proceeds to S103, and if the center point Po of the vehicle 1 is outside the recognition range 201a of the recognition position 201 (S102: No), the process returns to S101. Until the center point Po of the vehicle 1 is within the recognition range 201a, the processes of S101 and S102 are repeatedly executed.

  In the process of S103, since the center point Po of the vehicle 1 has moved into the recognition range 201a of the recognition position 201, a stop instruction is output to the wheel drive device 3 and the steering drive device 5 to stop the vehicle 1. (S103) Then, the periphery of the vehicle 1 is measured by the distance sensor 26a (S104), and a line segment forming the parking space 200 is extracted and divided based on the measurement result of the distance sensor 26a (S105).

  Here, with reference to FIG. 6, the measurement of the surroundings of the vehicle 1 performed by the process of S104 is demonstrated. FIG. 6A is a diagram showing a detection range in which the distance sensor 26a can detect the distance to the object, and FIG. 6B is a diagram showing a detection result by the distance sensor 26a. In FIG. 6B, the vehicle 1 is virtually illustrated by a broken line.

  As shown in FIG. 6A, the distance sensor 26a attached to the right front side of the vehicle 1 has a range of about 270 degrees from the rear to the left side of the vehicle 1 (a range indicated by an arrow A in FIG. 6A). It is comprised so that the distance with the target object located in can be detected. Moreover, the distance sensor 26a is within a range of about 270 degrees, and is up to an object located on a line extending radially every 0.5 degrees in the circumferential direction (for each range indicated by an arrow B in FIG. 6B). The distance can be detected.

  As shown in FIG. 6 (b), the detection result by the distance sensor 26a is extracted as a distance to the object detected every 0.5 degrees in the circumferential direction, and the extracted point is a distance. They are arranged in a coordinate system consisting of an x-axis and a y-axis in the order of detection by the sensor 26a.

  As shown in FIG. 6B, the extracted points are a group g1 corresponding to the extending wall 200d located in front of the vehicle 1 (in the arrow direction of the y axis), a group g2 corresponding to the back wall 200b, The group g3 corresponding to the side wall 200c located behind the vehicle 1 (counter arrow direction of the y axis) and the group g4 corresponding to the extended wall 200d located behind the vehicle 1 are mainly configured. In addition, since the side wall 200c located in front of the vehicle 1 is a blind spot from the distance sensor 26a, a point corresponding to the side wall 200c is not detected.

  Moreover, a point may be arranged in the position distant from the groups g1-g4 (one point of the upper right of FIG.6 (b)). This is the case, for example, when the detection result by the distance sensor 26a is detected as “maximum value” or “0” due to the influence of noise or the like generated between the distance sensor 26a, the surrounding environment monitoring device 26 and the CPU 91. is there. In addition, even when a person passes through the parking space 200 or an object is placed, the person or object is detected by the distance sensor 26a, and points not included in the groups g1 to g4 are arranged. Sometimes.

  Next, with reference to FIG. 7, the line segment extraction and division of the parking space 200 performed in the process of S105 will be described. FIG. 7A is a diagram showing a state in which a line segment is extracted from the detection result by the distance sensor 26a, and FIG. 7B is a diagram in which the line segment extracted from the detection result by the distance sensor 26a is divided. FIG. In FIG. 7A, the vehicle 1 is virtually illustrated by a broken line.

  As shown in FIG. 7A, when the points arranged in the coordinate system are connected from the detection result of the distance sensor 26a, a line segment parallel to the x-axis and the y-axis (the solid line segment in FIG. 7A) is obtained. ) And a line segment PQ0 in which a line segment oblique to the x-axis and the y-axis (a line segment illustrated by a dotted line in FIG. 7A) is continuously connected is extracted.

  Here, as described above, the distance sensor 26a detects the distance to each of the walls 200b to 200d in order to specify the shape of the parking space 200. Therefore, the distance between the points arranged in the coordinate system is extreme. If it is very wide, there is a possibility that the influence of noise or the like or a person or an object crossing the parking space has been detected, and it is preferable to invalidate an extremely wide point. Therefore, in the present embodiment, the point-to-point distance between the points arranged in the coordinate system is calculated, and when the point-to-point distance is equal to or greater than the predetermined distance, the points arranged at extremely distant positions are invalidated. (Point invalidation means). That is, one of the points constituting the line segment indicated by the dotted line shown in FIG. 7A is invalidated, and the line segment indicated by the dotted line is invalidated.

  The state shown in FIG. 7B shows a line segment that is extracted when a point at which the distance between points is extremely wide is invalid, and is a state in which the line segment PQ0 is divided. This division of the line segment PQ is obtained by dividing the portion corresponding to the boundary between the walls 200b to 200d surrounding the parking space 200. Specifically, the direction has changed extremely for each portion that is substantially a straight line. The part is divided. As shown in the figure, the line segment PQ0 is divided into dividing line segments PQ1 to PQ4, the dividing line segments PQ1 and PQ4 correspond to the extending wall 200d, the dividing line segment PQ2 corresponds to the back wall 200b, and the dividing line The minute PQ3 corresponds to the side wall 200c.

  Therefore, the extremely wide point-to-point distance between points arranged in the coordinate system can be invalidated according to the detection result of the distance sensor 26a. It is possible to invalidate abnormal points when a crossing person or object is detected. Therefore, the dividing line segments PQ1 to PQ4 that divide the parking space 200 can be suppressed from being extremely different from the actual shape of the parking space 200, so that the shape of the parking space 200 can be accurately recognized.

  Returning to FIG. When the surrounding measurement (S104) by the distance sensor 26a is completed and line segments corresponding to the walls 200b to 200d covering the parking space 200 are extracted and divided based on the measurement result (S105), next, other noise A filtering process for removing a line segment that may be affected by the above is executed (S106).

  Here, the filtering process executed in the process of S106 will be described with reference to FIG. When the filtering process is executed, first, a reference line segment is searched (S201), and a length filter is applied to all the divided line segments (S202) and an inclination filter is applied (S203). The reference line segment search performed in the process of S201 is a process of setting a reference line segment to be used as a reference in the filtering process among the divided line segments PQ1 to PQ4 divided in the process of S105. , Applying the length filter and the gradient filter performed in S203 is a process of removing unnecessary dividing line segments. The processing of S201 to S203 will be described with reference to FIGS.

  FIG. 8A is a diagram showing a reference line segment search method, FIG. 8B is a diagram showing calculation directions of the length filter and the gradient filter, and FIG. It is a figure showing conditions. Note that the y-axis of the coordinate system shown in FIG. 8B is arranged to be parallel to the reference line segment.

  As shown in FIG. 8A, the reference line segment search performed in the process of S201 searches the vehicle 1 for the end point P1 having the shortest distance from the end points of the dividing line segments PQ1 to PQ4. In this process, the dividing line segment having the end point P1 is set as the reference line segment. In the state of FIG. 8A, the end point P1 of the dividing line segment PQ1 indicated by the solid line arrow is the shortest. Therefore, in the following description, the dividing line segment PQ1 will be described as the reference line segment PQ1.

  As shown in FIG. 8B, the process of applying the length filter performed in the process of S202 calculates the length of each of the divided line segments PQ1 to PQ4 extracted and divided in the process of S105, and the division. When the lengths of the line segments PQ1 to PQ4 are equal to or shorter than the predetermined length, the line segments are invalidated and removed.

  For example, if the coordinates of the end points of the dividing line segment PQ5 located on the upper side of FIG. 8B are (Xs1, Ys1) and (Xe1, Ye1), the distance difference (Xs1-Xe1) in the x-axis direction is 2 The line segment length C can be calculated by adding the value obtained by squaring the distance difference (Ys1-Ye1) in the y-axis direction to the squared value. When the minimum line segment length is Cmin and the maximum line segment length is Cmax, a line segment that deviates from the condition of Cmin ≦ C ≦ Cmax is invalidated.

  Therefore, an extremely short dividing line segment can be invalidated among the dividing line segments. For example, a dividing line segment generated when the distance sensor 26a detects a human figure or an object in the parking space 200, noise, etc. A dividing line segment generated by the influence can be removed. Therefore, it can suppress that the virtual dividing line segment which divides the parking space 200 differs from the shape of the actual parking space 200 extremely, and can recognize the shape of the parking space 200 correctly.

  Next, in the process of applying an inclination filter performed in the process of S203, the inclination θ (−90 ≦ θ ≦ 90) of the remaining divided line is calculated by applying the length filter in the process of S202, and the divided line segment is calculated. In the process of removing the divided line segment as invalid when the inclination of the line segment is tilted by a predetermined angle or more with respect to the reference line segment PQ1 or tilted by a predetermined angle or more with respect to the orthogonal line segment orthogonal to the reference line segment PQ1. is there.

For example, the slope θ1 (0 ≦ θ ≦ 90) of the dividing line segment PQ5 located on the upper side of FIG. 8B can be calculated by the following equation (1), and the division located on the lower side of FIG. The slope θ2 (−90 ≦ θ ≦ 0) of the line segment PQ6 can be calculated by the following equation (2). The coordinates of the end points of the dividing line segment PQ6 are (Xs2, Ys2) and (Xe2, Ye2).
θ1 = tan−1 ((Ye1−Ys1) / (Xe1−Xs1)) (1)
θ2 = tan−1 ((Ye2−Ys2) / (Xe2−Xs2)) (2)
As shown in FIG. 8B, the angles θ1 and θ2 are calculated based on the smaller x value of the dividing line segments PQ5 and PQ6 (the left end point in FIG. 8B). The inclination θ1 has a relation of 0 ≦ θ1, the inclination θ2 has a relation of θ2 ≦ 0, and a relation of −90 ≦ θ ≦ 90.

  FIGS. 9A to 9D show the allowable ranges E1 to E3 of the inclinations θ1 and θ2 of the dividing line segments PQ5 and PQ6. FIG. 9A shows the allowable ranges E1 and E2 when the reference line segment PQ1 substantially overlaps the Y axis. As shown in FIG. 9A, the allowable range E1 in which the slopes θ1 and θ2 of the dividing line segments PQ5 and PQ6 are on the + side (0 ≦ θ) is within a range that is −θmin with respect to the y-axis. Yes. The reason why the range of + θmin with respect to the y axis is excluded from the allowable range is that the relationship of −90 ≦ θ ≦ 90 is established with respect to the inclinations θ1 and θ2. On the other hand, the allowable range E2 in which the inclinations θ1 and θ2 of the dividing line segments PQ5 and PQ6 are on the negative side (θ ≦ 0) is within the range of + θmin with respect to the y axis, and −90 ≦ θ ≦ 90. From the relationship, a range of −θmin with respect to the y-axis is excluded from the allowable range. Further, as shown in FIG. 9 (b), the range from −θmin to + θmin with respect to the perpendicular line segment (the line segment indicated by the thin line in FIG. 9 (b)) in the direction perpendicular to the reference line segment PQ1. The allowable range E3.

  FIG. 9C shows the allowable ranges E1 and E2 when the reference line segment PQ1 does not overlap the y-axis. As shown in FIG. 9C, the allowable ranges E1 and E2 are the reference line segment PQ1. Is within the range of -θmin to + θmin. Further, as shown in FIG. 9 (d), the range from −θmin to + θmin with respect to the perpendicular line segment (the line segment indicated by the thin line in FIG. 9 (d)) perpendicular to the reference line segment PQ1. The allowable range E3. Even when the reference line segment PQ1 does not overlap the y-axis, the relationship of −90 ≦ θ ≦ 90 is established, and when the reference line segment PQ1 deviates from −90 ≦ θ ≦ 90, the range is excluded.

  9C and 9D illustrate the case where the slope θ of the reference line segment PQ1 is 0 ≦ θ, the same applies when the slope θ of the reference line segment PQ1 is θ ≦ 0. In addition, the allowable ranges E1 and E2 are within the range of −θmin to + θmin with respect to the reference line segment PQ1, and the allowable range E3 is a range of −θmin to + θmin with respect to the orthogonal line segment perpendicular to the reference line segment PQ1. Inside.

  Therefore, by providing the permissible ranges E1 to E3 in the inclinations θ1 and θ2 of the dividing line segments PQ5 and PQ6, the dividing line segment that is substantially parallel to the reference line segment PQ1 and the vertical direction of the reference line segment PQ1 are approximately. Other than the parallel dividing line segments, that is, the dividing line segments that are extremely inclined with respect to the reference line segment PQ1, and the dividing line segments that are extremely inclined with respect to the vertical direction of the reference line segment PQ1, can be invalidated. That is, since the dividing line segment generated when the distance sensor 26a detects a human figure or an object in the parking space 200, or the dividing line segment generated due to the influence of noise or the like can be removed, the virtual partitioning the parking space 200 is possible. Therefore, it is possible to prevent the actual dividing line segment from being extremely different from the actual shape of the parking space 200 and to accurately recognize the shape of the parking space 200.

  Returning to FIG. Once the inclination filter is applied by the processing of S203, next, an end point filter is applied to extract an orthogonal line segment located in a direction orthogonal to the reference line segment PQ1 (S204). The end point filter performed in the process of S204 will be described with reference to FIG.

  FIG. 10 is a diagram illustrating a method of extracting orthogonal line segments orthogonal to the reference line segment PQ1. As shown in FIG. 10, it is determined whether or not there are end points of other dividing line segments PQ2 to PQ4 in the search range P1a that is a predetermined range centering on the end point P1 of the reference line segment PQ1 that is the shortest point from the vehicle 1. To detect. The search range P1a centered on the end point P1 of the reference line segment PQ1 is a range in which, for example, about 1 m is added (+ α) to the width of the vehicle 1 in the left-right direction (arrow LR direction shown in FIG. 1). (If you know the width of the parking lot entrance, it is desirable to set the parking entrance width + α). That is, the end points of the dividing line segments PQ2 to PQ4 corresponding to the walls 200b to 200d are detected. In FIG. 10, since one end point of the dividing line segments PQ2 to PQ4 is located within the search range P1a, all of the dividing line segments PQ2 to PQ4 are candidates for orthogonal line segments, but with respect to the reference line segment PQ1. Since the dividing line segment orthogonal to each other becomes the dividing line segment PQ3, the dividing line segment PQ3 is extracted as the orthogonal line segment. In the following description, the dividing line segment PQ3 is referred to as an orthogonal line segment PQ3.

  Further, for example, when there is no end point of another line segment within the predetermined range P1a centering on the end point P1 of the reference line segment PQ1, the direction (approach direction) of the vehicle 1 with respect to the parking space 200 may be different. Therefore, the search is performed within the search range P2a (within the circle indicated by the dotted line in FIG. 10) centered on the other end point P2 of the reference line segment PQ1. For example, if the end points of the line segments PQ7 and PQ8 can be detected by this search, the orthogonal line segment PQ7 orthogonal to the reference line segment PQ1 can be extracted.

  Returning to FIG. When an end point filter for searching whether there are end points of other divided line segments PQ2 to PQ4 within the search range P1a centering on the end point P1 of the reference line segment PQ1 by the processing of S204, the end point filter repeat flag 93a is on. If the end point filter repeat flag 93a is off (S205: No), it is determined whether there is one or more orthogonal line segments extracted by the end point filter (S206).

  As a result of checking in the process of S206, if there are one or more orthogonal line segments in the line segment extracted by the end point filter (S206: Yes), the shape of the parking space 200 can be specified, so this process is terminated as it is, If the orthogonal line segment cannot be extracted (S206: No), after turning on the end point filter repeat flag 93a (S207), the process returns to S204, and the end point filter is applied again. When the end point filter is applied again, as described above, the process of searching the search range P2a centered on the other end point P2 of the reference line segment PQ1 is executed.

  In the process of S205, when the end point filter is applied again, since the end point filter repeat flag 93a is turned on (S205: Yes), the process ends without executing the process of S206. In addition, if the orthogonal line segment cannot be searched even when the end point filter is applied again, the parking space 200 cannot be recognized in the process of S107 described later (S207: No), and automatic parking without automatic parking is performed. The process ends. Therefore, by making the process of applying the end point filter up to a maximum of two times, it is possible to suppress the end point search range from becoming too wide, thereby reducing the accuracy and reliability of the process, and reducing the processing burden. Can do. In addition, when the parking space 200 cannot be recognized, you may make it perform notification which urges | hangs the approach direction etc. of the vehicle 1, and performs an automatic parking process again.

  Returning to FIG. When the filtering process of S106 ends, the parking space 200 from which unnecessary line segments are removed by the filtering process is recognized (S107). The process which recognizes this parking space 200 is demonstrated with reference to FIGS.

  FIG. 11 is a diagram illustrating a method of extracting the intersection P3 between the reference line segment PQ1 and the orthogonal line segment PQ3. FIGS. 12-16 is the figure which showed the calculation method of the shape of the parking space 200 divided according to the approach direction of the vehicle 1 with respect to the parking space 200, and the position of the vehicle 1 with respect to the parking space 200. FIG.

  As shown in FIG. 11, if an orthogonal line segment PQ3 orthogonal to the reference line segment PQ1 can be searched by applying an end point filter in the process of S205, it is on an extension line of the reference line segment PQ1 (a dotted line extending from the reference line segment PQ1). Then, a point where the extension of the orthogonal line segment PQ3 (a dotted line extending from the orthogonal line segment PQ3) intersects is extracted as an intersection point P3. A space between the end point P1 of the reference line segment PQ and the extracted intersection point P3 is recognized as the width of the entrance 200a of the parking space 200.

  12 to 16, the distance from the detection result by the distance sensor 26a to a plurality of points (points shown in FIG. 6B) arranged in the coordinate system, and the circumference of the distance sensor 26a at each point. The approach direction of the vehicle 1 with respect to the parking space 200 and the position of the vehicle 1 with respect to the parking space 200 are calculated according to the angle in the direction, and the calculation is performed based on the calculation result.

  12 and 13 show a case where the vehicle 1 approaches the recognition position 201 in the state where the parking space 200 is located on the right side of the vehicle 1 as shown in FIG. The vertical width parkDepth (hereinafter referred to as “pD”) and the horizontal width parkWidth (hereinafter referred to as “pW”) of the parking space 200 illustrated in FIG. 12A are data stored in the HDD 96 in advance.

In the state shown in FIG. 12A, five dividing line segments PQ1 to PQ5 are extracted. A method for calculating the apexes P11 to P14 at the four corners that specify the shape of the parking space 200 in this state will be described. As described above, the vertex P11 is the intersection P3 of the reference line segment PQ1 and the orthogonal line segment PQ3. When the coordinates are represented as (Xp11, Yp11), the coordinates (Xp12, Yp12) of the vertex P12 are ( 3) and (4), the coordinates (Xp13, Yp13) of the vertex P13 can be calculated by the following expressions (5) and (6), and the coordinates (Xp14, Yp14) of the vertex P14 (end point P1) are (7) and (8).
Xp12 = Xp11−pD · cos θ (3)
Yp12 = Yp11 + pD · sin θ (4)
Xp13 = Xp12 + pW · sin θ (5)
Yp13 = Yp12 + pW · cos θ (6)
Xp14 = Xp13 + pD · cos θ (7)
Yp14 = Yp13−pD · sin θ (8)
When the coordinates of the vertices P12 to P14 are calculated by the above equations (3) to (8), the shape of the virtual parking space 200 as indicated by a thin line can be specified.

  In the state shown in FIG. 12B, the vehicle 1 is positioned without tilting with respect to the parking space 200. Therefore, since the angle θ = “0”, in the above (3) to (8), the vertices P12 to P14 can be calculated with sin θ = “1” and cos θ = “0”.

In the state shown in FIG. 13A, the vertices P12 to P14 can be calculated by the following equations (9) to (14).
Xp12 = Xp11 + pD · cos (π / 2 + θ) (9)
Yp12 = Yp11 + pD · sin (π / 2 + θ) (10)
Xp13 = Xp12 + pW · sin (π / 2 + θ) (11)
Yp13 = Yp12−pW · cos (π / 2 + θ) (12)
Xp14 = Xp13−pD · cos (π / 2 + θ) (13)
Yp14 = Yp13−pD · sin (π / 2 + θ) (14)

In the state shown in FIG. 13B, the vertices P12 to P14 can be calculated by the following equations (15) to (20).
Xp12 = Xp11 + pD · cos (π / 2−θ) (15)
Yp12 = Yp11−pD · sin (π / 2−θ) (16)
Xp13 = Xp12−pW · sin (π / 2−θ) (17)
Yp13 = Yp12−pW · cos (π / 2−θ) (18)
Xp14 = Xp13−pD · cos (π / 2−θ) (19)
Yp14 = Yp13 + pD · sin (π / 2−θ) (20)

  14 and 15 show a case where the vehicle 1 approaches the recognition position 201 in a state where the parking space 200 is located on the left side of the vehicle 1.

In the state shown in FIG. 14A, the vertices P12 to P14 can be calculated by the following equations (21) to (26).
Xp12 = Xp11−pD · cos (π / 2−θ) (21)
Yp12 = Yp11 + pD · sin (π / 2−θ) (22)
Xp13 = Xp12−pW · sin (π / 2−θ) (23)
Yp13 = Yp12−pW · cos (π / 2−θ) (24)
Xp14 = Xp13 + pD · cos (π / 2−θ) (25)
Yp14 = Yp13−pD · sin (π / 2−θ) (26)

  In the state shown in FIG. 14B, the vehicle 1 is positioned without tilting with respect to the parking space 200. Therefore, the angle θ = “0”, and in the above (21) to (26), sin (π / 2−θ) = “1” and cos (π / 2−θ) = “0”, and the vertices P12˜ P14 can be calculated.

In the state shown in FIG. 15A, the vertices P12 to P14 can be calculated by the following equations (27) to (32).
Xp12 = Xp11 + pD · cos (π / 2 + θ) (27)
Yp12 = Yp11 + pD · sin (π / 2 + θ) (28)
Xp13 = Xp12−pW · sin (π / 2 + θ) (29)
Yp13 = Yp12 + pW · cos (π / 2 + θ) (30)
Xp14 = Xp13−pD · cos (π / 2 + θ) (31)
Yp14 = Yp13−pD · sin (π / 2 + θ) (32)

In the state shown in FIG. 15B, the vertices P12 to P14 can be calculated by the following equations (33) to (38).
Xp12 = Xp11 + pD · cos (π / 2−θ) (33)
Yp12 = Yp11−pD · sin (π / 2−θ) (34)
Xp13 = Xp12 + pW · sin (π / 2−θ) (35)
Yp13 = Yp12 + pW · cos (π / 2−θ) (36)
Xp14 = Xp13−pD · cos (π / 2−θ) (37)
Yp14 = Yp13 + pD · sin (π / 2−θ) (38)

  Returning to FIG. In the process of S107, the optimal state for the approach direction and position of the vehicle 1 is selected from the states shown in FIGS. 12A to 15B, and the vertices P11 to P14 are calculated and the parking space 200 is displayed. A virtual shape is calculated to determine whether or not the parking space 200 is recognized (S108). If the parking space 200 is recognized (S108: Yes), a parking process is executed (S109), and parking is performed. If the space 200 is not recognized (S108: No), this process is terminated as it is. The case where the parking space 200 cannot be recognized in the process of S107 means that when the filtering process (S106) is executed, the reference line segment and the orthogonal line segment cannot be extracted, and the shape of the parking space 200 cannot be calculated. Is the case.

  Here, with reference to FIG. 16, the parking process performed by the process of S109 is demonstrated. FIG. 16 is a diagram showing automatic parking of the vehicle 1 in the parking space 204. FIG. 16 (a) shows a state where the vehicle 1 moves to the parking start position, and FIG. A state in which the vehicle 1 turns and parks in the parking space 200 is shown.

  As shown in FIG. 8A, when the vehicle 1 automatically parks in the parking space 200, the CPU 91 outputs an instruction to the wheel drive device 3 and the steering drive device 5 and starts parking from the current position with respect to the parking space 200. The vehicle 1 is automatically moved to the position P21. Then, as shown in FIG. 8B, when the vehicle 1 reaches the parking start position P21, a turning operation is performed with reference to the turning center point P22, and automatic parking in the parking space 200 is performed. Driving is complete.

  As described above, the control device 100 uses the distance sensor to determine the distance to each of the walls 200b to 200d located on a line extending radially at a predetermined angle (in this embodiment, every 0.5 degrees) with the vehicle 1 as a reference. 26a, and the shape of the parking space 200 can be calculated from the detection result. Therefore, the CPU 91 of the control device 200 is compared with the case where the shape of the parking space 200 is calculated by analyzing the image input by the camera pixel by pixel. Can reduce the processing load. Further, the calculation of the shape of the parking space 200 uses a normal trigonometric function as described above, and the calculation itself is simplified, so that the processing load on the CPU 91 can be further reduced.

  The present invention has been described above based on the embodiments. However, the present invention is not limited to the above embodiments, and various improvements and modifications can be made without departing from the spirit of the present invention. It can be easily guessed.

  For example, in the above embodiment, the distance sensor 26a is configured by a laser range finder, but may be configured by an ultrasonic sensor that transmits ultrasonic waves, a millimeter wave sensor that transmits millimeter waves, or the like. Even in the case of an ultrasonic sensor or a millimeter wave sensor, the processing load on the CPU 91 of the control device 200 can be reduced as compared with the case where the image input by the camera is analyzed to calculate the shape of the parking space 200.

  In the above embodiment, the parking space 200 is partitioned by the walls 200b to 200d. However, the parking space 200 may be partitioned by a wire mesh, or may be partitioned by a partition member or the like slightly protruding from the ground. Or, conversely, it may be defined by a dent provided on the ground. That is, as long as the object can be detected by the distance sensor 26a, any member may be used to partition the parking space 200. Note that when the parking space 200 is partitioned by a wire mesh, the wire sensor may not be positioned on a line extending radially every predetermined angle by the distance sensor 26a. Since an extremely wide distance between points can be invalidated, the shape of the parking space 200 can be accurately calculated even when the distance sensor 26a cannot extract all the points corresponding to the shape of the wire mesh.

  Moreover, in the said embodiment, although it comprised so that the parking space 200 might be divided by the parking area for 1 unit | set by the walls 200b-200d, it comprised so that it might be divided by the parking area for 2 units | sets of the vehicle 1 Also good. This is because the distance to the wall can be detected by the distance sensor 26a in the parking area for two vehicles 1. In this configuration, when one vehicle 1 is already parked, the parked vehicle 1 can also be calculated as a wall. Therefore, when parking by automatic driving, the vehicle 1 collides with the parked vehicle 1. It is possible to prevent the occurrence of harmful effects.

  In the above embodiment, the shape of the parking space 200 partitioned by the walls 200b to 200d is calculated. However, the shape of the parking space while the vehicle is parked on both sides is calculated. Also good. In this configuration, for example, since the vehicle can be applied to one parking space in a public parking lot by automatic driving, the parking function by automatic driving can be used effectively.

  In the above embodiment, the distance sensor 26a is attached to the right front of the vehicle 1. However, the distance sensor 26a may be attached to the left front, the right rear, or the left rear of the vehicle 1. It may be attached to the rear left and right two places, the four corners of the vehicle 1 and the like. That is, the number and location of the distance sensors 26a may be any number and location.

  Moreover, in the said embodiment, although the movement to the recognition position 201 of the vehicle 1 shall be performed by automatic driving | operation, you may comprise so that a driver | operator may drive and move to the parking space 200 vicinity. In this configuration, when the vehicle 1 reaches within the permissible range 201a of the recognition position 201, it may be configured to prompt the vehicle 1 to stop by voice or video.

  Further, in the above embodiment, the parking lot data of the parking lot 200, the position data of the parking space 200 where the vehicle 1 can be parked are stored in the HDD 96 in advance, but wirelessly from an external device at the entrance of the parking lot. You may comprise so that it may receive by communication and memorize | store in RAM95. In this configuration, since it is not necessary to store necessary data in the HDD 96 in advance, the storage area of the HDD 96 can be used effectively, and parking by automatic driving in another parking lot can be easily performed.

  In the above-described embodiment, the processing order in which the slope filter is applied after the length filter is applied in the filtering process is described, but the processing order in which the length filter is applied after the slope filter is applied may be adopted. Even if the processing order is different, it is possible to remove the dividing line segment generated when the distance sensor 26a detects a human figure or an object in the parking space 200, or the dividing line segment generated due to the influence of noise or the like. It can suppress that the virtual dividing line segment which divides the space 200 differs extremely from the shape of the actual parking space 200, and can recognize the shape of the parking space 200 correctly. Further, one of the length filter and the gradient filter may be applied.

  Moreover, in the said embodiment, although the vehicle 1 was back parking, it is good also as front parking and according to the position and surrounding environment of the parking space 204, you may comprise so that a parking method can be changed suitably.

1 Vehicle 26 Surrounding environment monitoring device (part of distance detection means)
26a Distance sensor (part of distance detection means)
100 Control device 200 Parking space 200b-200d Wall (object)
S104 Point extraction means S105 Part of generation means, point invalidation means, division means S107 Part of generation means, position calculation means S201 Line segment extraction means S202 Second line segment invalidation means S203 Line segment length calculation means, first line segment Invalid means

Claims (4)

In a control device mounted on a vehicle that can be moved by automatic driving in a predetermined parking space,
A distance detection means for detecting a distance to an object located on a line extending radially every predetermined angle with the vehicle as a reference; and
Point extraction means for extracting the distance to the object detected at predetermined angles by the distance detection means as a point in the coordinate system;
Generating means for generating a virtual line segment that divides the parking space, based on the points of the coordinate system extracted by the point extracting means;
And a position calculating means for calculating an estimated position where the vehicle is estimated to be located with respect to the parking space based on the points of the coordinate system extracted by the point extracting means. apparatus. The generating means includes
When the first point extracted by the point extraction means is used as a reference, a point-to-point distance with the second point extracted immediately before or after the first point is calculated, and the calculation is performed. A point invalidating means for invalidating at least one of the first and second points when the distance between the points is not less than a predetermined value;
2. The control according to claim 1, wherein a virtual line segment that divides the parking space is generated by virtually connecting points other than the points invalidated by the point invalidating means. apparatus. A dividing unit that divides a line segment that is virtually connected by the generating unit and generates a divided line segment for each portion that is substantially a straight line;
Of the end points of the dividing line segment divided by the dividing means, a line segment extracting means for extracting a dividing line segment having a point with the shortest distance from the vehicle as a reference line segment;
A dividing line segment having a certain angle or more with respect to the reference line segment extracted by the line segment extracting means, and a dividing line having a certain angle or more with respect to the orthogonal line segment orthogonal to the reference line segment 3. The control apparatus according to claim 2, further comprising first line segment invalidating means for invalidating the minutes. A dividing unit that generates a dividing line segment by dividing the dividing line segment that is virtually connected by the generating unit into portions that are substantially straight;
Line segment length calculating means for calculating the length of the divided line segments divided by the dividing means;
And a second line segment invalidating unit that invalidates the divided line segment when the length of the divided line segment calculated by the line segment length calculating unit is equal to or less than a predetermined length. The control device according to claim 2 or 3.

Images