Path planning method and device, nonvolatile storage medium and electronic equipment

Abstract

The invention discloses a path planning method and device, a nonvolatile storage medium and electronic equipment. The method comprises the following steps: acquiring first motion state data of a target vehicle in a rectangular coordinate system and second motion state data of an obstacle vehicle of the target vehicle in the rectangular coordinate system; converting the first motion state data into first target motion state data in a curve coordinate system, and converting the second motion state data into second target motion state data in the curve coordinate system; predicting motion state data of the obstacle vehicle at a second moment in the curve coordinate system; determining the static risk field intensity and the dynamic risk field intensity of the obstacle vehicle; and determining a target path corresponding to the target vehicle according to the first target motion state data and the total collision risk. According to the method and the device, the technical problem of inaccurate path planning when the vehicle faces the obstacle vehicle caused by insufficient accuracy of predicting the motion state of the obstacle vehicle and incapability of comprehensively evaluating the collision risk is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of unmanned path planning, in particular, to a path planning method and device, a non-volatile storage medium and an electronic device. BACKGROUND

[0002] With the development of intelligent transportation systems, especially the increasing maturity of autonomous driving technology, ensuring the safe driving of vehicles in various complex environments has become a research focus. On the highway, vehicles need to face rapidly changing traffic flow, including but not limited to the acceleration, deceleration, lane changing of the preceding vehicle and other unexpected situations. These dynamic changes pose high requirements on the path planning of autonomous vehicles, which need to be able to instantly identify and predict the motion state of the obstacle vehicle in order to make safe and efficient driving decisions.

[0003] Most related path planning methods are based on obstacle avoidance logic of static obstacles, for example, using a rectangular coordinate system to describe the position and motion of the vehicle. This coordinate system is suitable for simple and static obstacle environments. However, when facing dynamic obstacle vehicles in high-speed driving scenarios, the limitations of these methods become apparent. Since the rectangular coordinate system is difficult to directly express the motion of the vehicle along a specific road, there may be a large deviation when predicting the future position of the obstacle vehicle. Especially when the obstacle vehicle performs complex actions such as sudden lane changing and emergency braking, the prediction model based on the rectangular coordinate system often fails to update the prediction results in time, resulting in lag in updating the path planning and failing to quickly respond to rapidly changing road conditions.

[0004] In addition, for risk assessment of dynamic obstacle vehicles, related methods rely on the instantaneous position and speed of the obstacle vehicle, while ignoring its future motion trend and potential threat level. This simplified assessment method is difficult to comprehensively and accurately predict the collision threat that the obstacle vehicle may pose to the autonomous vehicle. In particular, on the highway, the relative speed between vehicles is fast, and even a small change in motion state within a short period of time can bring about a huge risk difference, therefore, a method is needed that can prospectively assess risks to ensure that path planning not only considers the current position of the obstacle vehicle, but also predicts its future motion state, thereby better avoiding potential collision risks.

[0005] In view of the above problems, no effective solutions have been proposed so far. SUMMARY

[0006] The present application provides a path planning method, device, non-volatile storage medium and electronic device to at least solve the technical problem of inaccurate path planning of vehicles when facing obstacle vehicles due to insufficient accuracy in predicting the motion state of obstacle vehicles and the inability to comprehensively assess collision risks.

[0007] According to an aspect of the present application, a path planning method is provided, comprising: obtaining first motion state data of a target vehicle in a rectangular coordinate system and second motion state data of an obstacle vehicle of the target vehicle in the rectangular coordinate system; converting the first motion state data into first target motion state data in a curvilinear coordinate system, and converting the second motion state data into second target motion state data in the curvilinear coordinate system, wherein an origin of the curvilinear coordinate system is a projection point of the target vehicle on a reference trajectory, the reference trajectory is used to represent an ideal motion path of the target vehicle, and the projection point is a point on the reference trajectory closest to the target vehicle; predicting motion state data of the obstacle vehicle at a second time in the curvilinear coordinate system according to motion state data of the obstacle vehicle at a first time in the curvilinear coordinate system, wherein the second time is a future time of the first time; determining static risk field intensity and dynamic risk field intensity of the obstacle vehicle according to the motion state data of the obstacle vehicle at the second time, wherein the static risk field intensity is used to evaluate the intensity of a collision threat posed by the obstacle vehicle in a static state to the target vehicle, and the dynamic risk field intensity is used to evaluate the intensity of a collision threat posed by the obstacle vehicle in a dynamic state to the target vehicle; determining a total collision risk according to the static risk field intensity and the dynamic risk field intensity, and determining a target path corresponding to the target vehicle according to the first target motion state data and the total collision risk.

[0008] Optionally, converting the first motion state data into first target motion state data in the curvilinear coordinate system comprises: determining a longitudinal displacement of the target vehicle in the curvilinear coordinate system as 0; determining a longitudinal velocity of the target vehicle in the curvilinear coordinate system according to the speed of the target vehicle, the heading angle deviation, a curvature of the reference trajectory at the projection point, and a lateral displacement of the target vehicle relative to the reference trajectory; determining a derivative of the heading angle deviation of the target vehicle in the curvilinear coordinate system according to the curvature of the reference trajectory at the projection point, the lateral displacement of the target vehicle relative to the reference trajectory, and the heading angle deviation; determining a longitudinal acceleration of the target vehicle in the curvilinear coordinate system according to the acceleration of the target vehicle, the heading angle deviation, the longitudinal velocity, a first derivative of the lateral displacement, the derivative of the heading angle deviation, the curvature of the reference trajectory at the projection point, a first derivative of the curvature of the reference trajectory at the projection point, and the lateral displacement; determining a lateral displacement of the target vehicle in the curvilinear coordinate system according to the global coordinates of the target vehicle and the coordinates of the projection point, wherein a sign of the lateral displacement of the target vehicle in the curvilinear coordinate system is used to represent a relative direction between the target vehicle and the reference trajectory; determining a lateral velocity of the target vehicle in the curvilinear coordinate system according to the speed of the target vehicle and the heading angle deviation; determining a lateral acceleration of the target vehicle in the curvilinear coordinate system according to the longitudinal acceleration of the target vehicle in the curvilinear coordinate system, the longitudinal velocity, a second derivative of the lateral displacement, and the first derivative of the lateral displacement; determining a first derivative of the lateral displacement of the target vehicle in the curvilinear coordinate system according to the curvature of the reference trajectory at the projection point, the lateral displacement, and the heading angle deviation; and determining a second derivative of the lateral displacement of the target vehicle in the curvilinear coordinate system according to the curvature of the reference trajectory at the projection point, the first derivative of the curvature of the reference trajectory at the projection point, the lateral displacement, the first derivative of the lateral displacement, the heading angle deviation, and the derivative of the heading angle deviation.

[0009] Optionally, converting the second motion state data into second target motion state data in the curvilinear coordinate system comprises: determining a longitudinal displacement of the obstacle vehicle in the curvilinear coordinate system according to a distance and a direction between a projection point of the obstacle vehicle on the reference trajectory and a projection point of the target vehicle on the reference trajectory; determining a lateral displacement of the obstacle vehicle in the curvilinear coordinate system according to a global coordinate of the obstacle vehicle and a coordinate of the projection point of the obstacle vehicle on the reference trajectory, wherein a sign of the lateral displacement of the obstacle vehicle in the curvilinear coordinate system is used to represent a relative direction between the obstacle vehicle and the reference trajectory; determining a heading angle deviation of the obstacle vehicle in the curvilinear coordinate system according to a heading angle of the obstacle vehicle and a tangent angle at the projection point of the obstacle vehicle on the reference trajectory; determining a lateral velocity of the obstacle vehicle in the curvilinear coordinate system according to a velocity of the obstacle vehicle and the heading angle deviation of the obstacle vehicle in the curvilinear coordinate system; and determining a relative longitudinal velocity of the obstacle vehicle in the curvilinear coordinate system relative to the target vehicle according to the velocity of the obstacle vehicle, the heading angle deviation of the obstacle vehicle in the curvilinear coordinate system, a curvature of the reference trajectory at the projection point of the obstacle vehicle, the lateral displacement of the obstacle vehicle, and a longitudinal velocity of the target vehicle in the curvilinear coordinate system.

[0010] Optionally, determining the static risk field intensity and the dynamic risk field intensity of the obstacle vehicle according to the motion state data of the obstacle vehicle at the second time comprises: determining the static risk field intensity of the obstacle vehicle according to a mass, a length and a width of the obstacle vehicle, a relative distance between the target vehicle and the obstacle vehicle in a longitudinal direction, and a relative distance between the target vehicle and the obstacle vehicle in a lateral direction; and determining the dynamic risk field intensity of the obstacle vehicle according to the mass, the length and the width of the obstacle vehicle, the relative distance between the target vehicle and the obstacle vehicle in the longitudinal direction, the relative distance between the target vehicle and the obstacle vehicle in the lateral direction, a longitudinal velocity of the obstacle vehicle, a lateral velocity of the obstacle vehicle, a longitudinal displacement of the target vehicle in the curvilinear coordinate system, and a longitudinal displacement of the obstacle vehicle in the curvilinear coordinate system.

[0011] Optionally, determining the target path corresponding to the target vehicle according to the first target motion state data and the total collision risk comprises: setting a plurality of path points in the longitudinal direction of the curvilinear coordinate system at a preset distance according to the first target motion state data, and setting a plurality of path points in the lateral direction of the curvilinear coordinate system based on a center line of a lane; dividing the path points at different longitudinal positions into a plurality of intervals, and generating a plurality of candidate paths by connecting the path points of adjacent intervals; determining an offset cost corresponding to the candidate paths according to a lateral displacement of the target vehicle in the curvilinear coordinate system; constructing a planning objective function based on the total collision risk and the offset cost corresponding to the candidate paths; and determining the target path that minimizes the planning objective function among the plurality of candidate paths.

[0012] Optionally, based on the total collision risk and the deviation cost corresponding to the candidate path, a planning objective function is constructed, including: based on the total collision risk and the deviation cost corresponding to the candidate path, a lateral planning objective function is constructed; according to the relative distance deviation of the candidate path and the reference trajectory in the longitudinal direction, the relative speed deviation of the longitudinal speed of the candidate path and the preset reference speed, and the longitudinal acceleration variation of the candidate path, a longitudinal planning objective function is constructed; the lateral planning objective function and the longitudinal planning objective function are weighted and combined to obtain the planning objective function.

[0013] Optionally, according to the motion state data of the obstacle vehicle at the first time in the polar coordinate system, the motion state data of the obstacle vehicle at the second time in the polar coordinate system is predicted, including: inputting the motion state data of the obstacle vehicle at the first time in the polar coordinate system into the pre-trained long short-term memory network, controlling the memory degree of the current input information through the input gate in the long short-term memory network, controlling the retention degree of the historical state information through the forgetting gate, and controlling the output degree of the current time state information through the output gate; learning the motion law of the obstacle vehicle in the polar coordinate system by using the long short-term memory network to extract the time sequence features of the obstacle vehicle; based on the time sequence features, the predicted motion state data of the obstacle vehicle at the second time in the polar coordinate system is output through the full connection layer of the long short-term memory network.

[0014] According to still another aspect of the present application, a path planning device is also provided, comprising: an acquisition module configured to acquire first motion state data of a target vehicle in a rectangular coordinate system and second motion state data of an obstacle vehicle of the target vehicle in the rectangular coordinate system; a conversion module configured to convert the first motion state data into first target motion state data in a polar coordinate system, and convert the second motion state data into second target motion state data in the polar coordinate system, wherein the origin of the polar coordinate system is a projection point of the target vehicle on a reference trajectory, the reference trajectory is used to represent an ideal motion path of the target vehicle, and the projection point is the closest point on the reference trajectory to the target vehicle; a prediction module configured to predict motion state data of the obstacle vehicle at a second time in the polar coordinate system according to motion state data of the obstacle vehicle at a first time in the polar coordinate system, wherein the second time is a future time of the first time; a first determination module configured to determine static risk field intensity and dynamic risk field intensity of the obstacle vehicle according to the motion state data of the obstacle vehicle at the second time, wherein the static risk field intensity is used to evaluate the intensity of the collision threat posed by the obstacle vehicle in a static state to the target vehicle, and the dynamic risk field intensity is used to evaluate the intensity of the collision threat posed by the obstacle vehicle in a dynamic state to the target vehicle; and a second determination module configured to determine a total collision risk according to the static risk field intensity and the dynamic risk field intensity, and determine a target path corresponding to the target vehicle according to the first target motion state data and the total collision risk.

[0015] According to still another aspect of the present application, a non-transitory storage medium is also provided, which includes a stored program, wherein the program, when executed, controls a device in which the storage medium is located to perform the path planning method.

[0016] According to still another aspect of the present application, an electronic device is also provided, which includes a memory and a processor, the processor being configured to execute a program stored in the memory, wherein the program, when executed, performs the path planning method.

[0017] According to still another aspect of the present application, a computer program is also provided, wherein the computer program, when executed by a processor, implements the path planning method.

[0018] According to still another aspect of the present application, a computer program product is also provided, which includes a non-transitory computer readable storage medium, wherein the non-transitory computer readable storage medium stores a computer program, and the computer program, when executed by a processor, implements the path planning method.

[0019] In the present application, first motion state data of a target vehicle in a rectangular coordinate system and second motion state data of an obstacle vehicle of the target vehicle in the rectangular coordinate system are obtained; the first motion state data is converted into first target motion state data in a curvilinear coordinate system, and the second motion state data is converted into second target motion state data in the curvilinear coordinate system, wherein an origin of the curvilinear coordinate system is a projection point of the target vehicle on a reference trajectory, the reference trajectory is used to represent an ideal motion path of the target vehicle, and the projection point is a point on the reference trajectory closest to the target vehicle; the motion state data of the obstacle vehicle at a second time in the curvilinear coordinate system is predicted according to the motion state data of the obstacle vehicle at a first time in the curvilinear coordinate system, wherein the second time is a future time of the first time; static risk field intensity and dynamic risk field intensity of the obstacle vehicle are determined according to the motion state data of the obstacle vehicle at the second time, wherein the static risk field intensity is used to evaluate the intensity of a collision threat posed by the obstacle vehicle in a static state to the target vehicle, and the dynamic risk field intensity is used to evaluate the intensity of a collision threat posed by the obstacle vehicle in a dynamic state to the target vehicle; total collision risk is determined according to the static risk field intensity and the dynamic risk field intensity, and a target path corresponding to the target vehicle is determined according to the first target motion state data and the total collision risk, thereby achieving the technical effects of improving the accuracy of predicting the future motion state of the obstacle vehicle and comprehensively quantifying the collision risk, and further solving the technical problem of inaccurate path planning of the vehicle when facing the obstacle vehicle due to insufficient accuracy of predicting the motion state of the obstacle vehicle and inability to comprehensively evaluate the collision risk. BRIEF DESCRIPTION OF DRAWINGS

[0020] The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiments of the application and together with the description serve to explain the application. In the drawings:

[0021] Figure 1 is a flow chart of a path planning method according to an embodiment of the application;

[0022] Figure 2 is a schematic diagram of coordinate conversion;

[0023] Figure 3 is a schematic diagram of coordinate conversion according to an embodiment of the application;

[0024] Figure 4 is a dynamic programming schematic diagram according to an embodiment of the application;

[0025] Figure 5 is a structural diagram of a path planning device according to an embodiment of the application;

[0026] Figure 6 is a hardware structural block diagram of a computer terminal of a path planning method according to an embodiment of the application. DETAILED DESCRIPTION

[0027] In order to enable persons skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work should fall within the scope of protection of the present application.

[0028] It should be noted that the terms "first", "second", and the like in the specification and claims of the present application and the above-described drawings are used to distinguish similar objects, and do not necessarily indicate a specific order or a chronological sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but can include other steps or units that are not clearly listed or inherent to the process, method, product or device.

[0029] According to the embodiment of the present application, a method embodiment of a path planning method is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described herein can be executed in an order different from that shown.

[0030] Figure 1 is a flowchart of a path planning method according to an embodiment of the present application, as shown in Figure 1 the method comprises the following steps:

[0031] Step S101, acquiring first motion state data of the target vehicle in the Cartesian coordinate system and second motion state data of the obstacle vehicle of the target vehicle in the Cartesian coordinate system.

[0032] Wherein, the Cartesian coordinate system is, for example, Cartesian coordinate system.

[0033] In step S101, the first motion state data of the target vehicle in the Cartesian coordinate system is collected, wherein the first motion state data includes but is not limited to the real-time position, speed, acceleration and other key information of the target vehicle. The first motion state data is expressed by x, y, z coordinates of the Cartesian coordinate system, which ensures the accuracy and integrity of the data. The second motion state data of the obstacle vehicle in the Cartesian coordinate system is collected, which includes but is not limited to the position, speed, acceleration and other dynamic parameters of the obstacle vehicle, which is also expressed by x, y, z coordinates, which ensures the comprehensive monitoring and description of the motion state of the obstacle vehicle.

[0034] Wherein, the first motion state data and the second motion state data can be obtained by various sensors and vehicle-mounted devices, such as radar, camera, GPS module, etc.

[0035] Step S102, converting the first motion state data into first target motion state data in the curvilinear coordinate system, and converting the second motion state data into second target motion state data in the curvilinear coordinate system, wherein the origin of the curvilinear coordinate system is the projection point of the target vehicle on the reference trajectory, the reference trajectory is used to represent the ideal motion path of the target vehicle, and the projection point is the nearest point on the reference trajectory to the target vehicle.

[0036] Wherein, the curvilinear coordinate system is, for example, Frenet coordinate system.

[0037] In step S102, the first motion state data of the target vehicle is converted into corresponding first target motion state data in the curvilinear coordinate system. The core of the conversion is to locate the origin of the curvilinear coordinate system, which is set as the projection point of the target vehicle on the reference trajectory. The reference trajectory represents the ideal driving route of the target vehicle, and the projection point is the point on the reference trajectory closest to the target vehicle. The dynamic information of the obstacle vehicle in the Cartesian coordinate system is converted into second target motion state data in the curvilinear coordinate system. This conversion step can effectively map the position, speed and acceleration information of the obstacle vehicle to the curvilinear coordinate system with the target vehicle projection point as the origin, facilitating subsequent risk assessment and path planning operations.

[0038] By converting the motion state data of the target vehicle and the obstacle vehicle into the curvilinear coordinate system, the driving characteristics of the vehicle along the curved road can be better adapted, and the errors that may be introduced by the Cartesian coordinate system when processing curved motion can be reduced. The introduction of the curvilinear coordinate system not only simplifies the calculation, but also improves the accuracy of the description of the vehicle motion state, especially in evaluating the relative position of the two vehicles and its dynamic change. The data in this coordinate system can provide more intuitive and accurate information.

[0039] In step S103, the motion state data of the obstacle vehicle at the second time in the curvilinear coordinate system is predicted based on the motion state data of the obstacle vehicle at the first time in the curvilinear coordinate system, wherein the second time is a future time of the first time.

[0040] In step S103, the motion state data of the obstacle vehicle at the second time in the curvilinear coordinate system is predicted based on the motion state data of the obstacle vehicle at the first time in the curvilinear coordinate system, wherein the second time is a future time of the first time.

[0041] The above prediction process is based on the curvilinear coordinate system, in which the origin of the coordinate is set as the projection point of the target vehicle on the reference trajectory. The advantage of this setting is that it can more closely associate the dynamics of the obstacle vehicle with the driving trajectory of the target vehicle, reduce the prediction error caused by the difference in coordinate systems, and enhance the real-time and accuracy of the prediction results.

[0042] Step S104, according to the motion state data of the obstacle vehicle at the second time, determine the static risk field intensity and the dynamic risk field intensity of the obstacle vehicle, wherein the static risk field intensity is used to evaluate the intensity of the collision threat posed by the obstacle vehicle in a static state to the target vehicle, and the dynamic risk field intensity is used to evaluate the intensity of the collision threat posed by the obstacle vehicle in a dynamic state to the target vehicle.

[0043] In step S104, according to the motion state data of the obstacle vehicle at the second time, i.e. at a certain time node in the future, the quantitative analysis of the static and dynamic risk field intensity posed by the obstacle vehicle is carried out. The evaluation of the static risk field intensity focuses on the degree of collision threat brought by the obstacle vehicle in a static state to the target vehicle. The calculation of the static risk field intensity mainly depends on the static properties of the obstacle vehicle such as mass, length, width and shape, and establishes a static risk field distribution related to the physical characteristics of the obstacle vehicle and the relative position of the target vehicle through Gaussian function or other appropriate mathematical models, so as to quantify the danger level that the static obstacle vehicle may cause.

[0044] On the other hand, the evaluation of the dynamic risk field intensity is a measure of the safety influence of the obstacle vehicle in motion on the target vehicle. The dynamic risk field intensity not only considers the static characteristics of the obstacle vehicle, but also takes into account the dynamic parameters such as longitudinal speed, lateral acceleration, heading angle change rate of the obstacle vehicle in the calculation. Through the dynamic risk field model, combined with the changes of the future predicted position, speed and direction of the obstacle vehicle, the immediate threat and potential avoidance difficulty brought by the moving obstacle vehicle to the target vehicle can be evaluated.

[0045] The determination of the static and dynamic risk field intensity is carried out in the ego curve coordinate system. By quantifying the static and dynamic risk field intensity, the potential collision danger can be evaluated comprehensively and in detail, and strong data support can be provided for the risk avoidance strategy in path planning, so as to ensure that the target vehicle can make safe and efficient driving decisions in complex traffic environment.

[0046] Step S105, according to the static risk field intensity and the dynamic risk field intensity, determine the total collision risk, and according to the first target motion state data and the total collision risk, determine the target path corresponding to the target vehicle.

[0047] Among them, the path planning process aims to find a path that can effectively reduce the total collision risk, and ensure the safety and smoothness of the target vehicle in the driving process. By integrating the risk value into the objective function of the optimization algorithm, such as using dynamic programming algorithm, the longitudinal path and the lateral path can be planned respectively.

[0048] In longitudinal planning, the distance, speed and acceleration of the vehicle are focused on to ensure that the vehicle can maintain an appropriate driving speed while meeting physical and road constraints, avoiding the risks caused by sudden acceleration and deceleration. In lateral planning, considering the need for obstacle avoidance and lane changing, an evaluation function can be designed to combine risk value and vehicle offset cost to ensure that the vehicle can avoid risks and maintain a comfortable and safe driving range when moving laterally. By using a multi-objective optimization strategy, the results of longitudinal and lateral planning can be coupled to generate an optimal target path that is safe, comfortable and efficient.

[0049] In summary, steps S101 and S102 achieve accurate collection and conversion of motion state data. Real-time motion state data of the target vehicle and its obstacle vehicles is obtained in the rectangular coordinate system, and then these data are converted to the curvilinear coordinate system with the projection point of the target vehicle on the reference trajectory as the coordinate origin. The conversion process optimizes the representation of the data, making it more suitable for subsequent dynamic planning and risk assessment algorithms, and improving the applicability and efficiency of the overall method.

[0050] Step S103 uses a prediction algorithm to predict the future motion state of the obstacle vehicle. The prediction is based not only on the current motion state of the obstacle vehicle, but also on its historical behavior patterns, enhancing the forward-looking nature of the path planning and enabling early response to uncertain behaviors of the obstacle vehicle, such as sudden lane changes or deceleration, thereby improving the decision-making quality and driving safety of the autonomous vehicle.

[0051] Step S104 uses the predicted motion state data to evaluate the intensity of static and dynamic risk fields. The static risk field is used to evaluate the collision threat posed by stationary obstacles, and the dynamic risk field is used to evaluate the safety hazards posed by moving obstacle vehicles to the target vehicle. By accurately calculating the intensity of the risk field, the potential risks in the driving environment can be more comprehensively understood, providing data support for decision-making.

[0052] In step S105, the total collision risk is determined by integrating the static and dynamic risk field intensities, and based on this risk and the motion state of the target vehicle in the curvilinear coordinate system, a dynamic programming algorithm is used to plan the optimal target path. The optimal target path not only seeks to avoid collisions, but also considers the physical constraints of the vehicle, road conditions and traffic efficiency, ensuring that the driving process is both safe and efficient.

[0053] Overall, the above steps constitute a closed-loop intelligent path planning system. From data collection, coordinate transformation, future prediction, risk assessment to path planning, each step revolves around the goal of improving vehicle driving safety and traffic efficiency. These steps are particularly suitable for high-speed scenarios, and their technical effect lies in enhancing the environmental perception capabilities of autonomous vehicles, improving the timeliness and accuracy of decision-making, thereby effectively reducing collision risks and ensuring stable vehicle operation in complex traffic environments. This represents a significant advancement in the safety and practicality of autonomous driving technology.

[0054] The following are Figure 1 The steps shown are illustrated and explained by way of example.

[0055] According to some optional embodiments of this application, converting the first motion state data into first target motion state data in a curvilinear coordinate system can be achieved by the following method: determining the longitudinal displacement of the target vehicle in the curvilinear coordinate system as 0; determining the longitudinal velocity of the target vehicle in the curvilinear coordinate system based on the target vehicle's velocity, heading angle deviation, curvature of the reference trajectory at the projection point, and lateral displacement of the target vehicle relative to the reference trajectory; determining the derivative of the heading angle deviation of the target vehicle in the curvilinear coordinate system based on the curvature of the reference trajectory at the projection point, lateral displacement of the target vehicle relative to the reference trajectory, and heading angle deviation; determining the longitudinal acceleration of the target vehicle in the curvilinear coordinate system based on the target vehicle's acceleration, heading angle deviation, longitudinal velocity, first derivative of lateral displacement, derivative of heading angle deviation, curvature of the reference trajectory at the projection point, first derivative of curvature of the reference trajectory at the projection point, and lateral displacement; and determining the longitudinal acceleration of the target vehicle based on the target vehicle's global... The lateral displacement of the target vehicle in the curvilinear coordinate system is determined using the coordinates of the reference trajectory and the coordinates of the projection point. The sign of the lateral displacement is used to indicate the relative direction between the target vehicle and the reference trajectory. The lateral velocity of the target vehicle in the curvilinear coordinate system is determined based on its velocity and heading angle deviation. The lateral acceleration of the target vehicle in the curvilinear coordinate system is determined based on its longitudinal acceleration, longitudinal velocity, the first derivative of its lateral displacement, and the second derivative of its lateral displacement. The first derivative of the lateral displacement of the target vehicle in the curvilinear coordinate system is determined based on the curvature of the reference trajectory at the projection point, its lateral displacement, and the heading angle deviation. The second derivative of the lateral displacement of the target vehicle in the curvilinear coordinate system is determined based on the curvature of the reference trajectory at the projection point, its first derivative, its lateral displacement, its first derivative, its heading angle deviation, and the derivative of its heading angle deviation.

[0056] In this embodiment, the longitudinal displacement of the target vehicle in the curvilinear coordinate system is first set to 0. This setting ensures that the target vehicle is always on the axis of the vehicle's Frenet coordinate system, simplifying subsequent calculations and enhancing the intuitive understanding of the vehicle's relative position.

[0057] Secondly, based on the target vehicle's speed, the deviation of the target vehicle's heading angle from the reference trajectory, and the curvature of the reference trajectory at the projection point, the longitudinal velocity of the target vehicle in the curvilinear coordinate system is calculated. The determination of the longitudinal velocity is based on the target vehicle's current driving state and the characteristics of the ideal driving path, providing dynamic longitudinal dimension information for path planning.

[0058] Then, using the curvature of the reference trajectory at the projection point, the lateral displacement of the target vehicle relative to the reference trajectory, and the heading angle deviation, the derivative of the heading angle deviation of the target vehicle in the curvilinear coordinate system is derived. This derivative reflects the trend of the vehicle's heading change and is an important basis for predicting future changes in the vehicle's direction.

[0059] Next, by combining the target vehicle's acceleration, current heading angle deviation, longitudinal velocity, first derivative of lateral displacement, derivative of heading angle deviation, and curvature and first derivative of the reference trajectory at the projection point, the longitudinal acceleration of the target vehicle in the curvilinear coordinate system is calculated. The longitudinal acceleration considers the effects of vehicle acceleration, deceleration, and path curvature, providing a core parameter for dynamic path planning. Simultaneously, based on the target vehicle's global coordinates and the projection point coordinates, the lateral displacement of the target vehicle in the curvilinear coordinate system is determined. The sign of the lateral displacement reveals the relative orientation between the vehicle and the reference trajectory, serving as an important reference for obstacle avoidance strategies in path planning.

[0060] Furthermore, based on the target vehicle's speed and heading angle deviation, the lateral velocity of the target vehicle in the curvilinear coordinate system can be determined. The lateral velocity reflects the vehicle's lateral movement trend and is crucial for predicting the dynamic behavior of obstacle vehicles and planning safe paths.

[0061] Using the target vehicle's longitudinal acceleration, longitudinal velocity, first derivative of lateral displacement, and second derivative of lateral displacement in the curvilinear coordinate system, the lateral acceleration of the target vehicle in this coordinate system is calculated. The calculation of lateral acceleration, combined with the vehicle's longitudinal dynamic changes and lateral stability control, ensures that the planned path both conforms to physical constraints and effectively avoids collisions.

[0062] Finally, by using the curvature, lateral displacement, and heading angle deviation of the reference trajectory at the projection point, the first derivative of the lateral displacement of the target vehicle in the curvilinear coordinate system is determined. Furthermore, by combining the first derivative of the curvature of the reference trajectory, the first derivative of the lateral displacement, the heading angle deviation, and their derivatives, the second derivative of the lateral displacement is calculated.

[0063] The above series of transformation and calculation steps not only include a comprehensive description of the target vehicle's motion state in the curvilinear coordinate system, but also fully consider the interaction between the vehicle and the reference trajectory, laying a solid foundation for autonomous vehicles to make optimal path planning decisions in complex high-speed scenarios.

[0064] In another aspect, converting the first motion state data into the first target motion state data in the curvilinear coordinate system can also be achieved by the following method.

[0065] A conventional Frenet coordinate conversion diagram is shown in Figure 2 , is the actual position of the vehicle in the Cartesian coordinate system, is the actual position of the vehicle is the projection point position on the reference trajectory, is the initial point position of the reference trajectory. and respectively represent the unit tangent vector of the trajectory point at the projection point and the vehicle coordinate position, and respectively represent the unit normal vector of the trajectory point at the projection point and the vehicle coordinate position, and may be represented by vectors and vectors respectively, representing the displacement vectors of the actual position and the projection position.

[0066] For Cartesian coordinate to Frenet coordinate conversion, in the Cartesian coordinate system, the actual position of the vehicle may be represented by coordinates , the projection point may be represented by coordinates , and the vehicle position in the Frenet coordinate system can be represented as . The difference between the actual position heading angle and the reference position heading angle is represented as , that is .

[0067] where x represents the coordinate value of the vehicle in the horizontal direction; y represents the coordinate value of the vehicle in the vertical direction; θ is the heading angle of the vehicle; v is the speed of the vehicle; a is the acceleration of the vehicle; is the curvature of the actual driving trajectory of the vehicle, describing the rate of change of the current direction of the vehicle, κ = 1 / R, where R is the curvature radius of the trajectory.

[0068] is the longitudinal displacement; is the longitudinal velocity in the Frenet coordinate system; is the longitudinal acceleration in the Frenet coordinate system; is the lateral displacement; is the lateral velocity in the Frenet coordinate system; is the lateral acceleration in the Frenet coordinate system; is the first derivative of the lateral displacement with respect to the longitudinal displacement; is the second derivative of the lateral displacement with respect to the longitudinal displacement.

[0069] According to the reference trajectory, the distance between the vehicle point is the vehicle reference point , the vehicle projection point is the point and the reference trajectory initial point The arc length between the point and the reference trajectory initial point is the longitudinal displacement in the Frenet coordinate system .

[0070] According to the geometric relationship, we can know that:

[0071] The transformation can be obtained:

[0072] The distance between the vehicle actual position and the reference point can be recorded as:

[0073]

[0074] In the Frenet coordinate system, the vector from each point to the vehicle projection point on the reference line is in the same direction or opposite direction with the normal vector of the vehicle projection point , that is and are in the same line, that is:

[0075]

[0076]

[0077] Based on the above, we can determine the size of , the positive and negative of to determine the positive and negative of , that is:

[0078]

[0079] Taking the derivative of both sides of , we can know that:

[0080]

[0081] wherein is the actual position vector of the vehicle, , is the position vector of the vehicle projection point, , that is: ​

[0082]

[0083] According to Frenet's formula, we know... , ,in Let be the curvature, then substituting it into the equation gives...

[0084]

[0085]

[0086] Equivalence Differentiating both sides yields .

[0087] Substitute Both sides ride at the same time It can be concluded that

[0088]

[0089] Simplifying it, we get:

[0090]

[0091] If is the first derivative of the lateral displacement with respect to the longitudinal coordinate, then it can be calculated as follows:

[0092]

[0093] If the second derivative of the lateral displacement with respect to the longitudinal coordinate is given, then the following can be calculated:

[0094]

[0095] in, .

[0096] because for and The included angle of the axes is the heading angle of vehicle P; for and The angle between the axes is the projection point. The heading angle, because:

[0097]

[0098] so:

[0099]

[0100] Substituting, we get:

[0101]

[0102] Because Therefore:

[0103]

[0104] Therefore:

[0105]

[0106] Lateral acceleration :

[0107]

[0108]

[0109] In summary, the formula for converting Cartesian coordinates to Frenet coordinates is as follows:

[0110]

[0111] For the conversion of Cartesian coordinates to Frenet coordinates, given the coordinates of the vehicle point in the Frenet coordinate system , the coordinates of the vehicle projection point point in the Cartesian coordinate system , the coordinates of the vehicle point in the Cartesian coordinate system are solved , that is, the conversion of in the Frenet coordinate system to in the Cartesian coordinate system.

[0112] As Figure 2 shown, it can be seen that:

[0113]

[0114] And , that is .

[0115] Heading angle in the geodetic coordinate system :

[0116]

[0117] Velocity in the geodetic coordinate system :

[0118]

[0119] Acceleration in the geodetic coordinate system :

[0120]

[0121] Curvature in the Cartesian coordinate system :

[0122]

[0123] In summary, the formula for converting Frenet coordinates to Cartesian coordinates is as follows:

[0124]

[0125] To reduce the calculation and accumulation of curve length for a long distance, thereby reducing the response time of the model, the ego vehicle is introduced into the coordinate origin of the Frenet coordinate system, and the introduction of the ego vehicle into the coordinate origin increases the position interaction between the ego vehicle and the obstacle vehicle.

[0126] As shown in Figure 3 , the ego vehicle Frenet coordinate conversion, where is the actual position of the ego vehicle in the Cartesian coordinate system, and the coordinate origin of the ego vehicle is is the projection point position of the ego vehicle on the reference trajectory in the Cartesian coordinate system, is the actual position of the obstacle vehicle in the Cartesian coordinate system, is the projection point position of the obstacle vehicle on the reference trajectory in the Cartesian coordinate system. The coordinates of the ego vehicle point in the Cartesian coordinate system can be expressed as , the coordinates of the obstacle vehicle point in the Cartesian coordinate system can be expressed as , the coordinates of the obstacle vehicle projection point point in the Cartesian coordinate system can be expressed as , the coordinates of the ego vehicle point in the ego vehicle Frenet coordinate system can be expressed as , the coordinates of the obstacle vehicle point in the ego vehicle Frenet coordinate system can be expressed as .

[0127] The coordinates of the ego vehicle in the Cartesian coordinate system are , which are converted according to the traditional Frenet coordinate system to , where is set to 0 to ensure that the longitudinal displacement of the ego vehicle is always 0, thereby making the ego vehicle in the On the axis, it can be known Therefore, the formula of converting the obstacle vehicle from Cartesian coordinate to ego vehicle Frenet coordinate is as follows:

[0128]

[0129] The coordinates of the obstacle vehicle in the Cartesian coordinate system are The coordinates of the obstacle vehicle in the Cartesian coordinate system are .

[0130] In the high-speed scenario, the road curvatures of the ego vehicle and the obstacle vehicle can be approximated as 0, and the longitudinal displacement between the two can be approximated as a straight line. As Figure 3 It can be known that

[0131]

[0132] The size of can be determined by , and the sign of can be determined by the sign of

[0133] , so

[0134]

[0135]

[0136] The Frenet lateral velocity of the obstacle vehicle in the ego vehicle coordinate system is

[0137]

[0138] The longitudinal velocity of the obstacle vehicle in the ego vehicle Frenet coordinate is the longitudinal velocity difference between the obstacle vehicle and the ego vehicle in the traditional Frenet coordinate, so it can be obtained that

[0139]

[0140] In summary, the formula of converting the obstacle vehicle from Cartesian coordinate to ego vehicle Frenet coordinate is as follows:

[0141]

[0142] According to some optional embodiments of the present application, the conversion of the second motion state data into the second target motion state data in the curvilinear coordinate system can be achieved by the following method: determining the longitudinal displacement of the obstacle vehicle in the curvilinear coordinate system according to the distance and direction between the projection point of the obstacle vehicle on the reference trajectory and the projection point of the target vehicle on the reference trajectory; determining the lateral displacement of the obstacle vehicle in the curvilinear coordinate system according to the global coordinates of the obstacle vehicle and the coordinates of the projection point of the obstacle vehicle on the reference trajectory, wherein the sign of the lateral displacement of the obstacle vehicle in the curvilinear coordinate system is used to represent the relative direction between the obstacle vehicle and the reference trajectory; determining the heading angle deviation of the obstacle vehicle in the curvilinear coordinate system according to the heading angle of the obstacle vehicle and the tangent angle at the projection point of the obstacle vehicle on the reference trajectory; determining the lateral velocity of the obstacle vehicle in the curvilinear coordinate system according to the speed of the obstacle vehicle and the heading angle deviation of the obstacle vehicle in the curvilinear coordinate system; and determining the relative longitudinal velocity of the obstacle vehicle in the curvilinear coordinate system relative to the target vehicle according to the speed of the obstacle vehicle, the heading angle deviation of the obstacle vehicle in the curvilinear coordinate system, the curvature of the reference trajectory at the projection point of the obstacle vehicle, the lateral displacement of the obstacle vehicle, and the longitudinal velocity of the target vehicle in the curvilinear coordinate system.

[0143] In the present embodiment, first, the longitudinal displacement of the obstacle vehicle in the curvilinear coordinate system is calculated by measuring the distance between the projection point of the obstacle vehicle on its reference trajectory and the projection point of the target vehicle, combined with the relative direction of the two. This calculation is based on the relative position of the two vehicles on the ideal driving path, and provides basic information in the longitudinal dimension for subsequent risk assessment.

[0144] Secondly, the lateral displacement of the obstacle vehicle in the curvilinear coordinate system is determined according to the global coordinates of the obstacle vehicle (in the Cartesian coordinate system) and the coordinates of the projection point of the obstacle vehicle on the reference trajectory. The sign of the lateral displacement clearly identifies the direction of the obstacle vehicle relative to the reference trajectory, which is a key indicator for determining whether the obstacle vehicle is deviating towards the driving route of the target vehicle, and is crucial for formulating obstacle avoidance strategies in path planning.

[0145] Then, the heading angle deviation of the obstacle vehicle in the curvilinear coordinate system is calculated by comparing the heading angle of the obstacle vehicle itself with the tangent angle of the reference trajectory at the projection point of the obstacle vehicle. The heading angle deviation is used to reflect the relationship between the driving direction of the obstacle vehicle and the ideal path, and is one of the important parameters for assessing dynamic risk.

[0146] Again, the lateral velocity of the obstacle vehicle in the curvilinear coordinate system is determined in combination with the speed of the obstacle vehicle and the calculated heading angle deviation. The quantification of the lateral velocity helps to understand the potential movement trend of the obstacle vehicle in the lateral direction, and is a necessary consideration when planning a safe driving path.

[0147] Finally, the relative longitudinal velocity of the obstacle vehicle in the curvilinear coordinate system relative to the target vehicle is calculated using the speed of the obstacle vehicle, the deviation of the heading angle of the obstacle vehicle in the curvilinear coordinate system, the curvature of the reference trajectory at the projection point of the obstacle vehicle, and the lateral displacement of the obstacle vehicle and the longitudinal velocity of the target vehicle in the curvilinear coordinate system. The determination of the relative longitudinal velocity comprehensively considers the dynamic changes of the obstacle vehicle and the target vehicle in the longitudinal direction and the influence of the lateral displacement of the obstacle vehicle, and provides detailed data support for constructing an accurate interaction model and effectively evaluating the collision risk.

[0148] The above steps can more accurately evaluate the collision threat brought by the obstacle vehicle by analyzing the position, direction and speed of the obstacle vehicle in the curvilinear coordinate system, and further plan a driving path that can avoid risks and maintain traffic efficiency.

[0149] On the other hand, the conversion of the second motion state data into the second target motion state data in the curvilinear coordinate system can be realized by the following method:

[0150] The coordinates of the obstacle vehicle in the Cartesian coordinate system are , which are converted into the Frenet coordinate system of the ego vehicle as .

[0151] In the high-speed scenario, the road curvatures of the ego vehicle and the obstacle vehicle can be approximated as 0, and the longitudinal displacement between the two can be approximated as a straight line. For example, Figure 3 , it can be known that

[0152]

[0153] The size of can be determined by , and the positive and negative signs can be determined by the positive and negative of , so

[0154] The lateral displacement of the obstacle vehicle :

[0155]

[0156] The Frenet lateral velocity of the obstacle vehicle in the ego vehicle coordinate system is

[0157]

[0158] The longitudinal velocity of the obstacle vehicle in the Frenet coordinate system of the ego vehicle is the longitudinal velocity difference between the obstacle vehicle and the ego vehicle in the traditional Frenet coordinate system, so it can be obtained that

[0159]

[0160] In summary, the formula for converting the obstacle vehicle in Cartesian coordinates into the ego vehicle Frenet coordinates is as follows:

[0161]

[0162] As another optional embodiment of the present application, according to the motion state data of the obstacle vehicle at the first time in the curvilinear coordinate system, the motion state data of the obstacle vehicle at the second time in the curvilinear coordinate system can be predicted by the following method: inputting the motion state data of the obstacle vehicle at the first time in the curvilinear coordinate system into the pre-trained long short-term memory network, controlling the memory degree of the current input information through the input gate in the long short-term memory network, controlling the retention degree of the historical state information through the forgetting gate, and controlling the output degree of the current time state information through the output gate; learning the motion law of the obstacle vehicle in the curvilinear coordinate system by using the long short-term memory network to extract the time sequence features of the obstacle vehicle; and outputting the predicted motion state data of the obstacle vehicle at the second time in the curvilinear coordinate system through the fully connected layer of the long short-term memory network based on the time sequence features.

[0163] In processing the motion state prediction of the obstacle vehicle, first, the motion state data of the obstacle vehicle at the first time in the curvilinear coordinate system, including but not limited to its position, velocity, acceleration and other information, is input into the pre-trained long short-term memory network. The long short-term memory network determines how much current input information will be stored in the cell state through its unique input gate mechanism. The forgetting gate controls the retention degree of the historical state information in the cell state. The output gate adjusts the output degree of the current time state information, and determines how much information is released from the cell state to the next level of the network, providing necessary internal state update for predicting the motion state at the next time.

[0164] Through the above mechanism, the long short-term memory network learns and masters the motion law of the obstacle vehicle in the curvilinear coordinate system, especially its time sequence features, i.e. how to predict the future state based on the past and current motion state. The above learning process relies on the iterative calculation of the multi-layer neural network, which gradually extracts and strengthens the internal correlation of the motion state change of the obstacle vehicle.

[0165] Based on the extracted time sequence features, the fully connected layer of the long short-term memory network is responsible for outputting the predicted motion state data of the obstacle vehicle at the second time in the curvilinear coordinate system. Through comprehensive analysis and comprehensive consideration of the internal state of the network, the fully connected layer generates the predicted values of the position, velocity, acceleration and other information of the obstacle vehicle at the next time, providing key future dynamic information of the obstacle vehicle for path planning.

[0166] On the other hand, according to the motion state data of the obstacle vehicle at the first time in the curvilinear coordinate system, the motion state data of the obstacle vehicle at the second time in the curvilinear coordinate system can also be predicted by the following method.

[0167] It can be understood that in the actual driving process, the uncertain risk mainly comes from the uncertain lane-changing behavior of the obstacle vehicle, and the ego vehicle should adopt the correct obstacle avoidance mode to avoid the driving risk problem caused by the uncertain behavior. In combination with the historical motion state information and the surrounding environment information of the vehicle, the trajectory relative displacement of the vehicle is predicted based on a long short-term memory network.

[0168] The input of the model is the predicted vehicle information and the predicted vehicle front vehicle information in the historical traffic scene for a period of time, and the output is the trajectory relative displacement of the predicted vehicle at the next time, and the input and the output are:

[0169]

[0170]

[0171]

[0172] Among them, is the historical time length of the input, represents the relative displacement in the x direction and the y direction at the current time relative to the last time.

[0173] Based on the prediction of the above vehicle trajectory relative displacement prediction model, the two relative displacements in the lateral and longitudinal directions at the next time and the current time can be obtained, and by calculating the position state at the current time, the predicted position, heading angle, speed, acceleration and curvature at the next time can be obtained, that is:

[0174]

[0175]

[0176]

[0177]

[0178]

[0179] Based on the above prediction model, by introducing the historical state of the obstacle vehicle, the trained vehicle trajectory relative displacement prediction model is loaded, and the position coordinates of the obstacle vehicle at the next time can be obtained , which is converted into the predicted coordinates of the obstacle vehicle in the ego vehicle Frenet coordinate system .

[0180] In some optional embodiments of the present application, determining the static risk field intensity and the dynamic risk field intensity of the obstacle vehicle according to the motion state data of the obstacle vehicle at the second time can be achieved by the following method: determining the static risk field intensity of the obstacle vehicle according to the mass, length and width of the obstacle vehicle, the relative distance between the target vehicle and the obstacle vehicle in the longitudinal direction, and the relative distance between the target vehicle and the obstacle vehicle in the lateral direction; and determining the dynamic risk field intensity of the obstacle vehicle according to the mass, length and width of the obstacle vehicle, the relative distance between the target vehicle and the obstacle vehicle in the longitudinal direction, the relative distance between the target vehicle and the obstacle vehicle in the lateral direction, the longitudinal speed of the obstacle vehicle, the lateral speed of the obstacle vehicle, the longitudinal displacement of the target vehicle in the curvilinear coordinate system, and the longitudinal displacement of the obstacle vehicle in the curvilinear coordinate system.

[0181] In the present embodiment, the calculation is first based on the mass, length, width of the obstacle vehicle, and the relative distance between the target vehicle and the obstacle vehicle in the longitudinal and lateral directions. This process links the physical dimensions of the obstacle vehicle with the size of its potential collision threat to the target vehicle, and converts these parameters into quantitative indicators of static risk field intensity through a Gaussian function or other suitable mathematical model. Among them, the evaluation of static risk field intensity focuses on the physical existence of static obstacles and their direct impact on the target vehicle's passage path, ensuring that the collision risk brought by stationary obstacles can be effectively avoided in path planning.

[0182] Then, the physical dimensions of the obstacle vehicle and the relative position between the target vehicle and the obstacle vehicle are considered, as well as the dynamic characteristics of the obstacle vehicle, including the longitudinal speed, lateral speed, and the longitudinal displacement of the target vehicle and the obstacle vehicle in the curvilinear coordinate system. By constructing a dynamic risk field model, the motion state and relative position information of the obstacle vehicle can be fused to quantify the immediate and potential threat posed by the moving obstacle vehicle to the target vehicle.

[0183] The dynamic risk field intensity aims to comprehensively evaluate the dynamic risks in the driving environment, including the speed changes, direction adjustments of the obstacle vehicle, and the relative motion trajectory with the target vehicle. The dynamic risk field intensity provides real-time safety evaluation for path planning, enabling the planning process to dynamically adapt to the behavior changes of the obstacle vehicle and timely adjust the driving strategy of the target vehicle to maintain a safe driving distance and avoid collisions. Through the comprehensive evaluation of static risk field intensity and dynamic risk field intensity, the risk of obstacles faced by the target vehicle in high-speed driving scenarios can be fully grasped, ensuring that path selection not only considers passage efficiency, but also focuses on driving safety.

[0184] On the other hand, determining the static risk field intensity and the dynamic risk field intensity of the obstacle vehicle according to the motion state data of the obstacle vehicle at the second time can also be achieved by the following method:

[0185] Based on the constructed ego vehicle Frenet coordinate system, a static obstacle vehicle risk field is described for a static vehicle parked on the roadside. Assuming that the position of the obstacle vehicle in the ego vehicle Frenet coordinate system is , the mass of the obstacle vehicle is , the length and width of the obstacle vehicle are and , and are high-order numbers:

[0186]

[0187] The dynamic obstacle vehicle risk field is related to the speed, heading angle and other change quantities in addition to the vehicle mass, length and width. Assuming that the position of the obstacle vehicle in the ego vehicle Frenet coordinate system is , the mass of the obstacle vehicle is , the longitudinal speed of the obstacle vehicle is , the length and width of the obstacle vehicle are and , and are high-order numbers:

[0188]

[0189] The dynamic obstacle vehicle risk distribution map is obtained by superimposing the static obstacle vehicle risk field and the dynamic risk field, that is: .

[0190] As some optional embodiments of the present application, the target path corresponding to the target vehicle can be determined according to the first target motion state data and the total collision risk by the following method: according to the first target motion state data, a plurality of path points are arranged at equal intervals in the longitudinal direction of the curve coordinate system according to a preset distance, and a plurality of path points are arranged in the transverse direction of the curve coordinate system based on the center line of the lane; the path points at different longitudinal positions are divided into a plurality of intervals, and a plurality of candidate paths are generated by connecting the path points of adjacent intervals; the offset cost corresponding to the candidate path is determined according to the transverse displacement of the target vehicle in the curve coordinate system; a planning objective function is constructed based on the total collision risk and the offset cost corresponding to the candidate path; and the target path that minimizes the planning objective function is determined from the plurality of candidate paths.

[0191] In the present embodiment, the equal-interval layout of the path points is performed in the longitudinal direction. Specifically, a plurality of path points are arranged along the extension direction of the reference trajectory at intervals of a preset distance, so as to ensure that the path planning covers a sufficient area in the forward direction of the vehicle. In the transverse direction, the path points are also arranged based on the center line of the lane, which aims to refine the transverse positioning of the vehicle in the lane.

[0192] The path points at different longitudinal positions are organized into multiple intervals. In each interval, the path points are connected to form a candidate path, which can achieve preliminary discretization of the path and facilitate subsequent optimization calculation. The connection process of the path needs to consider the physical limitations of the vehicle and the road conditions to ensure that the generated candidate path is feasible and reasonable.

[0193] According to the lateral displacement of the target vehicle in the curvilinear coordinate system, the offset cost of each candidate path is determined. The offset cost is used to reflect the degree of deviation of the vehicle from the center line of the lane, and by setting the offset function, the discomfort and safety reduction of the lateral movement of the vehicle can be quantified, and the constraint condition of the lateral movement is introduced for path planning.

[0194] Based on the above information, the planning objective function is constructed by integrating the total collision risk and the offset cost of the candidate path. The planning objective function is used to balance the safety and comfort of the vehicle, and by appropriately allocating the weights, it can ensure that the planning objective function can comprehensively reflect the pros and cons of the path.

[0195] Finally, among all the candidate paths, the target path that makes the planning objective function take the minimum value is selected. The optimization process aims to find the optimal driving path that can maximize the avoidance of the total collision risk and maintain a low offset cost, ensuring that the vehicle can complete the path planning in the most safe and comfortable way under the premise of meeting the physical constraints and road rules, and improving the driving efficiency and safety of the autonomous vehicle in complex traffic environment.

[0196] Specifically, based on the total collision risk and the offset cost corresponding to the candidate path, the planning objective function can be constructed by the following method: based on the total collision risk and the offset cost corresponding to the candidate path, a lateral planning objective function is constructed; according to the relative distance deviation of the candidate path and the reference trajectory in the longitudinal direction, the relative speed deviation of the longitudinal speed of the candidate path and the preset reference speed, and the longitudinal acceleration variation of the candidate path, a longitudinal planning objective function is constructed; the lateral planning objective function and the longitudinal planning objective function are combined by weighting to obtain the planning objective function.

[0197] In this embodiment, based on the total collision risk and the offset cost corresponding to the candidate path, a lateral planning objective function is constructed. The lateral planning objective function is used to quantify and optimize the movement strategy of the vehicle in the lateral direction, ensuring that it can effectively avoid potential collisions while maintaining the comfort and stability of the vehicle. The total collision risk, as the core indicator of safety, can reflect the possibility of the vehicle encountering a collision under a given path, and the offset cost considers the degree of deviation of the vehicle from the ideal driving lane center line, reflecting the comfort and difficulty of vehicle control during driving. The lateral planning objective function combines these two key factors through a suitable mathematical model, such as weighted summation or composite function, to form a comprehensive optimization target for lateral movement.

[0198] A longitudinal planning objective function is constructed according to the relative distance deviation of the candidate path from the reference trajectory in the longitudinal direction, the relative speed deviation of the longitudinal speed of the candidate path from the preset reference speed, and the longitudinal acceleration variation of the candidate path. The longitudinal planning objective function focuses on the speed control and acceleration variation of the vehicle in the forward direction, aiming to maintain the safe distance between the vehicle and the front vehicle, while keeping the speed as smooth as possible to avoid unnecessary acceleration or deceleration, so as to improve the driving efficiency and comfort. By calculating the deviation from the ideal driving state, the longitudinal planning objective function can evaluate the longitudinal performance of the candidate path, ensuring that the planned path is not only safe, but also efficient and smooth.

[0199] The lateral planning objective function and the longitudinal planning objective function are combined by weighting to form a comprehensive planning objective function. In the weighting combination process, the weight ratio between lateral risk and longitudinal comfort needs to be adjusted according to the specific application scenario and safety priority. By solving the optimal value of this planning objective function, an optimal path that meets the safety, comfort and efficiency requirements in both lateral and longitudinal directions can be found. The comprehensive planning objective function balances multiple dimensions of vehicle driving, ensuring that the driving decision of the vehicle in complex high-speed scenarios is scientific and comprehensive, greatly improving the driving safety and road traffic capacity of the autonomous vehicle.

[0200] On the other hand, according to the first target motion state data and the total collision risk, the target path corresponding to the target vehicle can also be determined by the following method:

[0201] As shown in Figure 4 , the DP dynamic programming algorithm theory can be introduced into the ego vehicle Frenet coordinate system to establish the DP dynamic programming under the ego vehicle Frenet, and by scattering points in the longitudinal direction and scattering points on the center line of each lane in the lateral direction in the ego vehicle Frenet coordinate system, the lateral scattering points and the lateral scattering points are respectively taken as each stage and each sub-structure system, and by linking the subsystem of the current stage with the subsystem of the next stage, a suitable path is planned.

[0202] To solve the safety and comfort problems of automatic driving lane changing, a suitable path planning trajectory is proposed to effectively improve the safety and comfort. A quintic polynomial is used to describe the lateral displacement of the trajectory, and a quartic polynomial is used to describe the longitudinal displacement of the trajectory in the longitudinal direction, to generate a candidate trajectory.

[0203]

[0204] The main function of the lateral planning is to avoid collision, change lane and consider the risk cost of the traffic environment, so the risk value is integrated into the objective function of the lateral planning. In the actual tracking, the offset cost is also considered in the lateral planning.

[0205]

[0206] wherein, is the lane width, and the DP scattering point coordinate is , and is the lane lateral coordinate on the left and right sides of the current DP scattering point coordinate.

[0207] Further, the offset function is defined to ensure the driving safety of the vehicle:

[0208]

[0209] wherein, is the lateral coordinate of the ego vehicle in the ego Frenet coordinate system.

[0210] The objective function of the lateral planning is defined as a multi-objective function including the risk cost and the offset cost:

[0211]

[0212] wherein, is the risk cost function weight, is the offset cost function weight.

[0213] The objective function of the longitudinal planning is defined as a target function considering the distance, speed and acceleration:

[0214]

[0215] wherein, is the relative distance weight of the trajectory and the DP planning trajectory, is the relative speed weight of the trajectory speed and the reference speed, is the acceleration weight.

[0216] The total planning function is: .

[0217] Finally, the most safe and comfortable target trajectory is planned through the multi-objective coupling weight function of the lateral planning and the longitudinal planning.

[0218] Figure 5 Fig. 1 is a structural diagram of a path planning device according to an embodiment of the present application, as shown in the figure, the device comprises: Figure 5

[0219] ​The acquisition module 51 is configured to acquire first motion state data of the target vehicle in the rectangular coordinate system and second motion state data of the obstacle vehicle of the target vehicle in the rectangular coordinate system.

[0220] The conversion module 52 is configured to convert the first motion state data into first target motion state data in a curvilinear coordinate system and convert the second motion state data into second target motion state data in the curvilinear coordinate system, wherein an origin of the curvilinear coordinate system is a projection point of the target vehicle on a reference trajectory, the reference trajectory is used to represent an ideal motion path of the target vehicle, and the projection point is a point on the reference trajectory closest to the target vehicle.

[0221] The prediction module 53 is configured to predict motion state data of the obstacle vehicle at a second time in the curvilinear coordinate system according to the motion state data of the obstacle vehicle at a first time in the curvilinear coordinate system, wherein the second time is a future time of the first time.

[0222] The first determination module 54 is configured to determine static risk field intensity and dynamic risk field intensity of the obstacle vehicle according to the motion state data of the obstacle vehicle at the second time, wherein the static risk field intensity is used to evaluate the intensity of a collision threat posed by the obstacle vehicle in a static state to the target vehicle, and the dynamic risk field intensity is used to evaluate the intensity of a collision threat posed by the obstacle vehicle in a dynamic state to the target vehicle.

[0223] The second determination module 55 is configured to determine a total collision risk according to the static risk field intensity and the dynamic risk field intensity, and determine a target path corresponding to the target vehicle according to the first target motion state data and the total collision risk.

[0224] Optionally, the first motion state data is converted into first target motion state data in the curvilinear coordinate system, specifically comprising the following steps: determining the longitudinal displacement of the target vehicle in the curvilinear coordinate system as 0; determining the longitudinal velocity of the target vehicle in the curvilinear coordinate system according to the speed of the target vehicle, the heading angle deviation, the curvature of the reference trajectory at the projection point, and the lateral displacement of the target vehicle relative to the reference trajectory; determining the derivative of the heading angle deviation of the target vehicle in the curvilinear coordinate system according to the curvature of the reference trajectory at the projection point, the lateral displacement of the target vehicle relative to the reference trajectory, and the heading angle deviation; determining the longitudinal acceleration of the target vehicle in the curvilinear coordinate system according to the acceleration of the target vehicle, the heading angle deviation, the longitudinal velocity, the first derivative of the lateral displacement, the derivative of the heading angle deviation, the curvature of the reference trajectory at the projection point, the first derivative of the curvature of the reference trajectory at the projection point, and the lateral displacement; determining the lateral displacement of the target vehicle in the curvilinear coordinate system according to the global coordinates of the target vehicle and the coordinates of the projection point, wherein the sign of the lateral displacement of the target vehicle in the curvilinear coordinate system is used to represent the relative direction between the target vehicle and the reference trajectory; determining the lateral velocity of the target vehicle in the curvilinear coordinate system according to the speed of the target vehicle and the heading angle deviation; determining the lateral acceleration of the target vehicle in the curvilinear coordinate system according to the longitudinal acceleration of the target vehicle in the curvilinear coordinate system, the longitudinal velocity, the first derivative of the lateral displacement, and the second derivative of the lateral displacement; determining the first derivative of the lateral displacement of the target vehicle in the curvilinear coordinate system according to the curvature of the reference trajectory at the projection point, the lateral displacement, and the heading angle deviation; and determining the second derivative of the lateral displacement of the target vehicle in the curvilinear coordinate system according to the curvature of the reference trajectory at the projection point, the first derivative of the curvature of the reference trajectory at the projection point, the lateral displacement, the first derivative of the lateral displacement, the heading angle deviation, and the derivative of the heading angle deviation.

[0225] Optionally, the second motion state data is converted into second target motion state data in a curvilinear coordinate system, specifically comprising the following steps: determining a longitudinal displacement of the obstacle vehicle in the curvilinear coordinate system according to a distance and a direction between a projection point of the obstacle vehicle on a reference trajectory and a projection point of the target vehicle on the reference trajectory; determining a lateral displacement of the obstacle vehicle in the curvilinear coordinate system according to a global coordinate of the obstacle vehicle and a coordinate of the projection point of the obstacle vehicle on the reference trajectory, wherein a sign of the lateral displacement of the obstacle vehicle in the curvilinear coordinate system is used to represent a relative direction between the obstacle vehicle and the reference trajectory; determining a heading angle deviation of the obstacle vehicle in the curvilinear coordinate system according to a heading angle of the obstacle vehicle and a tangent angle at the projection point of the obstacle vehicle on the reference trajectory; determining a lateral velocity of the obstacle vehicle in the curvilinear coordinate system according to a velocity of the obstacle vehicle and the heading angle deviation of the obstacle vehicle in the curvilinear coordinate system; and determining a relative longitudinal velocity of the obstacle vehicle in the curvilinear coordinate system relative to the target vehicle according to the velocity of the obstacle vehicle, the heading angle deviation of the obstacle vehicle in the curvilinear coordinate system, a curvature of the reference trajectory at the projection point of the obstacle vehicle, the lateral displacement of the obstacle vehicle, and a longitudinal velocity of the target vehicle in the curvilinear coordinate system.

[0226] Optionally, the static risk field intensity and the dynamic risk field intensity of the obstacle vehicle are determined according to the motion state data of the obstacle vehicle at the second time, specifically comprising the following steps: determining the static risk field intensity of the obstacle vehicle according to a mass, a length and a width of the obstacle vehicle, a relative distance between the target vehicle and the obstacle vehicle in a longitudinal direction, and a relative distance between the target vehicle and the obstacle vehicle in a lateral direction; and determining the dynamic risk field intensity of the obstacle vehicle according to the mass, the length and the width of the obstacle vehicle, the relative distance between the target vehicle and the obstacle vehicle in the longitudinal direction, the relative distance between the target vehicle and the obstacle vehicle in the lateral direction, a longitudinal velocity of the obstacle vehicle, a lateral velocity of the obstacle vehicle, a longitudinal displacement of the target vehicle in the curvilinear coordinate system, and a longitudinal displacement of the obstacle vehicle in the curvilinear coordinate system.

[0227] Optionally, the target path corresponding to the target vehicle is determined according to the first target motion state data and the total collision risk, specifically comprising the following steps: setting a plurality of path points in a longitudinal direction of the curvilinear coordinate system at a preset distance according to the first target motion state data, and setting a plurality of path points in a lateral direction of the curvilinear coordinate system based on a center line of a lane; dividing the path points at different longitudinal positions into a plurality of intervals, and generating a plurality of candidate paths by connecting the path points of adjacent intervals; determining an offset cost corresponding to the candidate paths according to a lateral displacement of the target vehicle in the curvilinear coordinate system; constructing a planning objective function based on the total collision risk and the offset cost corresponding to the candidate paths; and determining the target path that minimizes the planning objective function from the plurality of candidate paths.

[0228] Optionally, based on the total collision risk and the deviation cost corresponding to the candidate path, a planning objective function is constructed, specifically including the following steps: based on the total collision risk and the deviation cost corresponding to the candidate path, a lateral planning objective function is constructed; according to the relative distance deviation of the candidate path and the reference trajectory in the longitudinal direction, the relative speed deviation of the longitudinal speed of the candidate path and the preset reference speed, and the longitudinal acceleration variation of the candidate path, a longitudinal planning objective function is constructed; the lateral planning objective function and the longitudinal planning objective function are weighted and combined to obtain the planning objective function.

[0229] Optionally, the motion state data of the obstacle vehicle at the second time in the polar coordinate system is predicted according to the motion state data of the obstacle vehicle at the first time in the polar coordinate system, specifically including the following steps: the motion state data of the obstacle vehicle at the first time in the polar coordinate system is input into the pre-trained long short-term memory network, the memory degree of the current input information is controlled through the input gate in the long short-term memory network, the retention degree of the historical state information is controlled through the forgetting gate, and the output degree of the state information at the current time is controlled through the output gate; the motion law of the obstacle vehicle in the polar coordinate system is learned by using the long short-term memory network to extract the time sequence features of the obstacle vehicle; and the predicted motion state data of the obstacle vehicle at the second time in the polar coordinate system is output through the full connection layer of the long short-term memory network based on the time sequence features.

[0230] It should be noted that the above Figure 5 Each module in the above

[0231] It should be noted that Figure 5 The preferred embodiments of the embodiments shown in Figure 1 will not be described here.

[0232] Figure 6 A hardware structure block diagram of a computer terminal for implementing the path planning method is shown. As Figure 6As shown, the computer terminal 60 may include one or more processors 602 (shown as 602a, 602b, ..., 602n in the figure) 602 (processor 602 may include, but is not limited to, a microprocessor MCU or a programmable logic device FPGA, etc.), a memory 604 for storing data, and a transmission module 606 for communication functions. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of a BUS bus), a network interface, a power supply, and / or a camera. Those skilled in the art will understand that... Figure 6 The structure shown is for illustrative purposes only and does not limit the structure of the aforementioned electronic device. For example, the computer terminal 60 may also include... Figure 6 The more or fewer components shown, or having the same Figure 6 The different configurations shown.

[0233] It should be noted that the aforementioned one or more processors 602 and / or other data processing circuits are generally referred to herein as "data processing circuits". These data processing circuits may be embodied, in whole or in part, in software, hardware, firmware, or any other combination thereof. Furthermore, the data processing circuits may be a single, independent processing module, or may be integrated, in whole or in part, into any other element within the computer terminal 60. As described in the embodiments of this application, the data processing circuits serve as processor control (e.g., selection of a variable resistor termination path connected to an interface).

[0234] The memory 604 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the path planning method in the embodiments of this application. The processor 602 executes various functional applications and data processing by running the software programs and modules stored in the memory 604, thereby realizing the above-mentioned path planning method. The memory 604 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 604 may further include memory remotely located relative to the processor 602, and these remote memories can be connected to the computer terminal 60 via a network. Examples of the above-mentioned networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0235] The transmission module 606 is configured to receive or send data via a network. The network can include a wireless network provided by a communication provider of the computer terminal 60. In one example, the transmission module 606 includes a network interface controller (NIC) that can be connected to other network devices through a base station to communicate with the Internet. In one example, the transmission module 606 can be a radio frequency (RF) module configured to communicate with the Internet wirelessly.

[0236] The display can be a liquid crystal display (LCD) that is touch screen, for example, which can enable a user to interact with a user interface of the computer terminal 60.

[0237] It should be noted that in some optional embodiments, the above Figure 6 The computer terminal shown can include hardware elements (including circuitry), software elements (including computer code stored on a computer readable medium), or a combination of both hardware and software elements. It should be noted that in some embodiments, the functions of the computer terminal described above can be provided by one or more of the hardware elements and / or software elements. Figure 6 is merely one example of a particular implementation and is intended to illustrate the types of components that can be present in the computer terminal described above.

[0238] It should be noted that Figure 6 The computer terminal shown is configured to perform Figure 1 The path planning method shown, and thus the related explanations in the method of executing the above commands also apply to the electronic device, which will not be described again here.

[0239] The embodiments of the present application also provide a non-volatile storage medium, which includes a stored program, wherein the program controls a device where the storage medium is located to perform the path planning method described above when the program is executed.

[0240] The non-volatile storage medium executes a program to: acquire first motion state data of a target vehicle in a rectangular coordinate system and second motion state data of an obstacle vehicle of the target vehicle in the rectangular coordinate system; convert the first motion state data into first target motion state data in a curvilinear coordinate system, and convert the second motion state data into second target motion state data in the curvilinear coordinate system, wherein an origin of the curvilinear coordinate system is a projection point of the target vehicle on a reference trajectory, the reference trajectory is used to represent an ideal motion path of the target vehicle, and the projection point is a closest point on the reference trajectory to the target vehicle; predict, according to motion state data of the obstacle vehicle at a first time in the curvilinear coordinate system, motion state data of the obstacle vehicle at a second time in the curvilinear coordinate system, wherein the second time is a future time of the first time; determine static risk field intensity and dynamic risk field intensity of the obstacle vehicle according to the motion state data of the obstacle vehicle at the second time, wherein the static risk field intensity is used to evaluate intensity of a collision threat posed by the obstacle vehicle in a static state to the target vehicle, and the dynamic risk field intensity is used to evaluate intensity of a collision threat posed by the obstacle vehicle in a dynamic state to the target vehicle; determine a total collision risk according to the static risk field intensity and the dynamic risk field intensity, and determine a target path corresponding to the target vehicle according to the first target motion state data and the total collision risk.

[0241] The embodiments of the present application also provide an electronic device, comprising a memory and a processor, the processor is used to run a program stored in the memory, wherein the program performs the path planning method when running.

[0242] The processor is configured to run a program to perform the following functions: obtaining first motion state data of a target vehicle in a rectangular coordinate system and second motion state data of an obstacle vehicle of the target vehicle in the rectangular coordinate system; converting the first motion state data into first target motion state data in a curvilinear coordinate system, and converting the second motion state data into second target motion state data in the curvilinear coordinate system, wherein an origin of the curvilinear coordinate system is a projection point of the target vehicle on a reference trajectory, the reference trajectory is used to represent an ideal motion path of the target vehicle, and the projection point is a point on the reference trajectory closest to the target vehicle; predicting, according to the motion state data of the obstacle vehicle at a first time in the curvilinear coordinate system, motion state data of the obstacle vehicle at a second time in the curvilinear coordinate system, wherein the second time is a future time of the first time; determining a static risk field intensity and a dynamic risk field intensity of the obstacle vehicle according to the motion state data of the obstacle vehicle at the second time, wherein the static risk field intensity is used to evaluate the intensity of a collision threat posed by the obstacle vehicle in a static state to the target vehicle, and the dynamic risk field intensity is used to evaluate the intensity of a collision threat posed by the obstacle vehicle in a dynamic state to the target vehicle; determining a total collision risk according to the static risk field intensity and the dynamic risk field intensity, and determining a target path corresponding to the target vehicle according to the first target motion state data and the total collision risk.

[0243] The above sequence numbers of the embodiments of the present application are only for description, and do not represent the advantages and disadvantages of the embodiments.

[0244] In the above embodiments of the present application, the description of each embodiment has its own emphasis, and the parts not described in detail in a certain embodiment can be referred to the related description of other embodiments.

[0245] In the above embodiments of the present application, the collected information is information and data authorized by the user or authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of related data comply with relevant laws, regulations and standards, necessary protection measures are taken, do not violate public order and good customs, and provide corresponding operation portal for user to choose authorization or refusal.

[0246] In the several embodiments provided by the present application, it should be understood that the disclosed technology can be implemented in other ways. Of course, the unit described as the division is only a logic function division, and there can be other division ways during actual implementation, for example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some interface, unit or module, and can be electrical or other forms.

[0247] The units described as separate components may or may not be physically separate, and the components displayed as units may or may not be physical units, that is, may be located in one place, or may be distributed to multiple units. Part or all of the units can be selected according to actual needs to achieve the purpose of the embodiment scheme.

[0248] In addition, the functional units in each embodiment of the present application can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit.

[0249] The integrated unit, if realized in the form of a software functional unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application or the part that contributes to the related art or the whole or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The foregoing storage medium includes: a U disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a mobile hard disk, a magnetic disk or an optical disk, and various program code storage media.

[0250] The above is only the preferred embodiment of the present application, and it should be pointed out that for those skilled in the art, without departing from the principles of the present application, a number of improvements and refinements can be made, and these improvements and refinements should be considered as the protection scope of the present application.

Claims

1. A path planning method, characterized in that, include: Acquire the first motion state data of the target vehicle in a Cartesian coordinate system and the second motion state data of the obstacle vehicle of the target vehicle in a Cartesian coordinate system; The first motion state data is converted into first target motion state data in a curvilinear coordinate system, and the second motion state data is converted into second target motion state data in the curvilinear coordinate system. The origin of the curvilinear coordinate system is the projection point of the target vehicle on the reference trajectory. The reference trajectory is used to characterize the ideal motion path of the target vehicle, and the projection point is the point on the reference trajectory that is closest to the target vehicle. Based on the motion state data of the obstacle vehicle at a first moment in the curved coordinate system, predict the motion state data of the obstacle vehicle at a second moment in the curved coordinate system, wherein the second moment is a future moment of the first moment; Based on the motion state data of the obstacle vehicle at the second moment, the static risk field intensity and dynamic risk field intensity of the obstacle vehicle are determined. The static risk field intensity is used to assess the intensity of the collision threat posed by the obstacle vehicle to the target vehicle when the obstacle vehicle is stationary, and the dynamic risk field intensity is used to assess the intensity of the collision threat posed by the obstacle vehicle to the target vehicle when the obstacle vehicle is in motion. Based on the static risk field intensity and the dynamic risk field intensity, the total collision risk is determined, and based on the first target motion state data and the total collision risk, the target path corresponding to the target vehicle is determined.

2. The method according to claim 1, characterized in that, Converting the first motion state data into first target motion state data in a curvilinear coordinate system includes: The longitudinal displacement of the target vehicle in the curvilinear coordinate system is set to 0; Based on the target vehicle's speed, heading angle deviation, the curvature of the reference trajectory at the projection point, and the target vehicle's lateral displacement relative to the reference trajectory, the longitudinal velocity of the target vehicle in the curvilinear coordinate system is determined. Based on the curvature of the reference trajectory at the projection point, the lateral displacement of the target vehicle relative to the reference trajectory, and the heading angle deviation, determine the derivative of the heading angle deviation of the target vehicle in the curvilinear coordinate system. The longitudinal acceleration of the target vehicle in the curvilinear coordinate system is determined based on the target vehicle's acceleration, heading angle deviation, longitudinal velocity, first derivative of lateral displacement, derivative of heading angle deviation, curvature of the reference trajectory at the projection point, first derivative of curvature of the reference trajectory at the projection point, and lateral displacement. Based on the global coordinates of the target vehicle and the coordinates of the projection point, the lateral displacement of the target vehicle in the curvilinear coordinate system is determined, wherein the sign of the lateral displacement of the target vehicle in the curvilinear coordinate system is used to indicate the relative direction between the target vehicle and the reference trajectory; Based on the speed of the target vehicle and the heading angle deviation, determine the lateral velocity of the target vehicle in the curvilinear coordinate system; The lateral acceleration of the target vehicle in the curvilinear coordinate system is determined based on the longitudinal acceleration, longitudinal velocity, first derivative of lateral displacement, and second derivative of lateral displacement of the target vehicle in the curvilinear coordinate system. Based on the curvature of the reference trajectory at the projection point, the lateral displacement, and the heading angle deviation, determine the first derivative of the lateral displacement of the target vehicle in the curvilinear coordinate system. The second derivative of the lateral displacement of the target vehicle in the curvature coordinate system is determined based on the curvature of the reference trajectory at the projection point, the first derivative of the curvature of the reference trajectory at the projection point, the lateral displacement, the first derivative of the lateral displacement, the heading angle deviation, and the derivative of the heading angle deviation.

3. The method according to claim 1, characterized in that, Converting the second motion state data into the second target motion state data in the curvilinear coordinate system includes: Based on the distance and direction between the projection point of the obstacle vehicle on the reference trajectory and the projection point of the target vehicle on the reference trajectory, the longitudinal displacement of the obstacle vehicle in the curvilinear coordinate system is determined; Based on the global coordinates of the obstacle vehicle and the coordinates of the projection point of the obstacle vehicle on the reference trajectory, the lateral displacement of the obstacle vehicle in the curvilinear coordinate system is determined, wherein the sign of the lateral displacement of the obstacle vehicle in the curvilinear coordinate system is used to indicate the relative direction between the obstacle vehicle and the reference trajectory. The heading angle deviation of the obstacle vehicle in the curvilinear coordinate system is determined based on the heading angle of the obstacle vehicle and the tangential angle of the obstacle vehicle at the projection point of the obstacle vehicle on the reference trajectory. The lateral velocity of the obstacle vehicle in the curvilinear coordinate system is determined based on the speed of the obstacle vehicle and the heading angle deviation of the obstacle vehicle in the curvilinear coordinate system. The relative longitudinal velocity of the obstacle vehicle relative to the target vehicle in the curvilinear coordinate system is determined based on the speed of the obstacle vehicle, the heading angle deviation of the obstacle vehicle in the curvilinear coordinate system, the curvature of the reference trajectory at the projection point of the obstacle vehicle, the lateral displacement of the obstacle vehicle, and the longitudinal velocity of the target vehicle in the curvilinear coordinate system.

4. The method according to claim 1, characterized in that, Based on the motion state data of the obstacle vehicle at the second moment, the static risk field intensity and dynamic risk field intensity of the obstacle vehicle are determined, including: The static risk field intensity of the obstacle vehicle is determined based on the mass, length, and width of the obstacle vehicle, the relative distance between the target vehicle and the obstacle vehicle in the longitudinal direction, and the relative distance between the target vehicle and the obstacle vehicle in the lateral direction. The dynamic risk field intensity of the obstacle vehicle is determined based on the mass, length, and width of the obstacle vehicle, the relative distance between the target vehicle and the obstacle vehicle in the longitudinal direction, the relative distance between the target vehicle and the obstacle vehicle in the lateral direction, the longitudinal velocity of the obstacle vehicle, the lateral velocity of the obstacle vehicle, the longitudinal displacement of the target vehicle in the curvilinear coordinate system, and the longitudinal displacement of the obstacle vehicle in the curvilinear coordinate system.

5. The method according to claim 1, characterized in that, Based on the first target motion state data and the total collision risk, the target path corresponding to the target vehicle is determined, including: Based on the first target motion state data, multiple path points are set at equal intervals along the longitudinal direction in the curved coordinate system at a preset distance, and multiple path points are set along the transverse direction in the curved coordinate system based on the centerline of the lane. The path points at different vertical positions are divided into multiple intervals, and multiple candidate paths are generated by connecting the path points of adjacent intervals. The offset cost corresponding to the candidate path is determined based on the lateral displacement of the target vehicle in the curvilinear coordinate system. Based on the total collision risk and the offset cost corresponding to the candidate paths, a planning objective function is constructed; Among the candidate paths, the objective path that minimizes the planning objective function is determined.

6. The method according to claim 5, characterized in that, Based on the total collision risk and offset cost corresponding to the candidate paths, a planning objective function is constructed, including: Based on the total collision risk and offset cost corresponding to the candidate paths, a lateral planning objective function is constructed. Based on the relative distance deviation between the candidate path and the reference trajectory in the longitudinal direction, the relative velocity deviation between the longitudinal velocity of the candidate path and the preset reference velocity, and the change in longitudinal acceleration of the candidate path, a longitudinal planning objective function is constructed. The planning objective function is obtained by weighting and combining the horizontal planning objective function and the vertical planning objective function.

7. The method according to claim 1, characterized in that, Based on the motion state data of the obstacle vehicle at a first moment in the curved coordinate system, predict the motion state data of the obstacle vehicle at a second moment in the curved coordinate system, including: The motion state data of the obstacle vehicle in the first moment in the curved coordinate system is input into a pre-trained long short-term memory network. The input gate in the long short-term memory network controls the degree of memory of the current input information, the forget gate controls the degree of retention of historical state information, and the output gate controls the degree of output of the current state information. The motion patterns of the obstacle vehicle in the curved coordinate system are learned using the long short-term memory network to extract the temporal features of the obstacle vehicle; Based on the aforementioned temporal characteristics, the predicted motion state data of the obstacle vehicle in the curvilinear coordinate system at the second moment is output through the fully connected layer of the long short-term memory network.

8. A path planning device, characterized in that, include: The acquisition module is used to acquire the first motion state data of the target vehicle in a Cartesian coordinate system and the second motion state data of the obstacle vehicle of the target vehicle in a Cartesian coordinate system. The conversion module is used to convert the first motion state data into first target motion state data in a curve coordinate system, and to convert the second motion state data into second target motion state data in the curve coordinate system, wherein the origin of the curve coordinate system is the projection point of the target vehicle on the reference trajectory, the reference trajectory is used to characterize the ideal motion path of the target vehicle, and the projection point is the point on the reference trajectory closest to the target vehicle. The prediction module is used to predict the motion state data of the obstacle vehicle at a second moment in the curved coordinate system based on the motion state data of the obstacle vehicle at a first moment in the curved coordinate system, wherein the second moment is a future moment of the first moment; The first determining module is used to determine the static risk field intensity and dynamic risk field intensity of the obstacle vehicle based on the motion state data of the obstacle vehicle at the second moment. The static risk field intensity is used to assess the intensity of the collision threat posed by the obstacle vehicle to the target vehicle when the obstacle vehicle is stationary, and the dynamic risk field intensity is used to assess the intensity of the collision threat posed by the obstacle vehicle to the target vehicle when the obstacle vehicle is in motion. The second determining module is used to determine the total collision risk based on the static risk field intensity and the dynamic risk field intensity, and to determine the target path corresponding to the target vehicle based on the first target motion state data and the total collision risk.

9. A non-volatile storage medium, characterized in that, The non-volatile storage medium includes a stored program, wherein, when the program is executed, it controls the device containing the non-volatile storage medium to perform the path planning method according to any one of claims 1 to 7.

10. An electronic device, characterized in that, include: A memory and a processor, the processor being configured to run a program stored in the memory, wherein the program, when running, performs the path planning method according to any one of claims 1 to 7.

11. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the path planning method according to any one of claims 1 to 7.

Images