Search rule base design method for automatic driving track generation
By designing a variable neighborhood search rule library in an autonomous driving system, combining search methods such as rectangles, trapezoids, rings and sectors, and according to the environmental scene conversion search rules, the trajectory generation method in the existing technology cannot fully reflect differential neighborhood information. The problem of this achieves the effect of reducing accident risk and improving algorithm efficiency.
Patent Information
- Application Number
- CN202411020407.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-29
- Publication Date
- 2024-11-22
AI Technical Summary
In the prior art, the trajectory generation method cannot fully reflect the information of the differential neighborhood when searching for safe neighborhoods, resulting in large amounts of computational volume of search algorithms and easy to fall into local optimal solutions, increasing the risk of accidents.
A search rule library for autonomous driving trajectory is designed. By collecting environmental information around the vehicle, identifying roads, obstacles and traffic signs, and conducting variable neighborhood searches based on driving status and environmental scenes, using rectangles, trapezoids, and rings. Different safe neighborhood search methods such as sector shapes, generate trajectory candidate sets and evaluate them according to the search rules for environmental scene conversion.
It realizes the most suitable safe neighborhood search method in different environmental scenarios, reduces the incidence of smart car accidents, quickly converges to the global optimal solution, and improves the simplicity and reliability of the algorithm.
Smart Images

Figure CN119005464A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of autonomous driving, and in particular relates to a search rule base design method for autonomous driving trajectory generation. Background Art
[0002] At present, large cargo vehicles such as large trucks and port operation box trucks have begun to carry out demonstration applications of autonomous driving cargo loading in specific closed areas, such as ports and airports. This has promoted the automation upgrade of traditional container terminals, among which the application demonstration of unmanned electric container trucks has become an important content. However, in the trajectory planning of smart container trucks, the currently widely used single safety neighborhood search methods such as rectangles and sectors have shortcomings. This is because the length, wheelbase, and width of large trucks are significantly larger than those of conventional vehicles, resulting in a large space for turning and changing directions. Traditional single-type neighborhood or limited field simple combination search methods cannot fully reflect the information of differential neighborhoods, resulting in large search algorithm calculations and easy to fall into local optimal solutions. In order to overcome these problems, it is particularly important to introduce more complex and targeted search rules to combine different safety neighborhoods, thereby establishing a variable neighborhood search method. Summary of the invention
[0003] The purpose of the present invention is to provide a search rule base design method for autonomous driving trajectory generation, which is used to solve the technical problem that the trajectory generation method in the prior art cannot fully reflect the information of the differential neighborhood when performing a safe neighborhood search, resulting in a large amount of search algorithm calculation and easy to fall into a local optimal solution, thereby increasing the risk of accidents.
[0004] The method for designing a search rule base for autonomous driving trajectory generation comprises the following steps.
[0005] S1. Collect environmental information around the vehicle and identify roads, obstacles and traffic signs.
[0006] S2. Determine the scene where the vehicle is located based on the recognition results of roads, obstacles and traffic signs.
[0007] S3. Perform variable neighborhood search according to neighborhood search rules based on driving status and environmental scenarios.
[0008] A rule base for converting the neighborhood search method is pre-established in the controller of the vehicle's automatic driving. The rule base stores driving rules and conversion rules. The driving rules contain vehicle dynamics parameter constraints. The conversion rules are associated with the vehicle's driving state and environmental scenes. The shapes of the safety neighborhood are divided into at least two types.
[0009] S4. Generate a trajectory candidate set based on the variable neighborhood search results.
[0010] S5. Evaluate the candidate trajectory set and select the trajectory that meets the requirements.
[0011] Preferably, a rule base for converting the neighborhood search method is pre-established in the controller of the vehicle's automatic driving, wherein the rule base stores driving rules and conversion rules, wherein the driving rules include vehicle dynamics parameter constraints, and the conversion rules are associated with the vehicle's driving state and environmental scenarios; before establishing the conversion rules, a reliable environmental scenario is first established to simulate the working state of the work vehicle, thereby determining a neighborhood search method suitable for different environmental scenarios.
[0012] Preferably, in step S2, the environmental scene includes.
[0013] Simple straight road scenario: The vehicle travels in a straight line, the road environment is simple, there are no intersections and few obstacles.
[0014] Intersection scenario: There is an intersection and you need to pay attention to traffic dynamics in multiple directions.
[0015] Fast lane scenario: In the fast lane, vehicles are traveling at a faster speed and you only need to pay attention to maintaining a safe distance from the vehicle in front.
[0016] Parking area scenario: In parking areas, vehicles need to frequently change lanes, turn, and avoid other vehicles.
[0017] Temporary traffic sign scenario: There are temporary traffic signs on the road, and you need to pay close attention to the layout and changes in the construction area, as well as temporary traffic control measures.
[0018] Scenarios in crowded areas: In crowded areas, vehicles need to pay special attention to the movements of pedestrians and non-motor vehicles.
[0019] Preferably, in step S3, the conversion rules of the vehicle in this normal driving state include: using a rectangular neighborhood search method in simple straight road scenarios and fast lane scenarios, and using a larger trapezoidal neighborhood search method in intersection scenarios, complex road scenarios with traffic signals, parking area scenarios, temporary traffic sign scenarios and densely populated areas with vehicles and people.
[0020] Preferably, in step S3, the conversion rule of the vehicle in the reversing state includes: combining the annular neighborhood and the trapezoidal neighborhood as a variable neighborhood search method in the reversing state in the parking area scene.
[0021] Preferably, in step S3, the conversion rule of the vehicle in a normal driving state includes: when the steering angle of the vehicle is less than a certain threshold when changing direction or turning, a variable neighborhood search is performed using a fan-shaped neighborhood search method.
[0022] Preferably, in step S4, after determining the safe neighborhood search method for different areas in the vehicle's surrounding environment, a vehicle body coordinate system is established; the width of the rectangular search neighborhood is determined according to the width of the work vehicle, and the width of the smaller base of the trapezoidal neighborhood is kept the same as the width of the rectangular search neighborhood; a mathematical model is established to determine the neighborhood range that the vehicle's driving trajectory can pass through in the coordinate system; the length of the feasible domain search is determined according to the determined neighborhood width and the geometric characteristics of the rectangular neighborhood; the initial point and end point of the trajectory in the corresponding safe neighborhood are determined by the area divided by the variable neighborhood search result, and a trajectory candidate set is generated through the currently adapted trajectory planning algorithm based on the determined initial point positions and end point positions.
[0023] Preferably, in step S4, the specific algorithm for generating the trajectory candidate set by the trapezoidal neighborhood search method is as follows.
[0024] The mathematical expression of the trapezoidal neighborhood is O(x 0 ,y 0 ) represents the initial position of the smart car, and the side of the neighborhood l 1 , l 2 The length of the trapezoidal neighborhood is r, and the length of the top side is l 3 , the height is s, the length of the base is d, then the expression of the trapezoidal safe neighborhood is as follows:
[0025]
[0026] Then, the initial and end positions of the trajectory in each safe neighborhood are determined, and the trajectory is planned using a quintic polynomial, and the following quintic polynomial curve formula is defined:
[0027]
[0028] Where x and y are the position coordinates of the vehicle in the Cartesian coordinate system: 0 ,a 1 ,…,a s and b 0 ,b 1 ,…,b s is the unknown coefficient, and the vehicle obstacle avoidance start time is t 0 , the vehicle position, velocity, and acceleration at the beginning are x 0 ,y 0 ; The obstacle avoidance end time is t 1 , the vehicle parameter at the end time is x 1 ,y 1 ; Therefore, the boundary condition of the quintic polynomial path can be obtained:
[0029]
[0030] Combining the above equations, we can solve the unknown coefficients of the quintic polynomial and obtain the trajectory posterior set.
[0031] The present invention has the following advantages:
[0032] 1. Determine the types of environmental scenarios that trapezoidal neighborhood search and rectangular neighborhood search are most suitable for. According to different environmental scenarios, the corresponding neighborhood search is transformed to achieve the coordination of rectangular safety neighborhood and trapezoidal safety neighborhood according to environmental scenarios. The rectangular safety neighborhood and the trapezoidal safety neighborhood complement each other. This scheme uses the trapezoidal information and topological structure of their differential neighborhoods for search. This method can help the algorithm jump out of the local optimum and quickly converge to the global optimal solution. It can effectively reduce the accident rate of smart cars. In addition to the trapezoidal field search method, circular and fan-shaped safety neighborhoods can also be used according to the driving status and scene environment, and the conversion rules based on the environmental scenario can increase the simplicity and reliability of the neighborhood search. BRIEF DESCRIPTION OF THE DRAWINGS
[0033] Figure 1 This is a basic flow chart of a search rule base design method for autonomous driving trajectory generation according to the present invention.
[0034] Figure 2 It is a mathematical simulation diagram of the surrounding environment when the vehicle is operating in an embodiment of the present invention.
[0035] Figure 3 It is a mathematical simulation diagram of the trapezoidal neighborhood search in an embodiment of the present invention.
[0036] Figure 4 It is a coordinate diagram of the rectangular safety neighborhood in the present invention.
[0037] Figure 5 It is a coordinate diagram of the trapezoidal safety neighborhood in the present invention.
[0038] Figure 6 This is a coordinate diagram of the trapezoidal safety neighborhood based on the mathematical description of the vehicle in the present invention.
[0039] Figure 7 Schematic diagram of generating trajectory candidate sets for the present invention. DETAILED DESCRIPTION
[0040] The specific implementation modes of the present invention will be further explained in detail below by describing the embodiments with reference to the accompanying drawings, so as to help those skilled in the art to have a more complete, accurate and in-depth understanding of the inventive concept and technical solution of the present invention.
[0041] Embodiment 1.
[0042] like Figure 1-Figure 7 As shown, the present invention provides a search rule base design method for autonomous driving trajectory generation, comprising the following steps.
[0043] S1. Collect environmental information around the vehicle and identify roads, obstacles and traffic signs.
[0044] The environmental information around the vehicle is collected by using image acquisition devices such as cameras, and then the corresponding recognition software is used to recognize the information such as roads, obstacles and traffic signs in the image. The recognition software uses existing recognition software with high real-time performance.
[0045] S2. Determine the scene where the vehicle is located based on the recognition results of roads, obstacles and traffic signs.
[0046] The environmental scene is judged based on the recognition results of the previous step. The classification of environmental scenes is pre-divided based on the experience of professionals in this field and verified and corrected by simulating the operation scenes of the corresponding vehicles. Environmental scenes are divided into simple straight road scenes, intersection scenes, fast lane scenes, traffic signal complex road scenes, parking area scenes, temporary traffic sign scenes and crowded areas with people and vehicles.
[0047] S3. Perform variable neighborhood search according to neighborhood search rules based on driving status and environmental scenarios.
[0048] A rule base for converting neighborhood search methods is pre-established in a controller for automatic vehicle driving. The rule base stores driving rules and conversion rules. The driving rules contain vehicle dynamics parameter constraints, and the conversion rules are associated with the vehicle's driving state and environmental scenarios.
[0049] In this solution, the neighborhood search method is divided into at least two types according to the shape of the safe neighborhood. For example, in the normal driving state, the neighborhood search method is divided into a rectangular neighborhood search method and a trapezoidal neighborhood search method. The normal driving state includes the vehicle going straight, turning, and changing direction. Before establishing the conversion rules, first establish a reliable environmental scenario to simulate the working state of the work vehicle, so as to determine the neighborhood search method suitable for different environmental scenarios. The conversion rules of the vehicle in this driving state are based on the relationship between the environmental scenario and the search method, as shown in Table 1.
[0050] Table 1: Relationship between environmental scenarios and search methods
[0051]
[0052] As can be seen from Table 1, in the conversion rules under normal driving conditions, a rectangular neighborhood search method is used in simple straight road scenes and fast lane scenes, while in intersection scenes, complex road scenes with traffic signals, parking area scenes, temporary traffic sign scenes, and crowded areas with people and vehicles, considering that the vehicle length, wheelbase, and width are large, and the scene information is relatively complex, in order to improve the safety of vehicle driving, a larger trapezoidal neighborhood search method is used.
[0053] The relationship between the specific meaning of the above environmental scenarios and the search methods is as follows:
[0054] Simple straight road scenario: In this case, the vehicle mainly travels in a straight line, the road environment is relatively simple, there are no intersections and few obstacles. In this case, rectangular neighborhood search is more suitable due to its simple shape and computational efficiency.
[0055] Intersection scenario: This scenario has an intersection, which is an area with complex traffic conditions. Vehicles need to pay attention to traffic dynamics in multiple directions at the same time. Trapezoidal neighborhood search is more suitable for application at intersections because it can adapt to search needs in different directions.
[0056] Fast lane scenario: In the fast lane, the vehicle is traveling at a high speed and only needs to maintain a safe distance from the vehicle in front. In this condition, the rectangular neighborhood search can efficiently scan the road ahead to ensure safe driving of the vehicle.
[0057] Complex road scene with traffic signals: This type of road usually contains multiple traffic lights, pedestrians, non-motor vehicles and other complex factors. Trapezoidal neighborhood search can more comprehensively cover traffic conditions in multiple directions and provide vehicles with sufficient decision-making information.
[0058] Parking area scenario: In parking areas, vehicles need to frequently change lanes, turn, and avoid other vehicles. Trapezoidal neighborhood search can better adapt to such complex and changing conditions due to its flexibility.
[0059] Temporary traffic sign scenario: In special situations such as construction sections and fixed obstacles on the road, there are usually temporary traffic signs on the road. In this case, the vehicle needs to pay close attention to the layout and changes of the construction area, as well as possible temporary traffic control measures. The use of trapezoidal neighborhood search can quickly identify the construction area and make corresponding driving decisions.
[0060] Scenarios in crowded areas: In crowded areas, vehicles need to pay special attention to the dynamics of pedestrians and non-motor vehicles. Trapezoidal neighborhood search can better cover traffic conditions in multiple directions, helping vehicles to identify and avoid pedestrians and non-motor vehicles in a timely manner.
[0061] The specific meaning of the above environmental scenes is not only used for conversion rules, but also applies to the setting of the standard for judging the environmental scenes in step S2. For example, judging whether it is in a temporary traffic sign scene by the presence or absence of temporary traffic signs; judging whether it is in a densely populated area scene by whether the number of vehicles and pedestrians on the road reaches a threshold.
[0062] S4. Generate a trajectory candidate set based on the variable neighborhood search results.
[0063] After determining the safe neighborhood search method for different areas in the vehicle's surrounding environment, establish a vehicle coordinate system, with the front axle of the vehicle as the x-axis and the vehicle's forward direction as the y-axis.
[0064] The width of the rectangular search neighborhood is determined according to the width of the work vehicle, and the width of the smaller base of the trapezoidal neighborhood is kept the same as the width of the rectangular search neighborhood. The width of the longer base of the trapezoidal neighborhood is determined according to the width of the road, and the length of the rectangular neighborhood is kept consistent with that of the trapezoidal neighborhood.
[0065] A mathematical model is established to determine the neighborhood range that the vehicle's driving trajectory can pass through in the coordinate system. For example, in the attribute model, a circle represents the work vehicle, a black rectangle represents the target position of the work vehicle, a black triangle represents a mobile obstacle in the work vehicle's driving environment, and a white rectangle represents a fixed obstacle in the work vehicle's driving environment.
[0066] The length of the feasible domain search is determined based on the determined neighborhood width and the geometric characteristics of the rectangular neighborhood. Taking the port operation box truck as an example, since the box body of the port operation box truck is relatively large, the capture of differential information by the rectangular traditional safety neighborhood when the box truck turns is insufficient compared to the trapezoidal new safety neighborhood. Therefore, this scheme proposes the necessity of adding the trapezoidal new safety neighborhood in this case, forming a neighborhood search scheme that cooperates with the trapezoid and the rectangle.
[0067] It should be noted that for port operation container trucks, the vehicle body types are divided into single-box (20GP) and double-box (40GP). The two types of port operation container trucks with different body types not only choose different paths, but also affect the length of the feasible neighborhood.
[0068] The initial point and the end point of the trajectory in the corresponding safe neighborhood are determined by the area divided by the variable neighborhood search results, and the trajectory candidate set is generated by the currently adapted trajectory planning algorithm based on the determined positions of each initial point and end point.
[0069] Taking the trapezoidal safe neighborhood as an example, the specific algorithm for generating trajectory candidate sets in this step is as follows: The mathematical expression of the trapezoidal neighborhood is expressed in O(x 0 ,y 0 ) represents the initial position of the smart car, and the side of the neighborhood l 1 , l 2The length of the trapezoidal neighborhood is r, and the length of the top side is l 3 , the height is s, the length of the base is d, then the expression of the trapezoidal safe neighborhood is as follows:
[0070]
[0071] Next, determine the initial position in each neighborhood. For example, for a single-box truck and a double-box truck in a port, the center position of the rear of the box is used as the initial position of the next safe neighborhood for a single-box truck, and the center position of the rear of the second box is used for a double-box truck. In this way, the initial and final positions of the trajectory in each safe neighborhood are determined.
[0072] After that, the candidate trajectory set is generated according to the commonly used trajectory planning algorithm. Taking the quintic polynomial trajectory planning as an example, the following quintic polynomial curve formula is defined:
[0073]
[0074] Where x and y are the position coordinates of the vehicle in the Cartesian coordinate system: 0 ,a 1 ,…,a s and b 0 ,b 1 ,…,b s is the unknown coefficient, and the vehicle obstacle avoidance start time is t 0 , the vehicle position, velocity, and acceleration at the beginning are x 0 ,y 0 ; The obstacle avoidance end time is t 1 , the vehicle parameter at the end time is x 1 ,y 1 ; Therefore, the boundary condition of the quintic polynomial path can be obtained:
[0075]
[0076] Combining the above equations, we can solve the unknown coefficients of the quintic polynomial and obtain the trajectory posterior set.
[0077] S5. Evaluate the candidate trajectory set and select the trajectory that meets the requirements.
[0078] After the candidate set of trajectories is determined, the vehicle evaluates each of them based on safety or mission requirements, including real-time evaluation based on current specific driving parameters, thereby determining the trajectory that best meets the requirements and maneuvering the vehicle along this trajectory.
[0079] In this method, rectangular neighborhood search is a search method based on rectangular areas. Its advantage is that the search space is relatively small and the optimal solution can be found quickly. However, when the search space is large and the search situation is complex, this method may not find the global optimal solution.
[0080] Trapezoidal neighborhood search is a search method based on trapezoidal regions. Its advantage is that it can consider more neighborhood information and can better find the optimal solution in a larger search space and less complex situations. However, it will cover a wider range and the amount of calculation is relatively large, which also affects the real-time performance of the calculation to a certain extent.
[0081] After adopting this method, based on the environmental scenario, the autonomous driving system switches to a rectangular neighborhood search method when the space is small and the environment is simple, and switches to a trapezoidal neighborhood search method when the space is large and the environment is complex. This achieves a selective and targeted search for safe neighborhoods, taking into account the advantages of both search methods while avoiding the impact of their disadvantages.
[0082] If the above method is applied to the vehicle's reversing state, especially the reversing state, the rectangular neighborhood search method can be replaced by the annular neighborhood search method, and the annular neighborhood and the trapezoidal neighborhood can be combined as a variable neighborhood search method in the reversing state. In addition, when the vehicle steering angle is less than a certain threshold (for example, 30 degrees) when changing direction or turning, the rectangular neighborhood search method can also be replaced by the fan-shaped neighborhood search method. The fan-shaped neighborhood search method is suitable for turning and changing direction. Although it depends on the initial solution, it can quickly locate the target area, and its search algorithm is simple and can quickly generate candidate solutions.
[0083] The present invention is described above by way of example in conjunction with the accompanying drawings. It is obvious that the specific implementation of the present invention is not limited to the above-mentioned method. As long as various non-substantial improvements are made using the inventive concept and technical solution of the present invention, or the inventive concept and technical solution are directly applied to other occasions without improvement, they are all within the protection scope of the present invention.
Claims
1. A search rule base design method for autonomous driving trajectory generation, Features: The following steps are involved: S1. Collect environmental information around the vehicle and identify roads, obstacles and traffic signs; S2, judging the scene where the vehicle is located based on the recognition results of roads, obstacles and traffic signs; S3, performing variable neighborhood search according to neighborhood search rules based on driving status and environmental scenarios; A rule base for converting the neighborhood search method is pre-established in a controller of the vehicle's automatic driving, wherein the rule base stores driving rules and conversion rules, wherein the driving rules include vehicle dynamics parameter constraints, and the conversion rules are associated with the vehicle's driving state and environmental scenes, and the shapes of the safety neighborhood are classified into at least two types; S4, generating a trajectory candidate set based on the variable neighborhood search results; S5. Evaluate the candidate trajectory set and select the trajectory that meets the requirements.
2. A search rule base design method for autonomous driving trajectory generation according to claim 1, Features: A rule base for converting neighborhood search methods is pre-established in a controller of the vehicle's automatic driving, wherein the rule base stores driving rules and conversion rules, wherein the driving rules include vehicle dynamics parameter constraints, and the conversion rules are associated with the vehicle's driving state and environmental scenarios; before establishing the conversion rules, a reliable environmental scenario is first established to simulate the working state of the work vehicle, thereby determining a neighborhood search method suitable for different environmental scenarios.
3. A search rule base design method for autonomous driving trajectory generation according to claim 2, Features: In step S2, the environmental scene includes: Simple straight road scenario: The vehicle travels in a straight line, the road environment is simple, there are no intersections and few obstacles; Intersection scenario: There is an intersection, and you need to pay attention to traffic dynamics in multiple directions; Fast lane scenario: In the fast lane, the vehicle is traveling at a high speed and you only need to keep a safe distance from the vehicle in front. Parking area scenario: In parking areas, vehicles need to frequently change lanes, turn, and avoid other vehicles; Temporary traffic sign scenario: There are temporary traffic signs on the road, and you need to pay close attention to the layout and changes of the construction area, as well as temporary traffic control measures; Scenarios in crowded areas: In crowded areas, vehicles need to pay special attention to the movements of pedestrians and non-motor vehicles.
4. A search rule base design method for autonomous driving trajectory generation according to claim 3, Features: In step S3, the conversion rules of the vehicle in this normal driving state include: using a rectangular neighborhood search method in simple straight road scenarios and fast lane scenarios, and using a larger trapezoidal neighborhood search method in intersection scenarios, complex road scenarios with traffic signals, parking area scenarios, temporary traffic sign scenarios and crowded areas with people and vehicles.
5. A search rule base design method for autonomous driving trajectory generation according to claim 3, Features: In step S3, the conversion rule of the vehicle in the reversing state includes: combining the annular neighborhood and the trapezoidal neighborhood as a variable neighborhood search method in the reversing state in the parking area scene.
6. A search rule base design method for autonomous driving trajectory generation according to claim 3 or 4, Features: In step S3, the conversion rule of the vehicle in the normal driving state includes: when the steering angle of the vehicle is less than a certain threshold when changing direction or turning, a variable neighborhood search is performed using a fan-shaped neighborhood search method.
7. A search rule base design method for autonomous driving trajectory generation according to claim 2, Features: In the step S4, after determining the safe neighborhood search method for different areas in the vehicle's surrounding environment, a vehicle body coordinate system is established; the width of the rectangular search neighborhood is determined according to the width of the work vehicle, and the width of the smaller base of the trapezoidal neighborhood is kept the same as the width of the rectangular search neighborhood; a mathematical model is established to determine the neighborhood range that the vehicle's driving trajectory can pass through in the coordinate system; the length of the feasible domain search is determined according to the determined neighborhood width and the geometric characteristics of the rectangular neighborhood; the initial point and end point of the trajectory in the corresponding safe neighborhood are determined by the area divided by the variable neighborhood search result, and a trajectory candidate set is generated through the currently adapted trajectory planning algorithm based on the determined initial point positions and end point positions.
8. A search rule base design method for autonomous driving trajectory generation according to claim 7, Features: In step S4, the specific algorithm for generating a trajectory candidate set by means of a trapezoidal neighborhood search is as follows: The mathematical expression of the trapezoidal neighborhood is O(x 0 ,y 0 ) represents the initial position of the smart car, and the side of the neighborhood l 1 , l 2 The length of the trapezoidal neighborhood is r, and the length of the top side is l 3 , the height is s, the length of the base is d, then the expression of the trapezoidal safe neighborhood is as follows: Then, the initial and end positions of the trajectory in each safe neighborhood are determined, and the trajectory is planned using a quintic polynomial, and the following quintic polynomial curve formula is defined: Among them, x and y are the position coordinates of the vehicle in the Cartesian coordinate system: 0 ,a 1 ,…,a s and b 0 ,b 1 ,…,b s is the unknown coefficient, and the vehicle obstacle avoidance start time is t 0 , the vehicle position, velocity, and acceleration at the beginning are x 0 ,y 0 ; The obstacle avoidance end time is t 1 , the vehicle parameter at the end time is x 1 ,y 1 ; Therefore, the boundary condition of the quintic polynomial path can be obtained: Combining the above equations, we can solve the unknown coefficients of the quintic polynomial and obtain the trajectory posterior set.
Citation Information
Cited By
Urban scene-oriented automatic driving variable neighborhood searching and trajectory planning method
CN119840661A
Vehicle track generation method and device based on variable neighborhood search, and medium
CN121469637A
