An unmanned vehicle path planning method and system

Abstract

The present application relates to a kind of unmanned vehicle path planning method and system, belong to vehicle path planning technical field.The present application is under the influence of one segment global path when obstacle, if obstacle can form a continuous position sequence, the position sequence is set as virtual boundary, in the case where it meets global path target direction, according to virtual boundary navigation, generate a new path along the virtual boundary fixed safety distance.By this way, the planning number of unmanned vehicle can be effectively reduced when encountering long obstacle in depth, the calculation efficiency is improved, and the vehicle is quickly bypassed obstacle.At the same time, for discrete obstacle, according to multiple evaluation indexes, an optimal path is selected from the alternative tentacle path, the rationality of optimal path selection is ensured, and safety and traffic efficiency can be considered.

Description

Technical Field

[0001] This invention relates to an unmanned vehicle path planning method and system, belonging to the field of vehicle path planning technology. Background Technology

[0002] Path planning is one of the core functions of autonomous vehicles. It is primarily responsible for planning the safest and most efficient optimal path for the vehicle. Path planning is generally divided into two types: global path planning and local path planning. Global path planning, also known as static planning, is mainly responsible for planning a global path from the starting point to the destination, guiding the autonomous vehicle to its final destination. Local path planning, also known as dynamic planning, is mainly responsible for planning obstacle avoidance paths when encountering obstacles. Local path planning often plans several paths at once, and then selects the safest, collision-free optimal path as the obstacle avoidance path based on an evaluation method.

[0003] In recent years, with the advancement of internet and intelligent vehicle technologies, the application scenarios for autonomous vehicles have gradually diversified, with applications such as unmanned delivery gradually expanding in some industrial parks. However, the phenomenon of haphazard parking of motor vehicles and non-motor vehicles on the roadside in some industrial parks poses new challenges to intelligent vehicle path planning. Faced with complex traffic conditions, a comparison of the characteristics of various algorithms reveals that a single path planning algorithm is insufficient to meet current safety and speed requirements. Furthermore, when avoiding obstacles, such as sequentially placed vehicles, autonomous vehicles often need to plan multiple times. For example, multiple alternative paths are generated each time an obstacle is encountered. If there are many obstacles, multiple planning processes will be required, resulting in low efficiency. Summary of the Invention

[0004] The purpose of this invention is to provide an unmanned vehicle path planning method and system to solve the problems of excessive planning times and low efficiency in the current path planning process.

[0005] To solve the above-mentioned technical problems, this invention provides an unmanned vehicle path planning method, which includes the following steps:

[0006] 1) Obtain the real-time location of the vehicle and plan a global path based on the vehicle's current location and target location;

[0007] 2) Determine whether obstacle avoidance is needed when using the global path as the next path based on the vehicle's current location and the surrounding environment.

[0008] 3) If obstacle avoidance is required, determine whether the obstacles to be avoided form a continuous sequence of positions;

[0009] 4) If possible, use the resulting position sequence as a virtual boundary, and control the vehicle to travel along the virtual boundary according to the set safe distance;

[0010] 5) If not, select the optimal path from the alternative tendril paths corresponding to the current vehicle speed according to the evaluation index, and control the vehicle operation based on the selected optimal path.

[0011] 6) Determine whether the vehicle has successfully avoided the obstacle, and return to the global path after successfully avoiding the obstacle;

[0012] 7) Repeat steps 2)-6) until the destination is reached.

[0013] Through the above process, when an obstacle affects a segment of the global path, if the obstacle can form a continuous position sequence, this position sequence is set as a virtual boundary. Navigation is then performed based on this virtual boundary, while still conforming to the target direction of the global path, generating a new path along the virtual boundary at a fixed safe distance. This method effectively reduces the number of planning steps required for autonomous vehicles when encountering obstacles with long depths, improves computational efficiency, and helps vehicles quickly bypass obstacles. Simultaneously, for discrete obstacles, an optimal path is selected from candidate tentacle paths based on evaluation metrics, ensuring the rationality of the optimal path selection.

[0014] Furthermore, the evaluation index includes at least two of the following: a safety distance parameter, a lateral safety parameter, a trajectory consistency parameter, and a trajectory angle parameter. The safety distance parameter is used to characterize the distance of obstacles on the candidate tentacle path; the lateral safety parameter is used to characterize the distance of obstacles around the candidate tentacle path; the trajectory consistency parameter is used to characterize the degree of fit between the candidate tentacle path and the global path; and the trajectory angle parameter is used to characterize the amount of angle change between the current candidate tentacle path and the previous path segment.

[0015] This invention selects the optimal path by designing multiple evaluation indicators, taking into account the influence of conventional safety distance parameters and vehicle width, as well as the degree of fit with the global path and the connectivity between adjacent path segments, so that the selected optimal path can balance driving safety and traffic efficiency.

[0016] Furthermore, the calculation formula used for the lateral safety parameters is as follows:

[0017]

[0018]

[0019] Where v flatness For side safety parameters, v flatness The smaller the value, the greater the distance the vehicle travels laterally. (C) flatnessThe influence factor is w(o), which is the sum of the weights of all obstacle points in the neighborhood of the candidate tentacle path, and ∑w is the sum of the weights of all sampling points in the neighborhood of the candidate tentacle path. The weight of a sampling point is inversely proportional to its distance from the candidate tentacle path.

[0020] This invention constructs lateral safety parameters based on the distribution of obstacles in the alternative tendril paths, so that the constructed lateral safety parameters can accurately reflect the distribution of obstacles in the alternative tendril paths, ensuring that the optimal path selected subsequently can safely guarantee vehicle travel.

[0021] Furthermore, the trajectory consistency parameter is calculated using the following formula:

[0022]

[0023] v dist =a+c α α

[0024] Where v trajectory Here, is the trajectory consistency parameter, 'a' is the distance between the candidate tentacle path and the global path, 'α' is the angle between the candidate tentacle path and the global path, and 'v' is the distance between the candidate tentacle path and the global path. max and v min The velocity range of the alternative tentacle paths is v. dist The maximum and minimum values, c α For coefficient factors.

[0025] This invention constructs trajectory consistency parameters based on the distance between alternative tendril paths and the global path to ensure that the selected optimal path is close to the optimal path, thereby improving vehicle traffic efficiency.

[0026] Furthermore, the formula for calculating the safety distance parameter is as follows:

[0027]

[0028]

[0029] Where v clearance (l o The shortest distance l from the alternative tentacle path to the obstacle. o The safe distance parameter, c clearance It is the impact factor.

[0030] This invention constructs a safe distance parameter that reflects the distance to obstacles, making the subsequently selected optimal path sufficiently safe.

[0031] Furthermore, the formula for calculating the trajectory angle parameter is as follows:

[0032]

[0033] Where v angle Here, θ is the trajectory angle parameter, where θ is the angle between the candidate tentacle path and the previous path segment.

[0034] This invention constructs trajectory angle parameters that reflect the angle between the candidate tentacle path and the previous path segment, thus avoiding situations where a planned step length cannot successfully avoid obstacles.

[0035] Furthermore, the optimal path in step 5) is determined based on the weighted sum of the evaluation indicators.

[0036] Furthermore, the weights of each evaluation indicator are determined based on actual needs. When safety requirements are high, the weights of the safety distance parameter and the lateral safety parameter are increased. When traffic efficiency requirements are high, the weights of the trajectory consistency parameter and the trajectory angle parameter are increased.

[0037] This invention selects candidate tendril paths by weighting and summing various evaluation indicators, which can comprehensively consider the influence of each indicator. At the same time, it can also adjust the weight of each evaluation indicator according to the actual application scenario to meet the specific requirements of each actual scenario.

[0038] Furthermore, in step 3), the continuous position sequence refers to the spatial distance between each obstacle being within a set range.

[0039] This invention determines whether a continuous position sequence can be formed based on the spatial distance between obstacles, enabling rapid and accurate determination of continuous position sequences.

[0040] The present invention also provides an unmanned vehicle path planning system, which includes an environmental perception module, a decision planning module, and a control execution module. The environmental perception module is used to acquire information about the surrounding environment and the vehicle's location, and send them to the decision planning module. The control execution module is used to control the vehicle's operation according to the path information planned by the decision planning module. The decision planning module is used to execute the unmanned vehicle path planning method of the present invention to perform path planning.

[0041] This invention addresses the issue of obstacles affecting a global path. If the obstacles form a continuous sequence of positions, this sequence is designated as a virtual boundary. Navigation is then performed based on this virtual boundary, ensuring it aligns with the global path's target direction, generating a new path along this boundary at a fixed safe distance. This approach effectively reduces the number of planning steps required for autonomous vehicles encountering long, deep obstacles, improving computational efficiency and enabling vehicles to quickly bypass obstacles. Furthermore, for discrete obstacles, an optimal path is selected from candidate paths based on evaluation metrics, ensuring the rationality of the optimal path selection. Attached Figure Description

[0042] Figure 1 This is a flowchart of the unmanned vehicle path planning method of the present invention;

[0043] Figure 2 This is a schematic diagram of the unmanned vehicle path planning system of the present invention;

[0044] Figure 3 This is a schematic diagram of the tentacle algorithm used in this invention;

[0045] Figure 4 This is a schematic diagram of driving trajectory parameters in an embodiment of the present invention;

[0046] Figure 5 This is a schematic diagram of the trajectory angle difference in an embodiment of the present invention. Detailed Implementation

[0047] The specific embodiments of the present invention will be further described below with reference to the accompanying drawings.

[0048] Example of Autonomous Vehicle Path Planning Method

[0049] This invention first determines whether obstacle avoidance is needed when using the global path as the next segment of the path based on the vehicle's current position and the surrounding environment. If obstacle avoidance is needed, it determines whether the obstacles to be avoided form a continuous position sequence. If so, the formed position sequence is used as a virtual boundary, and the vehicle is controlled to travel along this virtual boundary according to a set safety distance. If not, the optimal path is selected from the alternative trail paths corresponding to the vehicle's current speed according to evaluation indicators, and the vehicle is controlled to run based on the selected optimal path. The invention then determines whether the obstacle avoidance was successful and, upon successful obstacle avoidance, returns to the global path. The specific implementation process of this method is as follows: Figure 1 As shown, the implementation process will be explained in detail below.

[0050] 1. Obtain the real-time location of the vehicle and the global path planned based on the vehicle's current location and target location.

[0051] The unmanned vehicles of this invention generally refer to vehicles operating on a site, such as a factory area or industrial park. To facilitate accurate subsequent judgment, this invention equates the operating site of the unmanned vehicle to a rectangle and establishes a corresponding Cartesian coordinate system to achieve the gridding of the operating site. For example, its lower right corner is designated as the origin o, and the two sides connected to the origin o are the x-axis and y-axis, respectively, thereby establishing a Cartesian coordinate system oxy; in this Cartesian coordinate system oxy, the operating site is divided into an M×N grid map.

[0052] Autonomous vehicles are typically equipped with an environmental perception module to acquire information about their surroundings and their location. This module can use radar and vision sensors to obtain environmental information and IMU and GPS devices to acquire vehicle location information. Additionally, autonomous vehicles are equipped with vehicle status sensors to acquire vehicle status information, such as speed and orientation. Based on the vehicle's current position and target position, the autonomous vehicle performs global path planning. This embodiment uses the A* algorithm for global path planning, constructing a global path that guides the overall direction of the autonomous vehicle.

[0053] The A* algorithm's principle formula is F(n) = G(n) + H(n), where F(n) represents the estimated cost of point n, G(n) represents the actual cost from the starting point to point n, and H(n) represents the estimated cost from point n to the target point. Since actual paths rarely reach a target point in a straight line, this invention uses the Manhattan distance formula in the heuristic function part of F(n):

[0054] l = |x1 - x2| + |y1 - y2|

[0055] Where (x1, y1) are the coordinates of the vehicle's current point n, and (x2, y2) are the coordinates of the vehicle's target point.

[0056] 2. Determine whether obstacle avoidance is required when the vehicle uses the global path as the next segment of the path.

[0057] Step 1 obtains information about the vehicle's surrounding environment, primarily the presence of obstacles. This information can be acquired using LiDAR and cameras installed on the vehicle. In this embodiment, the operating area is gridded, allowing for the assessment of obstacles at each grid point. Cn represents the passability of the corresponding grid point (nx, ny), where Cn = 1 indicates an obstacle is present and Cn = 0 indicates no obstacle is present. Based on the passability of grid points within the next global path's defined range, it is determined whether this will affect normal communication within that global path segment. If any grid point within the next global path's defined range is considered non-communicable, then obstacle avoidance is required for the next global path segment. The defined range is determined based on the vehicle's size and the required safety distance.

[0058] 3. If obstacle avoidance is required, determine whether the obstacles to be avoided form a continuous sequence of positions.

[0059] When it is determined that an obstacle will affect the next segment of the global path of a vehicle, if there are multiple obstacles (at least three), it is determined whether these obstacles can form a continuous position sequence. Here, a continuous position sequence means that the spatial positions are adjacent, that is, the distance between any two adjacent obstacles is less than a set distance. For example, vehicles parked consecutively in the venue may form a continuous position sequence.

[0060] 4. Use the resulting position sequence as a virtual boundary, and control the vehicle to travel along the virtual boundary according to the set safe distance.

[0061] In the case of multiple obstacles forming a continuous position sequence, in order to effectively reduce the number of planning steps for autonomous vehicles when encountering obstacles with long depths, improve computational efficiency, and help vehicles quickly bypass obstacles, this invention virtualizes the formed continuous position sequence as a virtual boundary. Based on this virtual boundary, the vehicle performs path planning. Specifically, during planning, a path along the virtual boundary and maintaining a set safe distance from the virtual boundary is selected as the vehicle's next driving path.

[0062] The safety distance in this invention can be determined in two ways. One is to set the safety distance to 1.5 meters based on the vehicle's trajectory-keeping capability, positioning error, and braking energy. The other is to establish a safety distance model that allows the safety distance to be adaptively adjusted according to the vehicle's speed. This safety distance model is expressed as follows:

[0063]

[0064] Where e is the safety distance, in meters; v i The speed is expressed in m / s.

[0065] 5. For discrete obstacles, select the optimal path from the alternative tendril paths corresponding to the vehicle's current speed according to the evaluation index, and control the vehicle's operation based on the selected optimal path.

[0066] When detected obstacles do not form a continuous sequence, meaning the multiple obstacles affecting global driving are discrete, this invention employs a tentacle algorithm to generate multiple candidate tentacle paths, selecting the optimal one as the vehicle's next driving path. The tentacle algorithm is a type of biomimetic algorithm; its main principle is to mimic the tentacles of insects, forming a fan-shaped region in front of the vehicle's direction of travel, such as... Figure 3 and Figure 4 As shown, this invention enables real-time obstacle avoidance. Before the vehicle moves, the speed range is divided into 16 groups, and 81 tendril paths are calculated for each group. When in use, the candidate tendril path in the corresponding group is selected based on the vehicle's current speed.

[0067] The evaluation criteria are set, and the optimal path is selected from the candidate tendril paths based on the evaluation criteria. The evaluation criteria of this invention include at least two of the following: safety distance parameter, lateral safety parameter, trajectory consistency parameter, and trajectory angle parameter.

[0068] The safety distance parameter characterizes the distance to obstacles on the alternative cantilever path; if there are no obstacles, the value is 0. This parameter primarily ensures forward driving safety by judging the distance to obstacles and giving full consideration to the selection of cantilever paths to ensure that a sufficiently safe alternative cantilever path is chosen. The formula for calculating the safety distance parameter is:

[0069]

[0070]

[0071] Where v clearance (l o The shortest distance l from the alternative tentacle path to the obstacle. o The safe distance parameter, c clearance It is the impact factor.

[0072] Lateral safety parameters characterize the distance to obstacles surrounding candidate tentacle paths. In actual driving, autonomous vehicles are not point masses but have a certain width, so the influence of width must be considered when selecting tentacle paths. Obstacles at varying distances are likely distributed around the tentacles. Being too close to these obstacles may affect driving safety; therefore, a lateral safety parameter is introduced to further ensure the safety of candidate tentacle paths. The calculation formula used is:

[0073]

[0074]

[0075] Where v flatness For side safety parameters, v flatness The smaller the value, the greater the distance the vehicle travels laterally. (C) flatness The influence factor is w(o), which is the sum of the weights of all obstacle points in the neighborhood of the candidate tentacle path, and ∑w is the sum of the weights of all sampling points in the neighborhood of the candidate tentacle path. The weight of a sampling point is inversely proportional to its distance from the candidate tentacle path.

[0076] The trajectory consistency parameter characterizes the degree of fit between the candidate tendril path and the global path. The local path is a temporary path taken by the autonomous vehicle when encountering obstacles and needing to avoid them on its way to the target point. The ultimate goal is to reach the target point safely and efficiently. Therefore, the local path should not deviate too much from the global path so that it can quickly return to the global path after obstacle avoidance. The trajectory consistency parameter is calculated using the following formula:

[0077]

[0078] v dist =a+c α α

[0079] Where v trajectory Here, is the trajectory consistency parameter, 'a' is the distance between the candidate tentacle path and the global path, 'α' is the angle between the candidate tentacle path and the global path, and 'v' is the distance between the candidate tentacle path and the global path. max and v min The velocity range of the alternative tentacle paths is v. dist The maximum and minimum values, c α For coefficient factors.

[0080] The trajectory angle parameter characterizes the angle change between the current candidate tentacle path and the previous path segment. During obstacle avoidance, autonomous vehicles may encounter situations where a single planned step is insufficient to successfully avoid an obstacle, requiring multiple planning steps to successfully navigate around it. If the angle change between the end point of the previous path and the starting point of the next path is too large, the transition will be less smooth. Inappropriate heading changes and excessive steering during obstacle avoidance can even lead to brief stops, affecting driving safety and reducing traffic efficiency. Figure 5 As shown. The formula for calculating the trajectory angle parameter is:

[0081]

[0082] Where v angle Here, θ is the trajectory angle parameter, where θ is the angle between the candidate tentacle path and the previous path segment.

[0083] Based on the above evaluation indicators, this invention evaluates each candidate tentacle path. In this embodiment, the weighted sum of each evaluation indicator is calculated as the final evaluation parameter v. combined The specific calculation formula is as follows:

[0084] v combined =av clearance +bv flatness +cv trajectory +dv angle

[0085] Where a, b, c, and d represent the weights of each evaluation indicator. The weights of each evaluation indicator are determined based on actual needs. When safety requirements are high, such as when operating in densely populated areas like campuses or factories, the weights of the safety distance parameter and the lateral safety parameter can be increased. When traffic efficiency requirements are high, such as when operating in sparsely populated areas, the weights of the trajectory consistency parameter and the trajectory angle parameter can be increased.

[0086] Based on the evaluation parameter v combinedEach candidate tentacle path can be evaluated, and the optimal path (v) can be selected. combined The shortest path is selected as the optimal path and sent to the vehicle actuator. The vehicle actuator then controls the vehicle for the next segment according to this optimal path.

[0087] 6. Determine whether the vehicle has successfully avoided the obstacle, and return to the global path after successfully avoiding the obstacle.

[0088] Obstacle avoidance can be achieved under different circumstances based on steps 4 and 5. After the vehicle travels along the path planned in step 4 or step 5, it is necessary to determine whether the vehicle has successfully avoided the obstacle. That is, whether there is a risk of collision when the vehicle returns to the global path. If there is no risk, it means that the obstacle avoidance was successfully achieved through the path planning in step 4 or step 5. At this time, the vehicle can be controlled to return to the global path.

[0089] 7. Repeat steps 2-6 until the target location is reached.

[0090] During the autonomous vehicle's operation, the determination of the driving path is completed at set intervals. If the vehicle returns to the global path, it returns to step 2 to re-determine whether obstacle avoidance is needed, and repeats the above process until the vehicle reaches the target location.

[0091] Through the above process, this invention can group obstacle locations. If these locations form a continuous sequence, they are virtualized as a virtual boundary. Then, while conforming to the global path target direction, navigation is performed based on the virtual road boundary to generate a new path along the virtual boundary at a fixed safe distance. This method can effectively reduce the number of planning steps for autonomous vehicles when encountering obstacles with long depths, improve computational efficiency, and help vehicles quickly bypass obstacles. Simultaneously, for discrete obstacles, multiple evaluation indicators are designed to select the optimal path from candidate paths, ensuring that the selected optimal path balances driving safety and traffic efficiency.

[0092] Example of an autonomous vehicle path planning system

[0093] The unmanned vehicle path planning system of the present invention includes an environmental perception module, a decision planning module, and a control execution module, such as... Figure 2 As shown in the diagram. The environmental perception module acquires information about the vehicle's surrounding environment and its location, and sends this information to the decision-making and planning module. The control execution module controls the vehicle's operation based on the path information planned by the decision-making and planning module. The decision-making and planning module executes the autonomous vehicle path planning method of this invention to perform path planning. The specific process of path planning by the decision-making and planning module has been described in detail in the embodiments of the autonomous vehicle path planning method, and will not be repeated here.

[0094] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are 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. An unmanned vehicle path planning method, characterized by, The method comprises the following steps: 1) obtaining a real-time position of the vehicle, planning a global path based on a current position and a target position of the vehicle; 2) determining whether the global path needs to be avoided as a next path according to the current position of the vehicle and environment information where the vehicle is located; 3) if the global path needs to be avoided, determining whether the obstacles needing to be avoided form a continuous position sequence; 4) if yes, taking the formed position sequence as a virtual boundary, and controlling the vehicle to travel along the virtual boundary according to a set safety distance, wherein the safety distance is adaptively adjusted according to the speed of the vehicle; 5) if not, select the optimal path from the candidate whisker paths corresponding to the current speed of the vehicle according to evaluation indexes, and control the vehicle to run based on the selected optimal path; the evaluation indexes include at least two of the following: a safety distance parameter , a side safety parameter , a trajectory consistency parameter , and a trajectory angle parameter ​ 6) determining whether the vehicle successfully avoids the obstacles, and returning to the global path to travel after successfully avoiding the obstacles; 7) repeating steps 2) to 6) until the end point is reached; The calculation formula is: is an influence factor, is a sum of weights of all obstacle points in the neighborhood of the alternative whisker path, is a sum of weights of all sampling points in the neighborhood of the alternative whisker path; The calculation formula is as follows: the distance of the alternative whisker path to the global path, the angle of the alternative whisker path to the global path, and the maximum and minimum values, respectively, within the speed interval in which the alternative whisker path is located, the maximum and minimum values, respectively, within the speed interval in which the alternative whisker path is located, a coefficient factor; The calculation formula is: for an alternative whisker path to the closest distance to the obstacle, is an impact factor; The calculation formula is: is the angle between the alternative whisker path and the previous path. 2.The method of claim 1, wherein, The safety distance parameter is used to represent the distance of the obstacles on the candidate whisker path, the side safety parameter is used to represent the distance of the obstacles around the candidate whisker path, the trajectory consistency parameter is used to represent the fitting degree of the candidate whisker path and the global path, and the trajectory angle parameter is used to represent the angle change between the current candidate whisker path and the previous path. 3.The method of claim 1, wherein, Whether the next global path needs to be avoided is determined based on the passable state of the grid points in the set range of the next global path, and if there are grid points in the set range of the next global path in an impassable state, it is considered that the next global path needs to be avoided. 4.The method of claim 1, wherein, Global paths are derived using algorithmic planning. 5.The method of claim 1, wherein, The safety distance is calculated by using a safety distance model, and the safety distance model is: where e is the safety distance in meters, v i is the vehicle speed in m / s. 6.The method of claim 1, wherein, The candidate whisker path is generated by using a whisker algorithm.

7. The unmanned vehicle path planning method of any one of claims 1-6, wherein, The optimal path in step 5) is determined according to the size of the weighted sum of each evaluation index. 8.The method of claim 7, wherein, The size of the weight of each evaluation index is determined according to actual requirements, when the safety requirement is higher, the weight of the safety distance parameter and the side safety parameter is increased, and when the travel efficiency requirement is higher, the weight of the trajectory consistency parameter and the trajectory angle parameter is increased. 9.The method of claim 1, wherein, The continuous position sequence in step 3) refers to the spatial distance between the obstacles being within a set range.

10. An unmanned vehicle path planning system, characterized by, The system comprises an environment perception module, a decision planning module and a control execution module, the environment perception module is used to obtain the surrounding environment information of the vehicle and the vehicle positioning information, and send them to the decision planning module; the control execution module is used to control the vehicle to run according to the path information planned by the decision planning module; and the decision planning module is used to perform the path planning method of any one of claims 1-9.

Images