A multi-mode parking path planning method and system based on scene dynamic identification and a storage medium

Abstract

The application relates to a multi-mode parking path planning method and system based on scene dynamic identification and a storage medium, and the method comprises the following steps: U1: acquiring vehicle positioning information, parking space information and road information; U2: establishing a parking space coordinate system, acquiring path planning information from a vehicle starting position to a vehicle parking target point according to the vehicle positioning information, the parking space information and the road information, wherein the path planning information comprises a first path, a second path and a third path, the first path is that a vehicle starting position point p1 drives forward to a starting parking point p2, and the second path is that the starting parking point p2 reverses to the parking target point p3. The application not only improves the flexibility, fluency and intelligence of automatic parking, but also has strong adaptability to various parking initial poses, and avoids parking failure caused by the fact that the pose of a vehicle parked by a driver does not meet the requirements in a conventional parking method.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of unmanned path planning, in particular to a multi-mode parking path planning method and system based on scene dynamic identification and a storage medium. BACKGROUND

[0002] Due to the increase in the number of vehicles and the improvement in people's consumption level, parking in crowded areas such as urban areas, business districts and tourist attractions is a big problem. Generally, in urban areas, the number of parking spaces is very limited. In addition to the shortage of parking spaces, the technical requirements for drivers are high in the case of dense parking, and there is a strong pain point of "difficulty in parking and difficulty in taking the car". In addition, the safety problems caused by parking are increasing year by year. Automatic parking technology is born to solve the parking problem. After the driver parks the vehicle in front of the parking space, the automatic parking system completes the automatic parking process. The automatic parking system is composed of a parking path planning module and a parking control module. The parking path planning module generates a parking path curve that is continuous in curvature and meets the path constraints from the current position of the vehicle to the parking point of the parking space. The existing research generally limits the parking starting position more rigidly. When the driver does not accurately park the vehicle at the desired parking starting point, it is easy to cause parking planning failure or poor parking effect, which increases the operation difficulty of the driver during parking. In addition, the existing parking path planning action is fixed, either one-segment path parking, or two-segment path parking, or three-segment path parking. The whole parking action is rigid and fixed, and cannot reflect the flexibility and smoothness of automatic parking. SUMMARY

[0003] In view of the above problems, the present application provides a multi-mode parking path planning method and system based on scene dynamic identification and a storage medium, which not only improves the flexibility, smoothness and intelligence of automatic parking, but also has strong adaptability to various parking initial positions, avoiding parking failure caused by the driver's parking position not meeting the requirements in the conventional parking method.

[0004] To achieve the above objects and other related objects, the technical solutions of the present application are as follows:

[0005] A multi-mode parking path planning method based on scene dynamic identification, the method comprising:

[0006] U1: obtaining vehicle positioning information, parking space information and road information;

[0007] U2: establishing a parking space coordinate system, and obtaining path planning information from a vehicle starting position to a vehicle parking target point according to the vehicle positioning information, the parking space information and the road information, the path planning information including a first path, a second path and a third path, the first path being a forward driving from a vehicle starting position point p1 to a starting parking point p2, the second path being a reverse driving from the starting parking point p2 to a parking target point p3, and the third path being a driving from a parking termination point p7 to a parking adjustment target point p6;

[0008] U3: obtaining four parking modes according to the first path, the second path and the third path, the four parking modes including a first parking mode for performing parking path planning according to the second path, a second parking mode for performing parking path planning according to the first path and the second path, a third parking mode for performing parking path planning according to the second path and the third path, and a fourth parking mode for performing parking path planning according to the first path, the second path and the third path;

[0009] U4: setting path constraints based on the four parking modes, and obtaining a best parking path curve according to a cost function:

[0010] W i =k1p i +k2s i +k3y i , wherein k1, k2 and k3 are weight coefficients, p i is a minimum curvature of a path curve, s i is a curve length, and y i is a space size occupied by a parking path, to obtain the best parking path curve and complete parking.

[0011] Further, when the vehicle initial position and the parking target point p3 are connected according to the second path, and the actual parking space length is greater than a minimum parallel parking space length, the first parking mode is selected; when the vehicle initial position and the parking target point cannot be connected according to the second path, and the actual parking space length is greater than the minimum parallel parking space length, the second parking mode is selected.

[0012] Further, when the vehicle initial position and the parking target point p3 are connected according to the second path, and the actual parking space length is less than the minimum parallel parking space length, the third parking mode is selected; when the vehicle initial position and the parking target point cannot be connected according to the second path, and the actual parking space length is less than the minimum parallel parking space length, the fourth parking mode is selected.

[0013] Further, the minimum parallel parking space length is L pmin ,

[0014] Lpmin = x p5 + x p3 + L r + 2L safe , where x p5 is the abscissa of point p5 in the second path, x p3 is the abscissa of the parking target point p3, L r is the vehicle rear overhang, L safe is the safety distance of the vehicle rear edge from the rear line of the parking space.

[0015] Further, the second path comprises a first trajectory curve and a second trajectory curve, the first trajectory curve being y

[0016] y = a0x 5 + a1x 4 + a2x 3 + a3x 2 + a4x + a5, where a0, a1, a2, a3, a4, a5 are constant coefficients, and the first trajectory curve constraint condition is obtained according to the p2 point pose (x p2 , y p2 , θ p2 , Φ p2 ) and the p4 point pose (x p4 , y p4 , θ p4 , Φ p4 ):

[0017] The coordinates of the vehicle at the start point p2 and the end point p4 of the first trajectory curve are (x p2 , y p2 ) and (x p4 , y p4 ) respectively; the heading angles of the vehicle at the start point and the end point of the first trajectory curve are θ p2 and θ p4 respectively;

[0018] The equivalent front wheel steering angles of the vehicle at the start point and the end point of the first trajectory curve are Φ p2 and Φ p4 respectively.

[0019] Further, the second trajectory curve is

[0020] where x p3 is the abscissa of the parking target point p3, y p3 is the ordinate of the parking target point p3, and R min is the minimum turning radius of the vehicle.

[0021] Further, the first path is that the vehicle drives from the starting position point p1 to the starting parking point p2, and the obtained trajectory curve is

[0022] y = b0x 5 +b1x 4 +b2x 3 +b3x 2 +b4x+b5, wherein b0, b1, b2, b3, b4, and b5 are constant coefficients, and the trajectory curve constraint condition is obtained according to the p1 point pose (x p1 , y p2 , θ p1 , Φ p1 ) and the p2 point pose (x p2 , y p2 , θ p2 , Φ p2 ).

[0023] The coordinates of the vehicle at the starting point p1 and the ending point p2 of the first trajectory curve are (x p1 , y p2 ) and (x p2 , y p2 ) respectively; and the heading angles of the vehicle at the starting point and the ending point of the first trajectory curve are θ p1 and θ p2 respectively.

[0024] The equivalent front wheel steering angles of the vehicle at the starting point and the ending point of the first trajectory curve are Φ p1 and Φ p2 respectively.

[0025] Further, the path constraint includes a path curvature constraint, a collision between the left side of the vehicle body and the upper boundary of the road, and a collision between the right side of the vehicle body and the lower boundary of the road.

[0026] To achieve the above object and other related objects, the present application further provides a multi-mode parking path planning system based on scene dynamic identification, comprising a computer device programmed or configured to perform the steps of any one of the multi-mode parking path planning methods based on scene dynamic identification.

[0027] To achieve the above object and other related objects, the present application further provides a computer readable storage medium having stored thereon a computer program programmed or configured to perform any one of the multi-mode parking path planning methods based on scene dynamic identification.

[0028] The present application has the following positive effects:

[0029] 1. The present application selects a suitable parking mode through scene dynamic identification, thereby improving the flexibility, fluency, and intelligence of automatic parking.

[0030] 2.The application has strong adaptability to various parking initial poses, and avoids parking failure caused by the driver parking the vehicle in a pose that does not meet the requirements in the conventional parking method.

[0031] 3.The application reduces the length of the parallel parking space required for parking planning to a certain extent.

[0032] 4.The application samples the rectangular area above the parking space at the starting point of the second segment of the parking path, generates a series of parking planning paths, can adapt to the changing road environment of the parking lot, selects the optimal path, and makes the parking path planning have good robustness and not fail to generate a parking path due to changes in the road environment of the parking lot.

[0033] 5.The application uses a cost evaluation function, which can consider curve smoothness, curve length, parking space size, and other aspects to select the optimal parking path. BRIEF DESCRIPTION OF DRAWINGS

[0034] Figure 1 is a schematic diagram of the method of the application;

[0035] Figure 2 is a schematic diagram of the first mode parking trajectory of the application;

[0036] Figure 3 is a schematic diagram of the third mode parking trajectory of the application;

[0037] Figure 4 is a schematic diagram of the fourth mode parking trajectory of the application;

[0038] Figure 5 is a schematic diagram of the parking trajectory of the application. DETAILED DESCRIPTION

[0039] In order to more clearly illustrate the technical solutions in the embodiments of the application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are only some embodiments of the application, and for those skilled in the art, other drawings can be obtained without creative labor under the premise of the drawings.

[0040] Embodiment 1: as shown in Figure 1 or Figure 5 A multi-mode parking path planning method based on scene dynamic identification, the method comprises:

[0041] U1: obtaining vehicle positioning information, parking space information and road information;

[0042] U2: establishing a parking space coordinate system, and obtaining path planning information from a vehicle starting position to a vehicle parking target point according to the vehicle positioning information, the parking space information and the road information, the path planning information including a first path, a second path and a third path, the first path being a forward driving from a vehicle starting position point p1 to a starting parking point p2, the second path being a reverse driving from the starting parking point p2 to a parking target point p3, and the third path being a driving from a parking termination point p7 to a parking adjustment target point p6;

[0043] U3: obtaining four parking modes according to the first path, the second path and the third path, the four parking modes including a first parking mode being a parking path planning according to the second path, a second parking mode being a parking path planning according to the first path and the second path, a third parking mode being a parking path planning according to the second path and the third path, and a fourth parking mode being a parking path planning according to the first path, the second path and the third path;

[0044] U4: setting a path constraint based on the four parking modes, and obtaining a best parking path curve according to a cost function:

[0045] W i =k1p i +k2s i +k3y i , wherein k1, k2 and k3 are weight coefficients, p i is a minimum curvature of a path curve, s i is a curve length, and y i is a space size occupied by a parking path, to obtain the best parking path curve and complete parking.

[0046] In the embodiment, when the vehicle initial position and the parking target point p3 are connected according to the second path, and the actual parking space length is greater than a minimum parallel parking space length, the first parking mode is selected; and when the vehicle initial position and the parking target point cannot be connected according to the second path, and the actual parking space length is greater than the minimum parallel parking space length, the second parking mode is selected.

[0047] In the embodiment, when the vehicle initial position and the parking target point p3 are connected according to the second path, and the actual parking space length is less than the minimum parallel parking space length, the third parking mode is selected; and when the vehicle initial position and the parking target point cannot be connected according to the second path, and the actual parking space length is less than the minimum parallel parking space length, the fourth parking mode is selected.

[0048] In the embodiment, the minimum parallel parking space length is L pmin ,

[0049] Lpmin = x p5 + x p3 + L r + 2L safe , where x p5 is the abscissa of point p5 in the second path, x p3 is the abscissa of the parking target point p3, L r is the vehicle rear overhang, L safe is the safety distance of the vehicle rear edge from the rear line of the parking space.

[0050] In the embodiment, the second path comprises a first trajectory curve and a second trajectory curve, the first trajectory curve is y,

[0051] y = a0x 5 + a1x 4 + a2x 3 + a3x 2 + a4x + a5, where a0, a1, a2, a3, a4, a5 are constant coefficients, and the first trajectory curve constraint condition is obtained according to the p2 point pose (x p2 , y p2 , θ p2 , Φ p2 ) and the p4 point pose (x p4 , y p4 , θ p4 , Φ p4 ):

[0052] The coordinates of the vehicle at the start point p2 and the end point p4 of the first trajectory curve are (x p2 , y p2 ) and (x p4 , y p4 ) respectively; the heading angles of the vehicle at the start point and the end point of the first trajectory curve are θ p2 and θ p4 respectively.

[0053] The equivalent front wheel steering angles of the vehicle at the start point and the end point of the first trajectory curve are Φ p2 and Φ p4 respectively.

[0054] In the embodiment, the second trajectory curve is

[0055] where x p3 is the abscissa of the parking target point p3, y p3 is the ordinate of the parking target point p3, and R min is the minimum turning radius of the vehicle.

[0056] In the embodiment, the first path is that the vehicle drives from the starting position point p1 to the starting parking point p2, and the trajectory curve obtained is

[0057] y = b0x 5 +b1x 4 +b2x 3 +b3x 2 +b4x+b5, wherein b0, b1, b2, b3, b4, and b5 are constant coefficients, and the trajectory curve constraint condition is obtained according to the point pose (x p1 , y p2 , θ p1 , Φ p1 ) of p1 and the point pose (x p2 , y p2 , θ p2 , Φ p2 ) of p2.

[0058] The coordinates of the vehicle at the starting point p1 and the ending point p2 of the first trajectory curve are (x p1 , y p2 ) and (x p2 , y p2 ) respectively; the heading angles of the vehicle at the starting point and the ending point of the first trajectory curve are θ p1 and θ p2 respectively.

[0059] The equivalent front wheel steering angles of the vehicle at the starting point and the ending point of the first trajectory curve are Φ p1 and Φ p2 respectively.

[0060] In the embodiment, the path constraint includes a path curvature constraint, no collision between the left side of the vehicle body and the upper boundary of the road, and no collision between the right side of the vehicle body and the lower boundary of the road.

[0061] To achieve the above object and other related objects, the present application further provides a multi-mode parking path planning system based on scene dynamic identification, comprising a computer device programmed or configured to perform the steps of any one of the multi-mode parking path planning methods based on scene dynamic identification.

[0062] To achieve the above object and other related objects, the present application further provides a computer readable storage medium having stored thereon a computer program programmed or configured to perform any one of the multi-mode parking path planning methods based on scene dynamic identification.

[0063] Embodiment 2: Based on the multi-mode parking path planning method, system, and storage medium based on scene dynamic identification of embodiment 1, the present application is further described as follows.

[0064] As Figure 2 shown in Fig. 2, to reduce the requirement of parking space size for the vehicle when parking, the second path is designed as two curves, curve 2 is a circular arc with a radius equal to the minimum turning radius R min of the vehicle, p3(x p3 , y p3 , θ p3 , Φ p3 ) is the target point of parking, p4(x p4 , y p4 , θ p4 , Φ p4 ) is the right front corner of the vehicle, p5(x p5 , y p5 , θ p5 , Φ p5 ) is the coordinate of the center of the rear axle of the vehicle when the longitudinal coordinate of the vehicle is equal to that of the boundary line 1, curve 1 is a quintic polynomial curve connecting the pose (x p2 , y p2 , θ p2 , Φ p2 ) of p2 and the pose (x p4 , y p4 , θ p4 , Φ p4 ) of p4, when the vehicle travels along curve 1, the coordinates (x, y) of the center of the rear axle, the heading angle θ and the equivalent steering angle Φ of the front wheel of the vehicle all change continuously. The second path curve is calculated by using the back-stepping method, p3 is the end point of parking, its pose p3(x p3 , y p3 , θ p3 , Φ p3 ) in the local coordinate system is:

[0065]

[0066] where L safe is the safety distance of the rear edge of the vehicle from the rear boundary line of the parking space, L r is the rear overhang of the vehicle, and L p is the length of the parking space.

[0067] The equation of curve 2 is:

[0068]

[0069] When the longitudinal coordinate of the right front corner of the vehicle is equal to that of the boundary line 1 and the right front point of the vehicle is on the boundary line of the parking space, the pose (x p4 , y p4 , θ p4 , Φ p4 ) of p4 and the pose (x p5 , y p5 , θ p5 , Φp5 );

[0070] Let the expression of curve 1 be y = a0x 5 +a1x 4 +a2x 3 +a3x 2 +a4x+a5,

[0071] wherein a0, a1, a2, a3, a4, a5 are the coefficients of the first segment path expression, and the constraint conditions are:

[0072] The coordinates of the vehicle at the start point p2 and the end point p4 of curve 1 are (x p2 , y p2 ), (x p4 , y p4 ) respectively;

[0073] The heading angles of the vehicle at the start point and the end point of curve 1 are θ p2 , θ p4 respectively;

[0074] The equivalent front wheel steering angles of the vehicle at the start point and the end point of curve 1 are Φ p2 , Φ p4 respectively;

[0075] The above six constraints are solved simultaneously to obtain the quintic polynomial coefficients a0-a5.

[0076] Embodiment 3: As shown in Figure 3 , the rear line of the garage is moved back by L Δ , and the vehicle backs up along the second segment path to the parking end point The quintic polynomial curve is used to connect and the parking target point p6 (x p6 , y p6 , θ p6 , Φ p6 )

[0077] wherein: L Δ =L pmin -L p

[0078] is the pose of the vehicle when it is away from the rear line L safe of the parking space

[0079]

[0080] The quintic polynomial equation can be obtained by the same method as that for calculating the second segment path curve 1, which is:

[0081] y = b0x 5 +b1x 4 +b2x3 +b3x 2 +b4x+b5.

[0082] like Figure 4 As shown, the pose (x) of point p2, which is the end point of the first parking path and the starting point of the second parking path, is... p2 y p2 θ p2 Φ p2 By sampling within a certain rectangular area above the parking space, a series of parking paths can be obtained.

[0083] In summary, this invention not only improves the flexibility, smoothness, and intelligence of automatic parking, but also has a strong adaptability to various initial parking positions, avoiding parking failures caused by the driver's incorrect parking position in conventional parking methods.

[0084] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A multi-mode parking path planning method based on scene dynamic recognition, characterized in that, The method includes: U1. Obtain vehicle location information, parking space information, and road information; U2. Establish a parking space coordinate system. Based on the vehicle positioning information, the parking space information, and the road information, obtain path planning information from the vehicle's starting position to the vehicle's parking target point. The path planning information includes a first path, a second path, and a third path. The first path is for the vehicle to move forward from the starting position point p1 to the starting parking point p2. The second path is for the vehicle to reverse from the starting parking point p2 to the parking target point p3. The third path is for the vehicle to move from the parking termination point p7 to the parking adjustment target point p6. U3. Based on the first path, the second path, and the third path, four parking modes are obtained, including the first parking mode which is parking path planning based on the second path, the second parking mode which is parking path planning based on the first path and the second path, the third parking mode which is parking path planning based on the second path and the third path, and the fourth parking mode which is parking path planning based on the first path, the second path, and the third path. U4. Based on the four parking modes, set path constraints and apply the cost function: W i =k1ρ i +k2s i +k3y i Where k1, k2, and k3 are the weighting coefficients, and ρ i For the minimum curvature of the path curve, s i Let y be the curve length. i The optimal parking path curve is obtained based on the space occupied by the parking path, and parking is completed. When the vehicle's initial position is connected to the parking target point p3 according to the second path segment, and the actual parking space length is greater than the minimum parallel parking space length, the first parking mode is selected; when the vehicle's initial position cannot be connected to the parking target point according to the second path segment, and the actual parking space length is greater than the minimum parallel parking space length, the second parking mode is selected; when the vehicle's initial position is connected to the parking target point p3 according to the second path segment, and the actual parking space length is less than the minimum parallel parking space length, the third parking mode is selected; when the vehicle's initial position cannot be connected to the parking target point according to the second path segment, and the actual parking space length is less than the minimum parallel parking space length, the fourth parking mode is selected.

2. The multi-mode parking path planning method based on scene dynamic recognition according to claim 1, characterized in that: The minimum parallel parking space length is L pmin , ， Where x p5 Let x be the x-coordinate of point p5 in the second path. p3 Let L be the x-coordinate of the parking target point p3. r For the rear suspension of the vehicle, L safe This refers to the safe distance between the rear edge of the vehicle and the rear line of the parking space.

3. The multi-mode parking path planning method based on scene dynamic recognition according to claim 1, characterized in that: The second path includes a first trajectory curve and a second trajectory curve, wherein the first trajectory curve is y. y = a0x 5 +a1x 4 +a2x 3 +a3x 2 +a4x+a5, where a0, a1, a2, a3, a4, and a5 are constant coefficients, based on the pose of point p2 (x p2 y p2 θ p2 Φ p2 ) and p4 point pose (x p4 y p4 θ p4 Φ p4 The constraint conditions for the first trajectory curve are as follows: the coordinates of the vehicle at the starting point p2 and the ending point p4 of the first trajectory curve are (x, y, p2, p4) respectively. p2 y p2 ), (x p4 y p4 The heading angles of the vehicle at the start and end points of the first trajectory curve are θ and θ', respectively. p2 θ p4 The equivalent front wheel steering angles of the vehicle at the start and end points of the first trajectory curve are Φ and Φ, respectively. p2 Φ p4 .

4. The multi-mode parking path planning method based on scene dynamic recognition according to claim 3, characterized in that: The second trajectory curve is, ， Where x p3 Let y be the x-coordinate of the parking target point p3. p3 Let R be the ordinate of the parking target point p3. min This is the vehicle's minimum turning radius.

5. The multi-mode parking path planning method based on scene dynamic recognition according to claim 1, characterized in that: The first path involves the vehicle moving forward from its starting position p1 to its starting parking position p2, resulting in a trajectory curve y = b0x. 5 +b1x 4 +b2x 3 +b3x 2 +b4x+b5, where b0, b1, b2, b3, b4, and b5 are constant coefficients, based on the pose of point p1 (x p1 y p2 θ p1 Φ p1 p2 pose (x) p2 y p2 θ p2 Φ p2 The trajectory curve constraints are obtained as follows: the coordinates of the vehicle at the starting point p1 and ending point p2 of the first trajectory curve are (x, y, p1, p2, p2, p1 ... p1 y p2 ), (x p2 y p2 The heading angles of the vehicle at the start and end points of the first trajectory curve are θ and θ', respectively. p1 θ p2 The equivalent front wheel steering angles of the vehicle at the start and end points of the first trajectory curve are Φ and Φ, respectively. p1 Φ p2 .

6. The multi-mode parking path planning method based on scene dynamic recognition according to claim 1, characterized in that: The path constraints include path curvature constraints, ensuring that the left side of the vehicle does not collide with the upper boundary of the road, and ensuring that the right side of the vehicle does not collide with the lower boundary of the road.

7. A multi-mode parking path planning system based on scene dynamic recognition, comprising computer equipment, characterized in that, The computer device is programmed or configured to perform the steps of the multi-mode parking path planning method based on scene dynamic recognition as described in any one of claims 1 to 6.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that is programmed or configured to perform the multi-mode parking path planning method based on scene dynamic recognition as described in any one of claims 1 to 6.

Images