WO2020168769A1 - 垂直泊车方法 - Google Patents

垂直泊车方法 Download PDF

Info

Publication number
WO2020168769A1
WO2020168769A1 PCT/CN2019/120799 CN2019120799W WO2020168769A1 WO 2020168769 A1 WO2020168769 A1 WO 2020168769A1 CN 2019120799 W CN2019120799 W CN 2019120799W WO 2020168769 A1 WO2020168769 A1 WO 2020168769A1
Authority
WO
WIPO (PCT)
Prior art keywords
parking space
vehicle
vertical parking
coordinate information
corner
Prior art date
Application number
PCT/CN2019/120799
Other languages
English (en)
French (fr)
Inventor
胡坤福
罗作煌
范晓娟
史云鹏
Original Assignee
惠州市德赛西威智能交通技术研究院有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 惠州市德赛西威智能交通技术研究院有限公司 filed Critical 惠州市德赛西威智能交通技术研究院有限公司
Priority to EP19915876.7A priority Critical patent/EP3929046B1/en
Publication of WO2020168769A1 publication Critical patent/WO2020168769A1/zh

Links

Images

Classifications

    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B62—LAND VEHICLES FOR TRAVELLING OTHERWISE THAN ON RAILS
    • B62D—MOTOR VEHICLES; TRAILERS
    • B62D15/00—Steering not otherwise provided for
    • B62D15/02—Steering position indicators ; Steering position determination; Steering aids
    • B62D15/027—Parking aids, e.g. instruction means
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B62—LAND VEHICLES FOR TRAVELLING OTHERWISE THAN ON RAILS
    • B62D—MOTOR VEHICLES; TRAILERS
    • B62D15/00—Steering not otherwise provided for
    • B62D15/02—Steering position indicators ; Steering position determination; Steering aids
    • B62D15/027—Parking aids, e.g. instruction means
    • B62D15/0285—Parking performed automatically

Definitions

  • the invention relates to a parking method, in particular to a vertical parking method.
  • the object of the present invention is to provide a vertical parking method with a small error in the estimated parking space during parking.
  • a vertical parking method in the process of using sensors to detect or obtain from a third party the coordinate information of an initial vertical parking space and the outer boundary of the initial vertical parking space for parking, defining a location in the initial vertical parking A starting reference line other than the short side of the entrance of the position, the shortest distance from each point on the starting reference line to the short side of the entrance is greater than the trigger threshold of the detection distance of the vehicle angle ultrasonic sensor. And define the angle of the vehicle that will first cross the starting reference line during the parking process as the first angle.
  • the vertical parking method includes:
  • Step 1 When the coordinate information of a preset position of the vehicle close to the first corner is located on the starting reference line or on the side of the starting reference line close to the initial vertical parking space, when the When the ultrasonic sensor in the first corner of the vehicle detects an obstacle, locate the obstacle and perform step 2;
  • Step 2 Calculate the coordinate information of the point O where the obstacle is most likely to contact the vehicle. According to the coordinate information of the point O, the first car angle, and the outer boundary of the initial vertical parking space, calculate the target vertical parking space for a relocation Coordinate information of the outer boundary; and
  • Step 3 Continue to park according to the coordinate information of the outer boundary of the target vertical parking space.
  • step 2 the method for calculating the coordinate information of the outer boundary of the target vertical parking space includes:
  • the initial vertical parking space is translated in a direction away from the obstacle along the short side of the entrance, so that the corner point of the target vertical parking space close to the first corner is located at the intersection point, thereby obtaining the outer boundary of the target vertical parking space Coordinate information.
  • step 2 the step of calculating the coordinate information of the point O where the obstacle is most likely to contact the vehicle includes:
  • step 1 if the obstacle cannot be located, perform step 4: detecting the obstacle distance Dist, the first car angle, and the initial vertical parking space close to the first car angle according to the ultrasonic sensor of the first car angle.
  • step 4 detecting the obstacle distance Dist, the first car angle, and the initial vertical parking space close to the first car angle according to the ultrasonic sensor of the first car angle.
  • the method of calculating the coordinate information of the outer boundary of the target vertical parking space in the step 4 is: calculating the offset Delt of the target vertical parking space relative to the initial vertical parking space in the Y axis direction :
  • the X-axis coordinate value and the Y-axis coordinate value of the corner point B of the initial vertical parking space, Pd_x and Pd_y are the X-axis coordinate value and the Y-axis coordinate value of the first car angle, respectively; and the initial vertical parking space is short along the entrance
  • the offset Delt is translated in the direction away from the obstacle, so as to obtain the coordinate information of the outer boundary of the target vertical parking space.
  • an intermediate reference line is also defined, which is located between the short side of the entrance and the short side of the bottom of the initial vertical parking space
  • the vertical parking method further includes: when the coordinate information of the preset position of the vehicle is located The distance between the intermediate reference line and the short side of the bottom of the initial vertical parking space and the first angle of the vehicle or the second angle opposite to the ultrasonic sensor for detecting obstacles on both sides of the vehicle When they are DistR, DistL, and the first corner is the right rear corner and the second corner is the left rear corner, perform step 5: Determine the vehicle based on DistR, DistL and the trigger threshold DistTh_Up of the vehicle's corner ultrasonic sensor detection distance Whether there is enough space on at least one side for people to get on and off the car, and according to the judgment result, output a reminder that the parking space cannot be parked or calculate the coordinate information of the outer boundary of the target vertical parking space, and perform step 3 to make the car after parking is finished At least one side can be used for getting off.
  • step 5 the method for judging whether there is enough space on at least one side of the vehicle for people to get on and off the vehicle includes:
  • DistTh_Up is the trigger threshold of the detection distance of the vehicle's angle ultrasonic sensor
  • the preset position is the midpoint of the front edge of the vehicle; when the first vehicle angle is the rear angle of the vehicle, the preset position is The midpoint of the boundary at the rear of the vehicle.
  • an end reference line close to the bottom short side of the initial vertical parking space is also defined; when it is determined that the preset position reaches the end reference line, parking ends.
  • the middle reference line is 2.3 to 2.6 m from the short side of the entrance, and the distance between the end reference line and the middle reference line is 1.8 to 2.2 m.
  • the vertical parking method of the present invention continuously locates the surrounding obstacles through the vehicle ultrasonic sensor during the parking process, and adjusts the coordinates of the vertical parking space in real time according to whether the obstacle is detected and the distance from the obstacle to eliminate the original parking space. Estimate the error and improve the accuracy of parking the vehicle. It can also avoid the situation that passengers cannot get off normally due to the small distance between the vehicle and the left or right reference vehicle after the vehicle is parked, and it can also avoid the parking error caused by the parking error at the last step of parking the vehicle. When the vehicle is traveling, the brake is interrupted due to an obstacle encountered, which affects the user experience.
  • FIG. 1 is a schematic diagram of the positional relationship among the large coordinate system x-o-y, the vehicle, the initial vertical parking space, and the reference line of the medium vertical parking method according to an embodiment.
  • Fig. 2 is a schematic diagram of the layout of an ultrasonic sensor at the rear of a vehicle according to an embodiment.
  • Fig. 3 is a flowchart of an embodiment of a vertical parking method.
  • Fig. 4 is a schematic diagram of parking space relocation in a vertical parking method according to an embodiment.
  • Fig. 5 is a partial enlarged view of Fig. 4.
  • Fig. 6 is a schematic diagram of relocating parking spaces in area D2 in an embodiment.
  • the vehicle to which the method of the present invention can be applied should include an automatic or semi-autonomous driving system that can drive the vehicle with simplified driver input, including sensors for collecting information about the surrounding environment of the vehicle, such as acquiring environmental images, vehicle speed, surrounding obstacles, and temperature Based on the collected information, the driving system performs functions such as deceleration, steering, acceleration, braking, pedestrian recognition, parking space recognition, and automatic or semi-automatic parking.
  • the above-mentioned automatic or semi-automatic parking refers to an automatic or semi-automatic driving state.
  • the vehicle controls the movement of the vehicle (for example, steering, turning, acceleration, etc.) without direct driver input. Slow down, etc.) to park the vehicle in the parking space.
  • the vehicle may require the driver to enter certain confirmation passwords in some cases to continue to perform some unconventional movement controls of the vehicle to park the vehicle in the parking space.
  • the four corners of the vehicle are equipped with ultrasonic sensors.
  • repositioning and vertical parking of the vertical parking space based on the signal output by the ultrasonic sensor can be realized.
  • “Vertical parking” means that the parking entrance of the target parking space is located on the short side of the parking space, and adjacent to the long side of the target parking space may be other parking spaces or obstacles such as walls or pillars.
  • the parking space is called a vertical parking space.
  • the vertical parking method of the present invention allows the vehicle to be parked in a forward or reverse direction.
  • the ultrasonic sensors at the two front corners of the front of the vehicle are mainly used.
  • the ultrasonic sensors at the rear corners of the vehicle are mainly used.
  • the "front corner” of a vehicle refers to the part of the vehicle where the front surface (front boundary) and side surface (side boundary of the vehicle body) of the vehicle converge. It is usually a curved surface connecting the front and side surfaces of the vehicle, including the left front corner and the right front angle.
  • the "rear corner” of a vehicle refers to the part of the vehicle where the rear surface (rear boundary) and the side surface of the vehicle converge.
  • the vehicle is usually represented as a rectangular top view state (as shown in Figure 1) or a rectangular top view state with certain rounded corners.
  • this article takes the vehicle represented as a rectangular top view state as an example, and the front left corner, front right corner, rear left corner, and rear right corner of the vehicle are marked as Pb, Pa, Pc, and Pd, respectively.
  • the vehicle When in an automatic or semi-autonomous driving system, the vehicle is represented as a rectangular top view state with certain rounded corners, a point in the rounded corners should be defaulted as a corner point, and the same as the points Pb, Pa, Pc and Pd in Figure 1 Function.
  • the midpoint of the boundary of the rear of the vehicle is marked as M1
  • the midpoint of the boundary of the front of the vehicle is marked as M2.
  • ultrasonic sensors are provided at each corner of the vehicle.
  • other sensor combinations or integrated sensors with similar functions may be used.
  • Figure 2 Take the left and right rear corners as examples (the left and right front corners are similar to the left and right rear corners, and are basically symmetrical), the left rear corner is set with ultrasonic sensors SR_RL, LR_RL and SR_RML, right The rear angle is set with ultrasonic sensors SR_RR, LR_RR and SR_RMR.
  • Adjacent ultrasonic sensors can form a common detection FOV (Field of View), and at the same time, based on the ultrasonic single-shot and multiple-receive mechanism, the triangulation method is used to locate obstacles.
  • FOV Field of View
  • the ultrasonic sensors SR_RL and SR_RR are set at the corners closer to the sides of the vehicle, and are mainly used to detect obstacles on both sides of the rear of the vehicle.
  • Ultrasonic sensors SR_RML and SR_RMR are set at the midpoint M1 at the corner closer to the rear edge of the vehicle, and are mainly used to detect obstacles on the rear side of the vehicle.
  • Ultrasonic sensors LR_RL and LR_RR are close to the points Pc and Pd in Figure 1, respectively. The obstacle facing the rear corner of the rover.
  • the vehicle In the process of vertical parking, the vehicle generally enters the parking space in a state almost parallel to the initial vertical parking space, then the positional relationship between the large coordinate system xoy, the vehicle and the initial vertical parking space is generally shown in Figure 1 (in different scenarios In the middle, only the vehicle is inclined to the left or right, or the front or rear is entered), the X axis of the coordinate system is roughly parallel (substantially or visually parallel) to the long side of the initial vertical parking space, and the deviation is generally not more than 6 °.
  • the automatic or semi-autonomous driving system of the vehicle should use sensors to detect a four corners marked with A, B, C, and D, and the outer boundary (the connecting points A, B, C, and D in the figure are solid-line boxes).
  • the coordinates of the corner points Pb, Pa, Pc and Pd of the vehicle are (Pb_x, Pb_y), (Pa_x, Pa_y), (Pc_x, Pc_y) and (Pd_x, Pd_y), and the midpoint of the vehicle rear boundary is M1
  • the coordinate information of is (M1_x, M1_y, Theta)
  • the coordinate information of the midpoint M2 of the front of the vehicle is (M2_x, M2_y, Theta).
  • all the coordinate information following "_x" is the X-axis coordinate value
  • all the coordinate information following "_y” is the Y-axis coordinate value.
  • Theta is the angle formed by the vehicle direction (M1-M2) and the x-axis direction.
  • the reference lines L1, L2 and L3 are all parallel to the short sides AB (entry short side) and CD (bottom short side) of the initial vertical parking space. ), and the starting reference line L1 is located outside the short side AB of the entrance of the initial vertical parking space, and the distance from the short side AB of the entrance is 0.4 ⁇ 0.6m.
  • This distance is related to the trigger range of the ultrasonic sensor and should be greater than the ultrasonic sensor detection
  • the trigger threshold of the distance is 0.5 meters in this embodiment.
  • the middle reference line L2 and the end reference line L3 are both located in the initial vertical parking space, and the middle reference line L2 is located between the entrance short side AB and the end reference line L3, and the end reference line L3 is located between the middle reference line L2 and the short side CD .
  • the distance between the middle reference line L2 and the starting reference line L1 is 2.3 to 2.6 m, in this embodiment it is 2.5 meters, and the distance between the end reference line L3 and the middle reference line L2 is 1.8 to 2.2 m, in this embodiment
  • the middle is 2 meters.
  • the start reference line L1, the middle reference line L2 and the extension of the long sides AD and BC of the initial vertical parking space is marked as D1
  • the area enclosed by sides AD and BC is marked as D2.
  • the start reference line L1, the middle reference line L2, and the end reference line L3 may be all or part of a curve (non-straight line).
  • the starting reference line is a curve, it is located outside the short side of the entrance of the initial vertical parking space, and the shortest distance from each point on the starting reference line to the short side of the entrance (the length of the vertical line segment) is greater than the vehicle angle ultrasonic The trigger threshold of the sensor detection distance.
  • the middle reference line is used to trigger the second parking space relocation state. At this time, the vehicle and the parking space are basically parallel, so the middle reference line is substantially or roughly straight.
  • the corner of the vehicle that first crosses the starting reference line L1 during parking is set as the right rear corner Pd (first corner).
  • the vertical parking method of the present invention includes the steps:
  • S001 Enter the parking state and determine that it is a reverse parking state, and obtain coordinate information of the initial vertical parking space and the outer boundary of the initial vertical parking space;
  • step S003 when the detection obstacle distance Dist of the ultrasonic sensors SR_RR and LR_RR of the first vehicle angle (the right rear corner Pd in this embodiment) is less than the threshold DistTh_Up, that is, Dist falls within the sensor detection blind zone threshold DistTh_Down and the detection distance trigger threshold When between DistTh_Up (Dist ⁇ [DistTh_Down,DistTh_Up]), it is judged that the vehicle is likely to hit an obstacle in the initial vertical parking space, and step S004 is executed, otherwise, step S001 is returned.
  • the blind zone threshold DistTh_Down is 0.3m (when it is less than or equal to 0.3m, it is the detection blind zone of the ultrasonic sensor, and the output remains unchanged at 0.3m); the trigger threshold DistTh_Up of the ultrasonic detection distance is 0.4m.
  • step S004 based on the common detection FOV (Field of View) of the ultrasonic sensors SR_RR and LR_RR and the single-shot and multiple-receive mechanism of ultrasound, use triangulation to locate obstacles. If the ultrasonic detection distance is Dist ⁇ [DistTh_Down,DistTh_Up], it can be located When there is an obstacle, step S005 is executed, otherwise, step S013 is executed.
  • FOV Field of View
  • step S008 Determine the size of Var. If Var is greater than the preset value VarTh (Var>VarTh), it is considered that the point set W does not meet the calculation requirement, and step S001 is executed again; otherwise, step S009 is executed.
  • the preset value VarTh can be selected from [0.01, 0.03], preferably 0.01.
  • steps S005 to S009 essentially calculate the coordinate information of the point O where the obstacle is most likely to contact the vehicle. It is understandable that in other embodiments, other coordinate information that can determine the point O most likely to contact the vehicle may be used.
  • S010 Calculate the slope K 1 of a long side of the initial vertical parking space close to the obstacle.
  • Fig. 5 is an enlarged view of the part enclosed by the dashed circle 22 in Fig. 4, and the process of forming the intersection Cr can be seen.
  • the initial vertical parking space can be translated along the short side of the entrance to the direction away from the obstacle by the Y-axis offset.
  • steps S010 and S011 are a method for calculating the parking space offset under the condition that the coordinate information of point O can be obtained. The essence is to first obtain the first vehicle that passes through the first corner (right rear corner Pd) and has a slope of K1.
  • the ultrasonic sensor of the first car angle detects the obstacle distance Dist, the first car angle, and the corner point of the short side of the entrance of the initial vertical parking space close to the first car angle. Coordinate information, calculate the coordinate information of the outer boundary of the target vertical parking space.
  • Dist the coordinate values of the right rear corner Pd (first corner) and corner point B, Dist is taken as the length of the hypotenuse of the right-angled triangle, and the difference between the right rear corner Pd and the corner point B in the X-axis direction The value is taken as the length of the right side of the right triangle parallel to the X axis, find the length of the right side of the right triangle parallel to the Y axis, and use this length as the right rear corner Pd and the corner point B'of the target vertical parking space. The difference in the Y-axis direction.
  • Step S015 Define the obstacle detection distance of the ultrasonic sensor SR_RR in the right rear corner of the vehicle as DistR, and the obstacle detection distance of the ultrasonic sensor SR_RL in the left rear corner as DistL, then determine whether (DistL+DistR) is greater than 2*DistTh_Up, if so, Step S016 is executed, otherwise, step S021 is executed.
  • step S016 Determine whether DistL>DistTh_Up and DistR>DistTh_Up, if yes, it is determined that the parking can be continued, and step S001 is executed, otherwise, step S017 is executed.
  • step S017 Determine whether DistL>DistTh_Up and DistR ⁇ DistTh_Up, if yes, determine that there is an obstacle on the right side of the vehicle (please refer to the fan-shaped ultrasonic radiation range in FIG. 6), then perform step S018, otherwise, perform step S019.
  • step S019 Determine whether DistL ⁇ DistTh_Up and DistR>DistTh_Up, if yes, determine that there is an obstacle on the left side of the vehicle, and then perform step S020, otherwise, perform step S001.
  • S021 Determine that parking space is not allowed and output a reminder message.
  • Steps S015 to S021 essentially determine whether there is enough space on at least one side of the vehicle for people to get on and off the vehicle based on DistR, DistL, and the trigger threshold DistTh_Up of the vehicle's angle ultrasonic sensor detection distance, and output a reminder that parking spaces cannot be parked according to the judgment results Information or calculation of the coordinate information of the outer boundary of the target vertical parking space, so that at least one side of the vehicle is available for people to get off after the parking is completed. It is understandable that other preset values can be used instead of DistTh_Up to make judgments. For example, different values can be used according to the vehicle sales area, taking into account different races, different body types, or local vehicle standard parking spaces. As long as the result of the judgment indicates that there is enough space on at least one side of the vehicle for people to get on and off the vehicle.
  • step S200 Determine whether the coordinate information of the midpoint M1 of the vehicle rear boundary reaches the end reference line L3, and if so, the parking ends, otherwise, step S001 is executed.
  • step S200 of other embodiments it is also possible to determine whether the parking is completed by determining whether the front of the vehicle enters the short side of the parking space entrance. At this time, there is no need to determine the positional relationship between the point M1 and the end reference line L3, and the end reference line L3 can be cancel.
  • the parking end is determined by judging that the front point M2 or all points of Pa and Pb are within the short side of the parking space entrance, or the front point Pa and Pb are within the short side of the parking space entrance and the distance from the short side of the entrance is greater than 0.1 m is the end of parking.
  • the real parking space position can be predicted by the obstacle positioning method, and a real and effective parking space can be obtained, which can improve the parking accuracy and user experience;
  • the vehicle constantly relocates the parking space according to the surrounding obstacles, eliminating the initial parking space estimation error, and avoiding the vehicle parking due to the distance between the vehicle and the left or right reference vehicle It is smaller, which will cause the passengers to be unable to get off the car normally, and it can also avoid the situation where the vehicle is stopped due to obstacles when the vehicle is running due to the parking error at the last step of parking the vehicle, which affects the user experience .
  • steps S016 and S017 can be reversed.
  • the position of the initial vertical parking space (the coordinate information of the outer boundary) can be obtained by a third party, and converted into coordinates in the large coordinate system x-o-y by the driving system of the vehicle.
  • the final coordinate information of the target vertical parking space is calculated directly along the entrance of the initial vertical parking space.
  • the short side is translated in the direction away from the obstacle. It is understandable that when the offset is calculated, other algorithms can be used to obtain the final coordinates of the target vertical parking space.
  • the end reference line L3 can be omitted, that is, the solution of the present invention can be implemented without defining the end reference line L3.
  • step S002 as long as a preset point on the vehicle is located on the central axis of the vehicle (vertical to the central axis of the front and rear axles of the vehicle) and close to the front or rear of the vehicle,
  • the coordinate information of is located on the initial reference line L1 or has entered the area between the initial reference line L1 and the middle reference line L2 to execute.
  • the preset point is not the midpoint of the front boundary of the vehicle and the midpoint of the rear boundary of the vehicle, the position of the starting reference line and the middle reference line relative to the short side of the entrance should be modified in contrast.
  • the angle of the vehicle that first crosses the starting reference line L1 during parking is set as the right rear angle Pd (as the first angle). It is understandable that when the first corner is the left rear corner or the left front corner or the right front corner, the corner corresponding to the real "first corner" in the vertical parking method should be adjusted accordingly and close to the "first corner""Or the corresponding corner point of the initial vertical parking space and the target vertical parking space.
  • step S003 modify the "determining the distance between the ultrasonic sensors SR_RR and LR_RR of the right rear corner Pd (first car corner) to detect obstacles" in step S003 to "determine the left rear corner Pc (The first angle of the car) ultrasonic sensors SR_RL and LR_RL detection obstacle distance”; modify the "ultrasonic sensors SR_RR and LR_RR” in step S004 to “ultrasonic sensors SR_RL and LR_RL”; calculate the long side BC in step S010 Slope, amended to "calculate the slope of the long side AD", and "pass through the right rear corner Pd (first corner) and slope -- amended to "pass through the left rear corner Pc (first corner) and slope Is";
  • step S011 "especially the coordinate information (Pd_x, Pd_y) of the right rear corner Pd (first corner)” is modified to "especially the coordinate information of the left rear corner Pc (first corner) ( Pc_x,
  • first and second are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of indicated technical features. Thus, the features defined with “first” and “second” may explicitly or implicitly include one or more of these features. In the description of the present invention, “plurality” means two or more than two, unless specifically defined otherwise.

Landscapes

  • Engineering & Computer Science (AREA)
  • Chemical & Material Sciences (AREA)
  • Combustion & Propulsion (AREA)
  • Transportation (AREA)
  • Mechanical Engineering (AREA)
  • Traffic Control Systems (AREA)
  • Measurement Of Velocity Or Position Using Acoustic Or Ultrasonic Waves (AREA)

Abstract

一种垂直泊车方法，包括：步骤一，当车辆的靠近所述第一车角的一个预设位置的坐标信息位于所述起始参考线上或位于所述起始参考线的靠近初始垂直停车位一侧，且所述第一车角的超声波传感器检测到障碍物时，定位障碍物，执行步骤二；步骤二，计算所述障碍物最有可能与车辆接触的点O的坐标信息，根据点O、第一车角以及初始垂直停车位的外边界的坐标信息，计算目标垂直停车位的外边界的坐标信息；以及步骤三，根据所述目标垂直停车位的外边界的坐标信息继续进行自动或半自动泊车。本发明的垂直泊车方法可消除原始泊车位的估计误差，提高车辆泊车入位的准确率。

Description

垂直泊车方法 技术领域
本发明涉及泊车方法，特别涉及一种垂直泊车方法。
背景技术
随着智能驾驶技术的发展，自动泊车技术的研发也高速发展，例如专利申请号为201810031452.4，申请日为2018年1月12日的发明专利申请《车辆在垂直停车位中的自主泊车》就公开了一种自动泊车技术。现有的机动车大多都配备了超声波传感器进行障碍物检测，帮助驾驶员消除盲点和视力模糊，提高驾驶安全性。由于超声波传感器经济实惠且多功能，现有的自动泊车技术多采用以超声波进行自动泊车位搜索。然而，受限于超声波传感器本身的特性，以及声波传播对***环境强依懒性的特点，使得在利用超声波进行车位搜索时，往往会产生较大的车位位置估计误差，同时叠加车辆定位可能产生的偏离泊车预设轨迹误差，从而大大降低针对多复杂工况下泊车入位的成功率。
发明内容
本发明的目的在于提供一种泊车时车位预估误差小的垂直泊车方法。
一种垂直泊车方法，在利用传感器检测或从第三方获得一初始垂直停车位和所述初始垂直停车位的外边界的坐标信息以进行泊车的过程中，定义一位于所述初始垂直停车位的入口短边之外的起始参考线，所述起始参考线上的各点到所述入口短边的最短距离大于车辆的车角超声波传感器探测距离的触发阈值。并定义在泊车的过程中会最先越过所述起始参考线的车辆的车角为第一车角。所述垂直泊车方法包括：
步骤一，当车辆的靠近所述第一车角的一个预设位置的坐标信息位于所述起始参考线上或位于所述起始参考线的靠近初始垂直停车位一侧时，当所述第一 车角的超声波传感器检测到障碍物时，定位障碍物，执行步骤二；
步骤二，计算所述障碍物最有可能与车辆接触的点O的坐标信息，根据点O、第一车角以及初始垂直停车位的外边界的坐标信息，计算一重定位的目标垂直停车位的外边界的坐标信息；以及
步骤三，根据所述目标垂直停车位的外边界的坐标信息继续进行泊车。
作为一种实施方式，步骤二中，所述计算目标垂直停车位的外边界的坐标信息的方法包括：
计算所述初始垂直停车位的长边的斜率K1；
获得穿过所述第一车角且斜率为K1的第二直线；
计算第二直线与初始垂直停车位的入口短边的交点的坐标信息；以及
将所述初始垂直停车位沿入口短边向远离障碍物的方向平移，使得目标垂直停车位的靠近所述第一车角的角点位于所述交点，从而获得目标垂直停车位的外边界的坐标信息。
作为一种实施方式，步骤二中，所述计算所述障碍物最有可能与车辆接触的点O的坐标信息的步骤包括：
按照时间以滑动窗口的形式记录数据长度为N的障碍物坐标点集W＝{(obj_x(n),obj_y(n))|N＝1,2,3,...N}，N为自然整数，obj_x(n)为障碍物X轴坐标值，obj_y(n)为障碍物Y轴坐标值；
对所述点集W进行最小二乘直线拟合，获得直线L：y＝K*x+b；
求解所述点集W中的各点到直线L距离的均方根Var：
Figure PCTCN2019120799-appb-000001
判断Var大小，若Var大于预设值，则认为点集W不满足计算要求，则重新执行步骤一，否则，确定点O的横坐标O_x＝obj_x(N)，纵坐标O_y＝obj_y(N)。
优选的，所述步骤一中，若无法定位到障碍物，则执行步骤四：根据第一车角的超声波传感器探测障碍物距离Dist、第一车角以及靠近第一车角的初始垂直停车位的入口短边的角点的坐标信息，计算目标垂直停车位的外边界的坐标信息，然后执行步骤三。
作为一种实施方式，所述步骤四中所述计算目标垂直停车位的外边界的坐标信息的方法为：计算所述目标垂直停车位相对于初始垂直停车位的在Y轴方向的偏移量Delt：|B_y–Pd_y|-sqrt(Dist 2-|B_x–Pd_x| 2)，其中Y轴大体平行于所述初始垂直停车位的入口短边，B_x和B_y分别为靠近所述第一车角的初始垂直停车位的角点B的X轴坐标值和Y轴坐标值，Pd_x和Pd_y分别为第一车角的X轴坐标值和Y轴坐标值；以及将所述初始垂直停车位沿入口短边向远离障碍物的方向平移所述偏移量Delt，从而获得目标垂直停车位的外边界的坐标信息。
优选的，还定义一中间参考线，位于所述初始垂直停车位的入口短边和底部短边之间，所述的垂直泊车方法进一步包括：当车辆的所述预设位置的坐标信息位于所述中间参考线与初始垂直停车位的底部短边之间，且车辆的第一车角或与其相对的第二车角的用于探测车辆两侧的障碍物的超声波传感器的探测障碍物距离分别为DistR，DistL，且第一车角为右后角，第二车角为左后角时，则执行步骤五：根据DistR、DistL和车辆的车角超声波传感器探测距离的触发阈值DistTh_Up判断车辆至少一侧是否有足够的位置可供人员上下车，并根据判断结果输出不可泊入车位的提醒信息或计算目标垂直停车位的外边界的坐标信息，执行步骤三，以使得泊车结束后车辆至少一侧可供人员下车。
作为一种实施方式，步骤五中，所述判断车辆至少一侧是否有足够的位置可供人员上下车的方法包括：
(1)当(DistL+DistR)<2*DistTh_Up，则确定不可泊入车位并输出提醒信 息，其中DistTh_Up为车辆的车角超声波传感器探测距离的触发阈值；
(2)如果(DistL+DistR)≥2*DistTh_Up时：
(a)如果DistL>DistTh_Up且DistR>DistTh_Up，则确定可继续泊车；
(b)如果DistL>DistTh_Up且DistR≤DistTh_Up，则确定车辆右侧存在障碍物，则车位调整偏移量为Delt＝DistTh_Up–DistR，并计算目标垂直停车位的外边界的坐标信息，执行步骤三；
(c)如果DistL≤DistTh_Up且DistR>DistTh_Up，则认为左侧碰到障碍物，则确定需要调整的偏移量为Delt＝DistTh_Up–DistL，并计算目标垂直停车位的外边界的坐标信息，执行步骤三。
优选的，当所述第一车角为车辆前角时，所述预设位置为所述车辆车头边界的中点；当所述第一车角为车辆后角时，所述预设位置为所述车辆车尾边界的中点。
作为一种实施方式，还定义一靠近所述初始垂直停车位的底部短边的末端参考线；当判断所述预设位置到达末端参考线时，泊车结束。
作为一种实施方式，当判定车身全部进入车位入口短边之内则泊车结束。
作为一种实施方式，所述中间参考线距离所述入口短边2.3～2.6m米，所述末端参考线与中间参考线之间的距离为1.8～2.2m。
本发明的垂直泊车方法是在泊车的过程中通过车辆超声波传感器持续定位周围障碍物，并根据是否检测到障碍物、与障碍物的距离实时调整垂直停车位的坐标，消除原始泊车位的估计误差，提高车辆泊车入位的准确率。还可避免车辆泊车入位后，由于车辆与左侧或右侧参考车辆距离较小，从而导致乘客无法正常下车的情况，也可避免在车辆泊车入位的最后一步由于车位误差使得车辆行进过程中，因遇到到障碍物而中断刹停，影响用户体验的情况。
附图说明
图1为一实施例的中垂直泊车方法的大坐标系x-o-y、车辆、初始垂直停车位及参考线的位置关系示意图。
图2为一实施例的车辆车尾的超声波传感器的布局示意图。
图3为一实施例垂直泊车方法流程图。
图4为一实施例的垂直泊车方法中车位重定位示意图。
图5为图4的局部放大图。
图6为一实施例中在D2区域进行车位重定位的示意图。
具体实施方式
下面将结合具体实施例及附图对本发明垂直泊车方法和及作进一步详细描述。
可应用本发明方法的车辆应包括能够以简化的驾驶员输入来驾驶车辆的自动或半自动驾驶***，包括用于收集车辆周围环境的信息的传感器，例如获取环境图像、车速、周围障碍物、温度、湿度等的传感器，驾驶***基于收集到的信息来执行例如减速、转向、加速、制动、识别行人、识别停车位和自动或半自动泊车等功能。上述的自动或半自动泊车指的是一种自动或半自动驾驶状态，在自动驾驶状态下，车辆在不需要直接的驾驶员输入的情况下控制车辆的运动动作(例如，转向、转弯、加速、减速等)以将车辆停放在停车位中。在半自动驾驶状态下，车辆在某些情况下可能需要驾驶员输入某些确认口令来继续执行一些非常规的控制车辆的运动动作以将车辆停放在停车位中。
本发明中，车辆四角装载有超声波传感器，在进行垂直停车位泊车时，可实现基于超声波传感器输出的信号进行垂直泊车位的重定位和垂直泊车。“垂直泊车”指的是目标停车位的泊车入口位于停车位的短边，与目标停车位的长 边相邻的可能是其他停车位或墙、柱等障碍物。在此，为避免混淆，该停车位被称为垂直停车位。
进一步地，本发明的垂直泊车方法和允许车辆正向前进或倒向行驶的方向来进行泊车。在车辆正向前进进行泊车时，主要使用车辆车头两个前角处的超声波传感器。在车辆倒向行驶进行泊车时，主要使用车辆车尾两个后角处的超声波传感器。车辆的“前角”指的是车辆的前表面(车头边界)和侧表面(车身侧边界)会聚到的车辆部分，通常为连接车辆的前表面和侧表面的弯曲表面，包括左前角和右前角。车辆的“后角”指的是车辆的后表面(车尾边界)和侧表面会聚到的车辆部分，通常为连接车辆的后表面和侧表面的弯曲表面，包括左后角和右后角。在自动或半自动驾驶***中，车辆通常被表示为直角长方形俯视状态(如图1所示)或具有一定圆角的长方形俯视状态。为方便描述，本文以车辆被表示为直角长方形俯视状态为例进行说明，且车辆左前角、右前角、左后角和右后角分别被标注为Pb，Pa，Pc和Pd。当在自动或半自动驾驶***中，车辆被表示为具有一定圆角的长方形俯视状态时，应默认圆角中的一点为角点，并与图1中的点Pb，Pa，Pc和Pd具有同样的功能。此外，车辆车尾边界的中点被标注为M1，车辆车头边界的中点被标注为M2。
本实施例中，车辆每个角处均设置三颗超声波传感器，其他实施例中可采用功能类似的其他传感器组合或集成传感器。请参考图2，以左后角和右后角处为例(左前角和右前角与左后角和右后角相似，并基本对称)，左后角设置超声波传感器SR_RL、LR_RL和SR_RML，右后角设置超声波传感器SR_RR、LR_RR和SR_RMR。相邻的超声波传感器可形成公共探测FOV(Field of View)，同时基于超声波的单发多收机制，利用三角定位法对障碍物进行定位。其中超声波传感器SR_RL和SR_RR设置在角部更靠近车身两侧处，分别主要用于探测车 尾两侧的障碍物。超声波传感器SR_RML和SR_RMR设置在角部更靠近车尾边界的中点M1，分别主要用于探测车尾后侧的障碍物，超声波传感器LR_RL和LR_RR靠近图1中的点Pc和Pd，分别主要用于探测车尾角部正对的障碍物。
在进行垂直泊车的过程中，车辆一般以近乎平行于初始垂直停车位的状态进入车位，则大坐标系x-o-y、车辆及初始垂直停车位的位置关系一般如图1所示(在不同的情景中，仅仅车辆左倾斜或右倾斜或车头进入还是车尾进入的不同)，坐标系的X轴大体平行(实质上或视觉上基本平行)于初始垂直停车位的长边，一般偏差不超过6°。在泊车入位之前，车辆的自动或半自动驾驶***应利用传感器检测得到一个四角用A、B、C和D标记、外边界(图中连接点A、B、C和D的为实线框)如图1所示的初始垂直停车位，并检测到四角在大坐标系x-o-y下的坐标信息分别为(A_x,A_y)，(B_x,B_y)，(C_x,C_y)，(D_x,D_y)。此时，车辆的各角点Pb，Pa，Pc和Pd的坐标为(Pb_x,Pb_y)，(Pa_x,Pa_y)，(Pc_x,Pc_y)和(Pd_x,Pd_y)，车辆车尾边界的中点M1的坐标信息为(M1_x,M1_y,Theta)，车辆车头边界的中点M2的坐标信息为(M2_x,M2_y,Theta)。其中，上述坐标信息中，凡是跟着“_x”的坐标信息均为X轴坐标值，凡是跟着“_y”的坐标信息均为Y轴坐标值。Theta为车辆方向(M1-M2)与x轴方向所成的夹角。此外，定义一起始参考线L1、一中间参考线L2和一末端参考线L3，参考线L1、L2和L3均平行于初始垂直停车位的短边AB(入口短边)和CD(底部短边)，且起始参考线L1位于初始垂直停车位的入口短边AB之外，并距离入口短边AB的距离为0.4～0.6m，该距离与超声波传感器的触发范围有关，应大于超声波传感器探测距离的触发阈值，本实施例中为0.5米。中间参考线L2和末端参考线L3均位于初始垂直停车位内，且中间参考线L2位于入口短边AB和末端参考线L3之间，末端参考线L3位于中间参考线L2和短边CD之间。中 间参考线L2与起始参考线L1之间的距离为2.3～2.6m，本实施例中为2.5米，末端参考线L3与中间参考线L2之间的距离为1.8～2.2m，本实施例中为2米。其中，起始参考线L1、中间参考线L2和初始垂直停车位的长边AD、BC的延长线围成的区域标记为D1，末端参考线L3、中间参考线L2和初始垂直停车位的长边AD、BC围成的区域标记为D2。可以理解的，起始参考线L1、中间参考线L2和末端参考线L3可全部或部分为曲线(非直线)。当起始参考线为曲线时，其位于初始垂直停车位的入口短边之外，且起始参考线上的各点到入口短边的最短距离(垂线段的长度)大于车辆的车角超声波传感器探测距离的触发阈值。当末端参考线为非直线时，其与中间参考线L2之间的距离应落在1.8～2.2m内。中间参考线用于触发第二车位重定位状态，此时车辆与车位基本呈平行状态，因此中间参考线实质或大体呈直线。
为方便描述，本实施例中，设定在泊车的过程中最先越过起始参考线L1的车辆的车角为右后角Pd(第一车角)。
请参考图3至图6，一实施例中，本发明的垂直泊车方法包括步骤：
S001，进入泊车状态，且判断为倒车泊车状态，获取初始垂直停车位和初始垂直停车位的外边界的坐标信息；
S002，根据车辆车尾边界的中点M1的坐标信息为(M1_x,M1_y,Theta)判断点M1是否位于起始参考线L1上或已经进入区域D1(进入区域D1也即位于起始参考线L1和中间参考线L2之间)，如果是，则进入第一车位重定位状态，执行步骤S003，否则执行步骤S014。
S003，当第一车角(本实施例中为右后角Pd)的超声波传感器SR_RR和LR_RR的探测障碍物距离Dist小于阈值DistTh_Up，也即Dist落在传感器探测盲区阈值DistTh_Down和探测距离的触发阈值DistTh_Up之间(Dist∈ [DistTh_Down,DistTh_Up])时，判断车辆在初始垂直停车位内有撞到障碍物的可能性，执行步骤S004，否则返回步骤S001。本实施例中，盲区阈值DistTh_Down为0.3m(小于或等于0.3m时是超声波传感器的探测盲区，且保持输出为0.3m不变)；超声波探测距离的触发阈值DistTh_Up为0.4m。
S004，基于超声波传感器SR_RR及LR_RR的公共探测FOV(Field of View)以及超声波的单发多收机制，利用三角定位法对障碍物进行定位，若超声波探测距离Dist∈[DistTh_Down,DistTh_Up]，可定位到障碍物，则执行步骤S005，否则执行步骤S013。
S005，计算在大坐标系x-o-y下障碍物的坐标(obj_x,obj_y)，并按照时间以滑动窗口的形式(缓存一组固定大小的点，比如10个点；然后，每次刷新，就会更新一个点进入缓存，同时将最“旧”的点删除，这样在每个更新周期，都会有不断的新点进入缓存，不断的“旧”点删除，目的是始终保持的固定大小的数组，对历史数据进行处理，以便得到一个准确的结果)，记录数据长度为N的障碍物坐标点集W＝{(obj_x(n),obj_y(n))|N＝1,2,3,...N}，N为自然整数，obj_x(n)为障碍物X轴坐标值，obj_y(n)为障碍物Y轴坐标值，N优选为5。
S006，对点集W进行最小二乘直线拟合，获得直线L：y＝K*x+b。
S007，求解点集W中的各点到直线L距离的均方根Var：
Figure PCTCN2019120799-appb-000002
S008，判断Var大小，若Var大于预设值VarTh(Var>VarTh)，则认为点集W不满足计算要求，则重新执行步骤S001，否则，执行步骤S009。预设值VarTh可选自[0.01,0.03]，优选为0.01。
S009，确定点O的横坐标O_x＝obj_x(N)，纵坐标O_y＝obj_y(N)。
上述步骤S005至S009实质为计算障碍物最有可能与车辆接触的点O的坐 标信息。可以理解的其他实施例中，可以采用其他能确定最有可能与车辆接触的点O的坐标信息。
S010，计算初始垂直停车位的靠近障碍物的一条长边的斜率K 1，本实施例中为长边BC的斜率K 1，再获得穿过点O且斜率为-1/K 1的第一直线Lo：y＝(-1/K 1)*(x-O_x)+O_y，以及穿过右后角Pd(第一车角)且斜率为K 1的第二直线Lp：y＝K 1*(x–Pd_x)+Pd_y，计算第一直线Lo和第二直线Lp的交点Cr的坐标信息(Cr_x,Cr_y)。图5为图4中被虚线圆圈22圈住的部分的放大图，可以看出交点Cr形成的过程。
S011，根据点Cr的坐标信息(Cr_x,Cr_y)以及初始垂直停车位的外边界的坐标信息，计算目标垂直停车位(用A'、B'、C'、D'标记四角点，虚线标记四边)的相对于初始垂直停车位的偏移量。由于X轴大体平行于停车位长边则只需计算Y轴方向偏移量Delt_y＝|B_y–Cr_y|即可。
S012，根据计算得到的偏移量求得目标垂直停车位的四角的坐标信息A'-B'-C'-D'，然后执行步骤S001。
由于在测量初始垂直停车位时，对其入口处的短边所在直线的测量较为准确，且通常情况下初始垂直停车位和目标垂直停车位的长边基本平行于坐标系X轴，因此在计算得到初始垂直停车位的偏移量后，可对初始垂直停车位沿入口短边向远离障碍物的方向平移Y轴方向偏移量即可。
因此，在可确定点Cr坐标的情况下，可直接根据点O、第一车角以及初始垂直停车位的外边界的坐标信息，计算目标垂直停车位的外边界的坐标信息。本实施例中B'的坐标值为(B_x,Cr_y)，又因为Cr_y实质上等于Pd_y，B'的坐标值为(B_x,Pd_y)。也即，步骤S010和S011为在能够获得点O坐标信息下的一种计算车位偏移量的方法，其实质为先获得穿过第一车角(右后角Pd)且 斜率为K1的第二直线Lp，再获得第二直线Lp与初始垂直停车位的入口短边AB的交点B'的坐标，然后将初始垂直停车位沿入口短边向远离障碍物的方向平移，使得目标垂直停车位的靠近第一车角的角点位于交点B'，从而获得目标垂直停车位的外边界的坐标信息A'-B'-C'-D'。
S013，利用右后角Pd(第一车角)的坐标信息(Pd_x,Pd_y)、靠近右后角Pd(第一车角)的初始垂直停车位的入口短边的角点B的坐标信息(B_x,B_y)、以及右后角Pd(第一车角)的超声波传感器SR_RR和LR_RR的探测障碍物距离Dist计算目标垂直停车位相对于初始垂直停车位的Y轴方向的偏移量Delt＝|B_y–Pd_y|-sqrt(Dist 2-|B_x–Pd_x| 2)，然后执行步骤S012。
也即，在无法定位到障碍物的情况下，根据第一车角的超声波传感器探测障碍物距离Dist、第一车角以及靠近第一车角的初始垂直停车位的入口短边的角点的坐标信息，计算目标垂直停车位的外边界的坐标信息。具体利用了Dist，右后角Pd(第一车角)和角点B的坐标值，将Dist作为一直角三角形的斜边的长度，将右后角Pd和角点B的X轴方向的差值作为直角三角形的平行于X轴的直角边的长度，求取平行于Y轴的直角三角形的直角边的长度，并将该长度作为右后角Pd和目标垂直停车位的角点B'在Y轴方向的差值。
S014，根据车辆车尾边界的中点M1的坐标信息为(M1_x,M1_y,Theta)判断点M1是否位于中间参考线L2或末端参考线L3上或已经进入区域D2(也即位于中间参考线L2和末端参考线L3之间)，如果是，从第一车位重定位状态转入第二车位重定位状态，执行步骤S015，否则执行步骤S200。
S015，定义车辆右后角的超声波传感器SR_RR的探测障碍物距离为DistR，左后角的超声波传感器SR_RL的探测障碍物距离为DistL，则判断(DistL+DistR)是否大于2*DistTh_Up，如果是，执行步骤S016，否则执行步骤S021。
S016，判断是否DistL>DistTh_Up且DistR>DistTh_Up，如果是则判断可继续泊车，执行步骤S001，否则执行步骤S017。
S017，判断是否DistL>DistTh_Up且DistR≤DistTh_Up，如果是则判断车辆右侧存在障碍物(请参考图6中扇形的超声波辐射范围)，则执行步骤S018，否则执行步骤S019。
S018，计算车位调整偏移量为Delt＝DistTh_Up–DistR，执行步骤S019。
S019，判断是否DistL≤DistTh_Up且DistR>DistTh_Up，如果是，则判断车辆左侧存在障碍物，则执行步骤S020，否则执行步骤S001。
S020，确定需要调整的偏移量为Delt＝DistTh_Up–DistL，执行步骤S012。
S021，确定不可泊入车位并输出提醒信息。
步骤S015至S021实质为根据DistR、DistL和车辆的车角超声波传感器探测距离的触发阈值DistTh_Up判断车辆至少一侧是否有足够的位置可供人员上下车，并根据判断结果输出不可泊入车位的提醒信息或计算目标垂直停车位的外边界的坐标信息，以使得泊车结束后车辆至少一侧可供人员下车。可以理解的，可以用其他预设值取代DistTh_Up进行判断，例如根据车辆销售区域考虑到不同人种不同体型或不过地区车辆标准车位不同采用不同数值。只要判断的结果导向车辆至少一侧有足够的空间可供人员上下车即可。
S200，判断车辆车尾边界的中点M1的坐标信息是否到达末端参考线L3，如果是，泊车结束，否则执行步骤S001。其他实施例的步骤S200中，还可以通过判定车头是否进入车位入口短边之内来确定是否泊车结束，此时不需要判断点M1与末端参考线L3的位置关系，则末端参考线L3可取消。例如，判断车头点M2或Pa和Pb中所有点均进入车位入口短边之内即确定泊车结束，或车头点Pa和Pb在车位入口短边之内且距离入口短边的距离均大于0.1m则泊车结 束。
通过执行上述垂直泊车方法，可达到如下效果：
(1)在车辆泊入区域D1时，可通过障碍定位的方法预测真实车位位置，并得到真实有效的车位，可提高泊车入位准确性，也可改善用户用体验；
(2)在车辆泊入区域D2时，由于垂直停车位两侧车辆倾斜停放，导致车位尾部相对较窄，可利用超声波探测两侧障碍物距离，得到真实有效地车位，并保证泊车入位后，可正常开门下车；
(3)可减小车辆利用超声波定位所带来的车位检测误差，提高泊车入位成功率。
如此，车辆在自动或半自动泊车的过程中，不断根据周围障碍物状态进行车位重定位，消除初始车位预估误差，避免车辆泊车入位后，由于车辆与左侧或右侧参考车辆距离较小，从而导致乘客无法正常下车的情况，也可避免在车辆泊车入位的最后一步由于车位误差使得车辆行进过程中，因遇到到障碍物而中断刹停，影响用户体验的情况。
在其他实施例中，可以理解的，可将步骤S016和S017对调。
其他实施例中，可以理解的，初始垂直停车位的位置(外边界的坐标信息)可由第三方获得，并由车辆的驾驶***转换为大坐标系x-o-y下的坐标。
上述实施例中，考虑到在测量初始垂直停车位时对其入口处的短边所在位置的测量较为准确，因此在计算目标垂直停车位的最终坐标信息时，是直接对初始垂直停车位沿入口短边向远离障碍物的方向平移来实现。可以理解的，在计算出偏移量的情况下，可采用其他算法获得目标垂直停车位的最终坐标。
其他实施例中，可省略末端参考线L3，也即无需定义末端参考线L3也可实现本发明方案。
其他实施例中，步骤S002中，只要车辆上一个预设的点，只要是位于车辆中轴线(垂直于车辆前轴和后轴的中轴)上，且靠近车头部或车尾部即可，的坐标信息位于起始参考线L1上或已经进入起始参考线L1和中间参考线L2之间的区域即可执行。当该预设的点不是车辆车头边界的中点和车辆车尾边界的中点时，应对照修改起始参考线和中间参考线相对入口短边的位置。
上述实施例中，设定在泊车的过程中最先越过起始参考线L1的车辆的车角为右后角Pd(作为第一车角)。可以理解的，当第一车角为左后角或左前角或右前角时，只要对应调整垂直泊车方法中与真正的“第一车角”对应的车角及靠近该“第一车角”或其对应的初始垂直停车位和目标垂直停车位的角点即可。例如，当第一车角为左后角时，将步骤S003中的“判断右后角Pd(第一车角)的超声波传感器SR_RR和LR_RR的探测障碍物距离”修改为“判断左后角Pc(第一车角)的超声波传感器SR_RL和LR_RL的探测障碍物距离”；将步骤S004中的“超声波传感器SR_RR及LR_RR”修改为“超声波传感器SR_RL及LR_RL”；将步骤S010中计算长边BC的斜率，修改为“计算长边AD的斜率”，将“穿过右后角Pd(第一车角)且斜率为……”修改为“穿过左后角Pc(第一车角)且斜率为……”；将步骤S011中“特别是右后角Pd(第一车角)的坐标信息(Pd_x,Pd_y)”修改为“特别是左后角Pc(第一车角)的坐标信息(Pc_x,Pc_y)”；将步骤S013中的“利用右后角Pd(第一车角)的坐标信息(Pd_x,Pd_y)、靠近右后角Pd(第一车角)的初始垂直停车位的入口短边的角点B的坐标信息(B_x,B_y)、以及右后角Pd(第一车角)的超声波传感器SR_RR和LR_RR的探测障碍物距离Dist计算目标垂直停车位相对于初始垂直停车位的偏移量Delt＝|B_y–Pd_y|-sqrt(Dist 2-|B_x–Pd_x| 2)”修改为“利用左后角Pc(第一车角)的坐标信息(Pc_x,Pc_y)、靠近左后角Pc(第一车角)的初始垂直停车位的入口 短边的角点A的坐标信息(A_x,A_y)、以及左后角Pc(第一车角)的超声波传感器SR_RL和LR_RL的探测障碍物距离Dist计算目标垂直停车位相对于初始垂直停车位的偏移量Delt＝|A_y–Pc_y|-sqrt(Dist 2-|A_x–Pc_x| 2)”即可。但步骤S014至步骤S020中的参数不随第一车角变换为左后角或左前角或右前角而变化。
在本发明的描述中，需要理解的是，术语诸如“上”、“下”、“前”、“后”、“左”、“右”、“竖直”、“水平”、“顶”、“底”“内”、“外”等指示的方位或位置关系为基于附图所示的方位或位置关系，仅是为了便于描述本发明和简化描述，而不是指示或暗示所指的装置或元件必须具有特定的方位、以特定的方位构造和操作，因此不能理解为对本发明的限制。
此外，术语“第一”、“第二”仅用于描述目的，而不能理解为指示或暗示相对重要性或者隐含指明所指示的技术特征的数量。由此，限定有“第一”、“第二”的特征可以明示或者隐含地包括一个或者更多个该特征。在本发明的描述中，“多个”的含义是两个或两个以上，除非另有明确具体的限定。
虽然对本发明的描述是结合以上具体实施例进行的，但是，熟悉本技术领域的人员能够根据上述的内容进行许多替换、修改和变化、是显而易见的。因此，所有这样的替代、改进和变化都包括在附后的权利要求的精神和范围内。

Claims (10)

  1. 一种垂直泊车方法，其特征在于，在利用传感器检测或从第三方获得一初始垂直停车位和所述初始垂直停车位的外边界的坐标信息以进行泊车的过程中，定义一位于所述初始垂直停车位的入口短边之外的起始参考线，所述起始参考线上的各点到所述入口短边的最短距离大于车辆的车角超声波传感器探测距离的触发阈值；并定义在泊车的过程中会最先越过所述起始参考线的车辆的车角为第一车角；
    所述垂直泊车方法包括：
    步骤一，当车辆的靠近所述第一车角的一个预设位置的坐标信息位于所述起始参考线上或位于所述起始参考线的靠近初始垂直停车位一侧，且所述第一车角的超声波传感器检测到障碍物时，定位障碍物，执行步骤二；
    步骤二，计算所述障碍物最有可能与车辆接触的点O的坐标信息，根据点O、第一车角以及初始垂直停车位的外边界的坐标信息，计算一重定位的目标垂直停车位的外边界的坐标信息；以及
    步骤三，根据所述目标垂直停车位的外边界的坐标信息继续进行泊车。
  2. 根据权利要求1所述的垂直泊车方法，其特征在于，步骤二中，所述计算目标垂直停车位的外边界的坐标信息的方法包括：
    计算所述初始垂直停车位的长边的斜率K 1；
    获得穿过所述第一车角且斜率为K 1的第二直线；
    计算第二直线与初始垂直停车位的入口短边的交点的坐标信息；以及
    将所述初始垂直停车位沿入口短边向远离障碍物的方向平移，使得目标垂直停车位的靠近所述第一车角的角点位于所述交点，从而获得目标垂直停车位的外边界的坐标信息。
  3. 根据权利要求2所述的垂直泊车方法，其特征在于，步骤二中，所述计算所 述障碍物最有可能与车辆接触的点O的坐标信息的步骤包括：
    按照时间以滑动窗口的形式记录数据长度为N的障碍物坐标点集W＝{(obj_x(n),obj_y(n))|N＝1,2,3,...N}，N为自然整数，obj_x(n)为障碍物X轴坐标值，obj_y(n)为障碍物Y轴坐标值；
    对所述点集W进行最小二乘直线拟合，获得直线L：y＝K*x+b；
    求解所述点集W中的各点到直线L距离的均方根Var：
    Figure PCTCN2019120799-appb-100001
    判断Var大小，若Var大于预设值，则认为点集W不满足计算要求，则重新执行步骤一，否则，确定点O的横坐标O_x＝obj_x(N)，纵坐标O_y＝obj_y(N)。
  4. 根据权利要求1所述的垂直泊车方法，其特征在于，所述步骤一中，若无法定位到障碍物，则执行步骤四：
    根据第一车角的超声波传感器探测障碍物距离Dist、第一车角以及靠近第一车角的初始垂直停车位的入口短边的角点的坐标信息，计算目标垂直停车位的外边界的坐标信息，然后执行步骤三。
  5. 根据权利要求4所述的垂直泊车方法，其特征在于，所述步骤四中所述计算目标垂直停车位的外边界的坐标信息的方法为：计算所述目标垂直停车位相对于初始垂直停车位的在Y轴方向的偏移量Delt：|B_y–Pd_y|-sqrt(Dist 2-|B_x–Pd_x| 2)，其中Y轴大体平行于所述初始垂直停车位的入口短边，B_x和B_y分别为靠近所述第一车角的初始垂直停车位的角点B的X轴坐标值和Y轴坐标值，Pd_x和Pd_y分别为第一车角的X轴坐标值和Y轴坐标值；以及
    将所述初始垂直停车位沿入口短边向远离障碍物的方向平移所述偏移量Delt，从而获得目标垂直停车位的外边界的坐标信息。
  6. 根据权利要求1所述的垂直泊车方法，其特征在于，还定义一中间参考线，位于所述初始垂直停车位的入口短边和底部短边之间，所述的垂直泊车方法进一步包括：当车辆的所述预设位置的坐标信息位于所述中间参考线与初始垂直停车位的底部短边之间，且车辆的第一车角或与其相对的第二车角的用于探测车辆两侧的障碍物的超声波传感器的探测障碍物距离分别为DistR，DistL，且第一车角为右后角，第二车角为左后角时，则执行步骤五：根据DistR、DistL和车辆的车角超声波传感器探测距离的触发阈值DistTh_Up判断车辆至少一侧是否有足够的位置可供人员上下车，并根据判断结果输出不可泊入车位的提醒信息或计算目标垂直停车位的外边界的坐标信息，执行步骤三，以使得泊车结束后车辆至少一侧可供人员下车。
  7. 根据权利要求6所述的垂直泊车方法，其特征在于，步骤五中，所述判断车辆至少一侧是否有足够的位置可供人员上下车的方法包括：
    (1)当(DistL+DistR)<2*DistTh_Up，则确定不可泊入车位并输出提醒信息；
    (2)如果(DistL+DistR)≥2*DistTh_Up时：
    (a)如果DistL>DistTh_Up且DistR>DistTh_Up，则确定可继续泊车；
    (b)如果DistL>DistTh_Up且DistR≤DistTh_Up，则确定车辆右侧存在障碍物，则车位调整偏移量为Delt＝DistTh_Up–DistR，并计算目标垂直停车位的外边界的坐标信息，执行步骤三；
    (c)如果DistL≤DistTh_Up且DistR>DistTh_Up，则认为左侧碰到障碍物，则确定需要调整的偏移量为Delt＝DistTh_Up–DistL，并计算目标垂直停车位的外边界的坐标信息，执行步骤三。
  8. 根据权利要求1所述的垂直泊车方法，其特征在于，当所述第一车角为车辆 前角时，所述预设位置为所述车辆车头边界的中点；当所述第一车角为车辆后角时，所述预设位置为所述车辆车尾边界的中点。
  9. 根据权利要求8所述的垂直泊车方法，其特征在于，还定义一靠近所述初始垂直停车位的底部短边的末端参考线；当判断所述预设位置到达末端参考线时，泊车结束。
  10. 根据权利要求9所述的垂直泊车方法，其特征在于，当判定车身全部进入车位入口短边之内则泊车结束。
PCT/CN2019/120799 2019-02-19 2019-11-26 垂直泊车方法 WO2020168769A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
EP19915876.7A EP3929046B1 (en) 2019-02-19 2019-11-26 Vertical parking method

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201910121262.6A CN109927715B (zh) 2019-02-19 2019-02-19 垂直泊车方法
CN201910121262.6 2019-02-19

Publications (1)

Publication Number Publication Date
WO2020168769A1 true WO2020168769A1 (zh) 2020-08-27

Family

ID=66985641

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/120799 WO2020168769A1 (zh) 2019-02-19 2019-11-26 垂直泊车方法

Country Status (3)

Country Link
EP (1) EP3929046B1 (zh)
CN (1) CN109927715B (zh)
WO (1) WO2020168769A1 (zh)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113460039A (zh) * 2021-07-22 2021-10-01 上汽通用五菱汽车股份有限公司 泊车控制方法、车辆及存储介质
CN113830076A (zh) * 2021-09-17 2021-12-24 东风柳州汽车有限公司 泊车方法、装置、设备及存储介质
CN115331478A (zh) * 2022-08-11 2022-11-11 富赛汽车电子有限公司 避免车辆在斜坡停车的控制方法、装置、电子设备及介质
CN116653920A (zh) * 2023-04-18 2023-08-29 广州优保爱驾科技有限公司 一种基于4d毫米波雷达的自动泊车方法和***

Families Citing this family (24)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109927715B (zh) * 2019-02-19 2020-08-25 惠州市德赛西威智能交通技术研究院有限公司 垂直泊车方法
CN110271539B (zh) * 2019-07-22 2020-12-04 重庆长安汽车股份有限公司 一种自动垂直泊车控制***
CN110371109B (zh) * 2019-07-23 2020-09-04 国唐汽车有限公司 一种基于多传感器融合的电动公交车自动泊车方法
CN110444044B (zh) * 2019-08-27 2022-07-12 纵目科技（上海）股份有限公司 基于超声波传感器的车辆位姿检测***、终端和存储介质
CN110517507B (zh) * 2019-08-27 2022-10-11 纵目科技（上海）股份有限公司 基于超声波传感器的车辆位姿检测方法、***、终端和存储介质
CN110838244B (zh) * 2019-10-17 2021-02-02 惠州市德赛西威智能交通技术研究院有限公司 一种基于超声波的垂直泊车位的车位航向调整方法
CN110969655B (zh) * 2019-10-24 2023-08-18 百度在线网络技术（北京）有限公司 用于检测车位的方法、装置、设备、存储介质以及车辆
CN110956847B (zh) * 2019-12-20 2021-11-30 奇瑞汽车股份有限公司 车位的识别方法、装置及存储介质
CN111114538B (zh) * 2020-02-18 2021-11-05 威马智慧出行科技（上海）有限公司 自动泊车倒车入库车位判断方法、电子设备及汽车
CN111591295B (zh) * 2020-03-31 2022-06-14 宁波吉利汽车研究开发有限公司 一种目标车位确认方法、装置、电子设备和存储介质
CN111824130A (zh) * 2020-07-09 2020-10-27 广州小鹏车联网科技有限公司 一种泊车控制方法、装置、车辆、存储介质
CN112193241A (zh) * 2020-09-29 2021-01-08 湖北亿咖通科技有限公司 一种自动泊车方法
CN112744213A (zh) * 2021-01-07 2021-05-04 广州小鹏自动驾驶科技有限公司 一种自动泊车的方法和装置
JP7355043B2 (ja) * 2021-01-25 2023-10-03 トヨタ自動車株式会社 自動駐車システム
CN112937554A (zh) * 2021-01-30 2021-06-11 惠州华阳通用电子有限公司 一种泊车方法及***
CN112776797A (zh) * 2021-02-27 2021-05-11 重庆长安汽车股份有限公司 泊出原始车位创建方法、***、车辆和存储介质
CN115083199B (zh) * 2021-03-12 2024-02-27 上海汽车集团股份有限公司 一种车位信息确定方法及其相关设备
CN113077652B (zh) * 2021-03-19 2022-08-09 上海欧菲智能车联科技有限公司 车位自动搜索方法、装置、电子装置及存储介质
CN113511191B (zh) * 2021-05-12 2022-06-07 重庆长安汽车股份有限公司 垂直泊车完成判断***、方法、车辆及存储介质
CN113516870B (zh) * 2021-05-17 2022-11-11 上海欧菲智能车联科技有限公司 待泊车位确定方法、装置、自动泊车设备和存储介质
CN114274952B (zh) * 2021-12-24 2023-09-22 重庆长安汽车股份有限公司 垂直车位自主泊车方法、***、存储介质及电子设备
CN115214627B (zh) * 2022-05-10 2024-01-23 广州汽车集团股份有限公司 泊车提示方法、装置、电子设备及存储介质
CN115056765A (zh) * 2022-06-30 2022-09-16 重庆长安汽车股份有限公司 垂直与斜列车位自主泊车的路径规划方法及装置
CN117565860B (zh) * 2024-01-16 2024-03-19 江苏开沃汽车有限公司 一种远程自动泊车的方法、***、终端设备及存储介质

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH0628598A (ja) * 1992-07-08 1994-02-04 Nissan Motor Co Ltd 駐車アシスト装置
US20160280263A1 (en) * 2015-03-23 2016-09-29 Mitsubishi Electric Corporation Parking assist apparatus
CN106143609A (zh) * 2016-07-15 2016-11-23 江苏大学 一种泊车***及其垂直方向泊车方法
US20180297589A1 (en) * 2015-10-27 2018-10-18 Clarion Co., Ltd. Parking assistance device
CN109131317A (zh) * 2018-07-23 2019-01-04 同济大学 基于多段式规划与机器学习的自动垂直泊车***及方法
CN109927715A (zh) * 2019-02-19 2019-06-25 惠州市德赛西威智能交通技术研究院有限公司 垂直泊车方法

Family Cites Families (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8560169B2 (en) * 2007-12-18 2013-10-15 Honda Motor Co., Ltd. Vehicular parking feasibility determining system, vehicular parking space detection system and vehicular movable range detection system
CN102372000B (zh) * 2010-08-23 2014-02-05 北京经纬恒润科技有限公司 获取泊车位参数的装置和***
CN102431551B (zh) * 2011-11-18 2014-06-11 奇瑞汽车股份有限公司 一种泊车车位探测方法以及泊车车位探测***
CN103241239B (zh) * 2013-04-27 2015-10-14 重庆邮电大学 一种自动泊车***车位识别方法
DE102014200611A1 (de) * 2014-01-15 2015-07-16 Robert Bosch Gmbh Verfahren zum autonomen Parken eines Fahrzeugs, Fahrerassistenzvorrichtung zur Durchführung des Verfahrens, sowie Fahrzeug mit der Fahrerassistenzvorrichtung
CN105197010B (zh) * 2014-06-04 2018-03-27 长春孔辉汽车科技股份有限公司 辅助泊车***以及辅助泊车控制方法
DE102014220298A1 (de) * 2014-10-07 2016-04-07 Bayerische Motoren Werke Aktiengesellschaft Unfallvermeidung mit dem nachfolgenden Verkehr bei einem automatisierten Einparkvorgang in eine Querparklücke
CN107672585B (zh) * 2017-08-29 2019-09-17 广州小鹏汽车科技有限公司 一种自动泊车路径规划方法及***
CN108281041A (zh) * 2018-03-05 2018-07-13 东南大学 一种基于超声波和视觉传感器相融合的泊车车位检测方法

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH0628598A (ja) * 1992-07-08 1994-02-04 Nissan Motor Co Ltd 駐車アシスト装置
US20160280263A1 (en) * 2015-03-23 2016-09-29 Mitsubishi Electric Corporation Parking assist apparatus
US20180297589A1 (en) * 2015-10-27 2018-10-18 Clarion Co., Ltd. Parking assistance device
CN106143609A (zh) * 2016-07-15 2016-11-23 江苏大学 一种泊车***及其垂直方向泊车方法
CN109131317A (zh) * 2018-07-23 2019-01-04 同济大学 基于多段式规划与机器学习的自动垂直泊车***及方法
CN109927715A (zh) * 2019-02-19 2019-06-25 惠州市德赛西威智能交通技术研究院有限公司 垂直泊车方法

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113460039A (zh) * 2021-07-22 2021-10-01 上汽通用五菱汽车股份有限公司 泊车控制方法、车辆及存储介质
CN113460039B (zh) * 2021-07-22 2022-04-12 上汽通用五菱汽车股份有限公司 泊车控制方法、车辆及存储介质
CN113830076A (zh) * 2021-09-17 2021-12-24 东风柳州汽车有限公司 泊车方法、装置、设备及存储介质
CN113830076B (zh) * 2021-09-17 2023-12-19 东风柳州汽车有限公司 泊车方法、装置、设备及存储介质
CN115331478A (zh) * 2022-08-11 2022-11-11 富赛汽车电子有限公司 避免车辆在斜坡停车的控制方法、装置、电子设备及介质
CN116653920A (zh) * 2023-04-18 2023-08-29 广州优保爱驾科技有限公司 一种基于4d毫米波雷达的自动泊车方法和***
CN116653920B (zh) * 2023-04-18 2024-04-16 广州优保爱驾科技有限公司 一种基于4d毫米波雷达的自动泊车方法和***

Also Published As

Publication number Publication date
EP3929046A4 (en) 2022-05-04
EP3929046A1 (en) 2021-12-29
CN109927715A (zh) 2019-06-25
EP3929046B1 (en) 2023-10-04
CN109927715B (zh) 2020-08-25

Similar Documents

Publication Publication Date Title
WO2020168769A1 (zh) 垂直泊车方法
JP6418332B2 (ja) 車両位置推定装置、車両位置推定方法
CN105551302B (zh) 辅助停车过程的方法以及停车辅助装置
JP4461920B2 (ja) 駐車支援装置
US20220169280A1 (en) Method and Device for Multi-Sensor Data Fusion For Automated and Autonomous Vehicles
US20090121899A1 (en) Parking assistance device
KR20200109365A (ko) Slam을 사용한 차량의 위치 보정 기법
US20080238718A1 (en) Method for preventing lane departure for use with vehicle
WO2016117507A1 (ja) 区画線認識装置
CN109581389B (zh) 一种识别泊车车位边界的方法和装置
WO2007132860A1 (ja) 対象物認識装置
WO2006090736A1 (ja) 物体認識装置
JP3473378B2 (ja) 駐車空間検出装置
JP2017159879A (ja) 車両位置認識システム
CN112771591B (zh) 用于评价运输工具的环境中的对象对运输工具的行驶机动动作的影响的方法
KR20140104611A (ko) 차량의 자동 주차 장치 및 이를 이용한 방법
JP2002240661A (ja) 駐車支援装置
CN108149990A (zh) 自动泊车车位探测方法和探测***以及车辆
US12030518B2 (en) Lane changing based only on local information
WO2024001215A1 (zh) 垂直与斜列车位自主泊车的路径规划方法及装置
JP2002098758A (ja) 走行車線に関する自動車の位置を求めるための方法および装置
JP2006504962A (ja) 車両の駐車過程用の幾何学的形状データを検出するための方法
JP4110922B2 (ja) 車両用外界認識装置
CN114132305A (zh) 用于垂直与斜列泊车路径规划方法、***、存储介质及车辆
US20190064321A1 (en) Vehicle exterior environment recognition apparatus

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 19915876

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

ENP Entry into the national phase

Ref document number: 2019915876

Country of ref document: EP

Effective date: 20210920