Driving track planning method, device and equipment and readable storage medium

Abstract

The invention relates to the technical field of vehicle path planning, and discloses a driving track planning method, device and equipment and a readable storage medium, and the driving track planning method comprises the steps: obtaining vehicle state information, target driving information and environment perception information, and determining a vehicle drivable area according to the environment perception information; performing path search based on the vehicle state information and the target driving information in the vehicle driving area to generate path reference information; performing continuous processing on the path reference information, and constructing reference trajectory information for vehicle motion control; and establishing a vehicle motion control model based on the reference trajectory information, and executing trajectory optimization to generate a target driving trajectory. According to the method, the planning efficiency and the track smoothness and stability are effectively improved, and the adaptive capacity and the driving safety of the automatic driving vehicle in a complex scene are enhanced.

Description

Technical Field

[0001] This application relates to the field of vehicle route planning technology, and in particular to a driving trajectory planning method, apparatus, device and readable storage medium. Background Technology

[0002] With the development of autonomous driving technology, path planning, as one of its core modules, needs to generate smooth driving trajectories that conform to vehicle dynamics while ensuring driving safety. Currently widely used traditional methods, such as graph search or sampling algorithms like A* and RRT, can quickly find a feasible path to avoid obstacles in static environments, but their outputs are mostly discrete polylines, which suffer from abrupt changes in direction and discontinuous curvature, making them difficult to use directly for vehicle control. In contrast, optimization-based methods, such as quadratic programming (QP), can finely adjust the path by introducing constraints, possessing a certain constraint handling capability, and are more widely used in structured scenarios.

[0003] Despite continuous technological advancements, numerous challenges remain. While path planning methods based on the Frenet coordinate system can decompose complex motion into longitudinal and lateral dimensions, reducing modeling difficulty and improving computational efficiency, their constraint representation and solution accuracy are still insufficient when dealing with dynamic obstacles or strongly nonlinear constraints. On the other hand, while nonlinear optimization methods can more accurately characterize vehicle dynamics, they are prone to optimization failure or getting trapped in local optima when introducing hard constraints such as collision avoidance, resulting in trajectory jitter or even infeasibility, affecting system stability and safety. Summary of the Invention

[0004] In view of this, embodiments of this application provide a driving trajectory planning method, apparatus, device, and readable storage medium, which can effectively solve the problems of poor path smoothness, weak dynamic adaptation capability, and insufficient stability of planning results in the prior art.

[0005] In a first aspect, embodiments of this application provide a driving trajectory planning method, including: Acquire vehicle status information, target driving information, and environmental perception information, and determine the vehicle's drivable area based on the environmental perception information; Within the vehicle's drivable area, a path search is performed based on the vehicle's status information and the target driving information to generate path reference information; The path reference information is processed to be continuous to construct reference trajectory information for vehicle motion control; A vehicle motion control model is established based on the reference trajectory information, and trajectory optimization is performed to generate the target driving trajectory.

[0006] In some embodiments, after generating the target driving trajectory, the method further includes: Collision detection is performed on the target vehicle trajectory; If the detection results do not meet the preset safe distance requirements, the path reference information will be regenerated. The reference trajectory information is updated based on the regenerated path reference information, and trajectory optimization is performed on the updated reference trajectory information to generate the target driving trajectory.

[0007] In some embodiments, acquiring vehicle status information, target driving information, and environmental perception information, and determining the drivable area of ​​the vehicle based on the environmental perception information, includes: The vehicle's current location, heading angle, and speed are obtained to form the vehicle's state information; Obtain the target location coordinates in the preset navigation path to construct the target driving information; Acquire laser point cloud or obstacle projection data containing obstacle distribution, and perform rasterization processing on the laser point cloud or obstacle projection data to generate an environmental raster map to constitute the environmental perception information. The connected regions in the environmental grid map that are not occupied by obstacles are defined as the drivable areas for the vehicle.

[0008] In some embodiments, the step of performing path search based on the vehicle status information and the target driving information within the vehicle's drivable area to generate path reference information includes: The current location information in the vehicle status information is used as the starting point for path search; The target location coordinates in the target driving information are determined as the endpoint of the path search; Within the drivable area of ​​the vehicle, a heuristic search algorithm is used to explore the path, with the starting point and the ending point as boundary conditions, to generate a discrete node sequence, which is then used as path reference information.

[0009] In some embodiments, the step of performing continuous processing on the path reference information to construct reference trajectory information for vehicle motion control includes: Resampling is performed on the discrete node sequence in the path reference information to generate a set of resampled points; Based on the resampled point set, a cubic spline curve is constructed to obtain the trajectory function continuously expressed under the path length parameter; The position, direction, and curvature information of the trajectory function are extracted to form a reference trajectory representation; The reference trajectory representation is converted to the coordinate system used by the vehicle control system to form the reference trajectory information.

[0010] In some embodiments, establishing a vehicle motion control model based on the reference trajectory information and performing trajectory optimization to generate a target driving trajectory includes: Based on the reference trajectory information, determine the vehicle's motion state sequence in the discrete time domain; Determine the set of control variables used for vehicle control; Based on the motion state sequence and the set of control variables, a vehicle motion control model and a trajectory optimization objective function are established. Within the discrete time domain, an iterative optimization algorithm is used to solve the trajectory optimization objective function; The obtained state trajectory is used as the target driving trajectory.

[0011] In some embodiments, the process of solving the problem using an iterative optimization algorithm includes: Calculate the difference between the state trajectory generated in this iteration and the state trajectory generated in the previous iteration; If the difference is less than a preset convergence threshold, the iteration stops and the current state trajectory is taken as the target driving trajectory. Otherwise, continue iterative optimization until the maximum number of iterations is reached, and use the state trajectory generated in the last iteration as the target driving trajectory.

[0012] Secondly, embodiments of this application provide a vehicle trajectory planning device, comprising: The area determination module is used to acquire vehicle status information, target driving information and environmental perception information, and determine the driving area of ​​the vehicle based on the environmental perception information. The path search module is used to perform path search based on the vehicle status information and the target driving information within the vehicle's drivable area, so as to generate path reference information; The path processing module is used to perform continuous processing on the path reference information to construct reference trajectory information for vehicle motion control; The trajectory generation module is used to establish a vehicle motion control model based on the reference trajectory information and perform trajectory optimization to generate a target driving trajectory.

[0013] Thirdly, embodiments of this application provide a terminal device, the terminal device including a processor and a memory, the memory storing a computer program, and the processor executing the computer program to implement the driving trajectory planning method of the first aspect described above.

[0014] Fourthly, embodiments of this application provide a computer-readable storage medium, wherein when the computer program is executed on a processor, it implements the driving trajectory planning method of the first aspect described above.

[0015] The embodiments of this application have the following beneficial effects: acquiring vehicle status information, target driving information, and environmental perception information, and determining the drivable area of ​​the vehicle based on the environmental perception information; within the drivable area, performing path search based on the vehicle status information and target driving information to generate path reference information; performing continuous processing on the path reference information to construct reference trajectory information for vehicle motion control; establishing a vehicle motion control model based on the reference trajectory information, and performing trajectory optimization to generate the target driving trajectory.

[0016] The embodiments of this application rapidly obtain topologically feasible initial paths through path search, and generate geometrically smooth reference trajectories by combining continuous processing, providing high-quality initial values ​​for dynamic optimization. Trajectory optimization is then performed on this basis, effectively integrating vehicle motion characteristics and control constraints, thus improving the safety and executability of the planning results. This achieves a progressive planning approach from obstacle avoidance to smoothing and then to fine-tuning, significantly improving trajectory generation efficiency and stability, and enhancing the adaptability and driving comfort of the autonomous driving system in complex environments. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A flowchart of a driving trajectory planning method according to an embodiment of this application is shown; Figure 2 Another flowchart of the driving trajectory planning method in the embodiments of this application is shown; Figure 3 This illustrates yet another flowchart of the vehicle trajectory planning method according to an embodiment of this application; Figure 4 A schematic diagram of a vehicle trajectory planning method according to an embodiment of this application is shown. Detailed Implementation

[0019] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0020] The components of the embodiments of this application described and illustrated in the accompanying drawings can be arranged and designed in a variety of different configurations. Therefore, the following detailed description of the embodiments of this application provided in the drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0021] In the following text, the terms "comprising," "having," and their cognates, which may be used in various embodiments of this application, are intended only to indicate a particular feature, number, step, operation, element, component, or combination thereof, and should not be construed as primarily excluding the presence of one or more other features, numbers, steps, operations, elements, components, or combinations thereof, or adding the possibility of one or more combinations thereof. Furthermore, the terms "first," "second," "third," etc., are used only for distinguishing descriptions and should not be construed as indicating or implying relative importance.

[0022] Unless otherwise specified, all terms used herein (including technical and scientific terms) shall have the same meaning as commonly understood by one of ordinary skill in the art to which the various embodiments of this application pertain. Terms (such as those defined in commonly used dictionaries) shall be interpreted as having the same meaning as in their contextual meaning in the relevant technical field and shall not be construed as having an idealized or overly formal meaning, unless clearly defined in the various embodiments of this application.

[0023] The following detailed description of some embodiments of this application is provided in conjunction with the accompanying drawings. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0024] Considering the problems of poor path smoothness, weak dynamic adaptability, and insufficient stability of planning results in existing technologies, a vehicle trajectory planning method is proposed. This method determines the drivable area of ​​the vehicle by combining environmental perception information and performs path search within this area to generate a preliminary path reference. Then, the path reference information is processed to make it continuous, and a reference trajectory that meets basic motion continuity is constructed. Finally, a vehicle motion control model is established based on the reference trajectory and trajectory optimization is performed, realizing a progressively refined transformation from geometric path to executable driving trajectory, thereby improving the safety, smoothness, and system adaptability of the planning results.

[0025] The driving trajectory planning method will be explained below with reference to some specific examples.

[0026] Figure 1A flowchart illustrating a driving trajectory planning method according to an embodiment of this application is shown. Exemplarily, the driving trajectory planning method includes the following steps: Step S100: Obtain vehicle status information, target driving information and environmental perception information, and determine the drivable area of ​​the vehicle based on the environmental perception information.

[0027] Among them, vehicle state information is used to characterize the spatial pose and motion state of the autonomous vehicle at the current moment, target driving information indicates the endpoint of the planning task, and environmental perception information reflects the distribution of surrounding obstacles. By integrating the above multi-source inputs, a safety cognition basis for the local traffic environment is established, and a legal path search space is delineated accordingly to provide geographical constraint boundaries for subsequent trajectory generation, ensuring that the planning results have obstacle avoidance feasibility.

[0028] In an optional embodiment, step S100 includes the following sub-steps: S101, acquire the vehicle's current position information, heading angle and driving speed to form vehicle status information.

[0029] The current location information refers to the vehicle's Cartesian coordinates in the global or local map coordinate system. heading angle This indicates the angle of the vehicle's longitudinal axis relative to true north, and its speed. This represents the current instantaneous longitudinal velocity; these parameters collectively describe the vehicle's initial motion conditions. Exemplarily, vehicle status information is output in real time by a high-precision positioning module (such as a GNSS / IMU integrated navigation system) and transmitted to the path planning unit via the vehicle's CAN bus; for example, the data received at a certain moment might be: The initial state serves as the starting pose for path searching and directly affects the continuity and smoothness of the subsequent path.

[0030] S102, obtain the target location coordinates in the preset navigation path to form target driving information.

[0031] Among them, the target location coordinates The destination's geographical coordinates are calculated by the navigation system based on a high-precision map and the user's destination, representing the endpoint of this route planning task. For example, when a vehicle needs to change lanes from its current lane and exit the ramp, the target point issued by the navigation system is located on the center line of the exit ramp, with a specific value... The target point is used as the termination condition for the path search algorithm, guiding heuristic search methods such as A* in the correct direction and avoiding ineffective exploration.

[0032] S103: Acquire laser point cloud or obstacle projection data containing obstacle distribution, and perform rasterization processing on the laser point cloud or obstacle projection data to generate an environmental raster map to constitute environmental perception information.

[0033] The laser point cloud is a set of three-dimensional spatial sampling points collected by the vehicle-mounted LiDAR sensor. The obstacle projection data is a dynamic obstacle representation formed by projecting the fused radar and visual detection results onto a two-dimensional horizontal plane. Rasterization processing refers to dividing the continuous space into fixed-size grid units (e.g., 0.2m × 0.2m), with each unit marked as "occupied" or "free". Exemplarily, the raw point cloud stream from multiple sensors is received, and after denoising, clustering, and classification processing, obstacle entities such as pedestrians and vehicles are identified. Their occupied areas are mapped onto a local raster map to generate a binary map, where "1" represents a passable area and "0" represents an obstacle area. This environmental raster map constitutes the core expression of environmental perception information and is used for subsequent determination of drivable areas.

[0034] S104, define the connected areas in the environmental grid map that are not occupied by obstacles as the drivable areas for vehicles.

[0035] The connected region refers to the set of all grid cells that can be reached from the vehicle's current location via an unobstructed path, excluding isolated empty areas or spaces blocked by obstacles. Exemplarily, using the grid cell where the vehicle is currently located as the seed point, a breadth-first search (BFS) is performed, traversing all adjacent empty grid cells to form a connected region mask. This mask is then logically ANDed with the original grid map to obtain the final drivable area map.

[0036] For example, a binary grid of passable areas can be generated based on environmental perception information. This means that free space in the environment that is not occupied by obstacles and is connected to the starting point is marked as a drivable area for subsequent path search, thereby avoiding the planning of invalid paths that are unreachable or require detours through walls.

[0037] In other implementations, road semantic information (such as lane lines and traffic signs) can be further introduced to assist in determining drivability.

[0038] Step S200: Within the vehicle's drivable area, perform a path search based on the vehicle's status information and the target driving information to generate path reference information.

[0039] The path reference information is used to characterize a preliminary feasible path from the current vehicle position to the target position. Its role is to provide a topologically connected path solution under the premise of satisfying environmental safety constraints, as the initial input for subsequent continuous processing and trajectory optimization, thereby improving the convergence efficiency and stability of the overall planning algorithm.

[0040] In one alternative embodiment, such as Figure 2 As shown, step S200 includes the following sub-steps: S201, determine the current location information in the vehicle status information as the starting point for path search.

[0041] The current location information refers to the vehicle's spatial position in the local map coordinate system. The starting geographic reference is the initial location for the path planning process. Exemplarily, the acquired real-time vehicle pose data is spatially aligned with the environmental grid map to identify the corresponding grid cell and mark it as the starting node for path searching. For example, when the vehicle is in the center of the road, it is projected onto the nearest drivable grid cell, ensuring the starting point is within the effective drivable range. If the starting point does not accurately correspond to the vehicle's actual pose, the planned path may not align with the vehicle's direction of movement, leading to drastic fluctuations in control commands or trajectory jumps, affecting driving safety and comfort.

[0042] S202, the target location coordinates in the target driving information are determined as the endpoint of the path search.

[0043] Among them, the target location coordinates This represents the target geographical location for this path planning. Exemplarily, these coordinates are mapped to the corresponding spatial location on the environmental grid map and converted into a grid index, serving as the target endpoint for the search algorithm. If this location is occupied by obstacles or is not within a drivable area, the system automatically selects a nearby, idle, connected grid as an alternative endpoint to ensure the search task is solvable. For example, the reachability of the endpoint directly affects whether a valid solution exists for the path; therefore, it must be dynamically judged and adjusted based on the environmental conditions to prevent the algorithm from failing due to the endpoint being unreachable.

[0044] S203. Within the drivable area, a heuristic search algorithm is used to explore the path with the starting point and the ending point as boundary conditions, generating a discrete node sequence containing multiple intermediate nodes, and using the discrete node sequence as path reference information.

[0045] Among them, the heuristic search algorithm is a graph traversal method that uses a cost evaluation mechanism to guide the search direction, and can efficiently find a better path in complex obstacle environments.

[0046] As an example, the A* algorithm is used to perform the path exploration process, and its state space is defined as vehicle coordinates. The evaluation function is:

[0047] in, The actual path cost from the initial state to the current state is calculated by considering the distance traveled and the penalty for crossing high-risk areas. This is a heuristic function based on obstacle distance, balancing path length and safety constraints. The final output is a set of discrete path points connected in sequence. This constitutes path reference information.

[0048] Understandably, although the discrete node sequence has bends, it has avoided all static obstacles and maintained a basic alignment with the road direction, so that cubic spline interpolation can be performed in subsequent steps, significantly reducing the sensitivity of subsequent optimization to the initial value.

[0049] Step S300: The path reference information is processed to make it continuous, and reference trajectory information for vehicle motion control is constructed.

[0050] Among them, continuous processing refers to transforming discrete, non-smooth path reference information into a trajectory expression with geometric continuity, so as to meet the basic requirements of vehicle motion control system for path smoothness and differentiability. Through this processing, a reference trajectory with continuously changing position, direction and curvature is generated as the basis for subsequent dynamic modeling and optimization, effectively improving the feasibility of planning results and driving comfort.

[0051] In one alternative embodiment, such as Figure 3 As shown, step S300 includes the following sub-steps: S301, perform resampling processing on the discrete node sequence in the path reference information to generate a resampled point set.

[0052] Resampling involves inserting new sampling points between the original discrete nodes to make the path points more evenly distributed, avoiding oscillations or distortions in the interpolation curve caused by excessively low or uneven node density. For example, based on the overall path length and local curvature characteristics, intermediate points are inserted between nodes at preset intervals. For instance, when the distance between adjacent nodes exceeds a set threshold, several intermediate coordinates are automatically added along the straight or circular path to form a point set with consistent density. To illustrate, if the original A* output path has sparse points in curved areas and excessively dense points in straight areas, direct fitting can easily produce curvature jumps. Resampling to generate a reasonably distributed point set can significantly improve the quality of subsequent interpolation, providing data support for constructing high-order continuous trajectories.

[0053] S302, based on the resampled point set, constructs a cubic spline curve to obtain the trajectory function continuously expressed under the path length parameter.

[0054] A cubic spline curve is a piecewise cubic polynomial function that can precisely pass through all given points while ensuring C² continuity (i.e., continuity of position, velocity, and acceleration). Exemplarily, the resampled point set is used as control points, with arc length... Construct x and y coordinate functions in Cartesian coordinate system for the parameters. This allows the entire path to be represented as a vector function:

[0055] in, The total arc length of the path; this function is continuously differentiable throughout its domain and supports subsequent differentiation operations.

[0056] After adopting cubic spline interpolation, the path no longer has sharp angles, the curvature changes smoothly, and it can be smoothly tracked by the vehicle controller, avoiding frequent steering wheel vibration and improving the driving experience.

[0057] S303 extracts position, direction, and curvature information from the trajectory function to form a reference trajectory representation.

[0058] Here, direction refers to the tangent direction at each point on the path. Determined by the first derivative of the coordinate function; curvature This indicates the degree of path curvature, reflecting the magnitude of the lateral acceleration required for the vehicle to turn. Exemplarily, the trajectory function described above is used... Calculate its first and second derivatives, and substitute them into the following formula to solve for the geometric properties:

[0059] System parameters along arc length Multiple inspection points are sampled uniformly, and their corresponding values ​​are calculated point by point. and and location information Combined into a complete reference trajectory representation.

[0060] S304 converts the reference trajectory representation to the coordinate system used by the vehicle control system to form reference trajectory information.

[0061] Vehicle control systems often use the Frenet coordinate system for trajectory tracking. This coordinate system uses a reference line as a baseline and is decomposed into longitudinal components along the path. Lateral component of the off-path Demonstratively, a mapping relationship between Cartesian coordinates and Frenet coordinates is established: for any vehicle's current position... Projecting this onto the corresponding point on the nearest reference line forms the reference trajectory information. , .

[0062] In other implementations, the coordinates can be converted to other local coordinate frames according to the actual system requirements.

[0063] Step S400: Establish a vehicle motion control model based on the reference trajectory information and perform trajectory optimization to generate the target driving trajectory.

[0064] Among them, the vehicle motion control model is used to describe the dynamic law of the vehicle state evolution over time, and trajectory optimization refers to the fine adjustment of the initial path under the premise of satisfying dynamic constraints, so as to make it more in line with the actual operating capabilities of the vehicle. Through this process, a final driving trajectory that takes into account safety, smoothness and feasibility is obtained, achieving high-quality path tracking.

[0065] In an optional embodiment, step S400 includes the following sub-steps: S401 determines the motion state sequence of the vehicle in the discrete time domain based on reference trajectory information.

[0066] Here, the motion state sequence refers to a set of vehicle state vectors arranged in chronological order, with each state containing a longitudinal arc length. lateral deviation ,speed Heading error Related variables; discrete time domain refers to dividing the planning period into... There are 1 time steps, each with a length of 1. .

[0067] As an example, the state variables at each time step are initialized along the reference trajectory, and the start and end points are projected onto the Frenet coordinate system to obtain... and This forms the initial state trajectory. For example, this initial state sequence serves as the starting point for the optimization algorithm and directly affects the subsequent convergence speed and solution stability; if the initial value deviation is too large, it may lead to optimization failure or getting trapped in a local optimum.

[0068] S402, Determine the set of control variables used for vehicle control.

[0069] Among these, the control variables are adjustable parameters that affect the vehicle's motion behavior. As an example, longitudinal acceleration is selected. and yaw rate As a control input, denoted as These variables will be dynamically adjusted during the optimization process to minimize the overall cost function. For example, the selection of control variables must match the capabilities of the vehicle's actuators to ensure that the optimization results are physically feasible.

[0070] S403 establishes a vehicle motion control model and trajectory optimization objective function based on the motion state sequence and control variable set.

[0071] The vehicle motion control model is used to characterize the state transition relationships. The trajectory optimization objective function is used to quantify trajectory quality and is typically composed of a weighted average of multiple performance metrics. An approximate kinematic model in Frenet coordinates is used as an example:

[0072] in, It is approximately the angle between the heading and the reference tangent.

[0073] The trajectory optimization objective function is a quadratic cost function, and the stage cost is set as follows: Include: Lateral deviation and heading error costs:

[0074] Cost of longitudinal deviation:

[0075] Cost of control quantity:

[0076] Cost of controlling the rate of change:

[0077] Obstacle distance constraint cost:

[0078] The total cost function is:

[0079] in For terminal costs.

[0080] S404, in the discrete time domain, uses an iterative optimization algorithm to solve the trajectory optimization objective function, and uses the solved state trajectory as the target driving trajectory.

[0081] Iterative optimization algorithms are numerical methods that find the optimal solution through multiple approximations, and are suitable for handling nonlinear, multi-constraint trajectory planning problems. As an example, the iLQR (Iterative Linear Quadratic Regulator) algorithm is used to perform the solution process: First, the linearized motion model is then used to quadratize the cost function at the current trajectory point; Then, the feedback gain and feedforward term are calculated by backward Ricardi recursion. Update the state trajectory again using forward simulation; Repeat the above process until the cost decreases below the threshold or the maximum number of iterations is reached. The final output is the converged state sequence, which is the optimal driving trajectory.

[0082] For example, collision detection is performed simultaneously during the forward simulation. If a trajectory segment is found to enter an obstacle area, the step size is appropriately reduced and the simulation is repeated to ensure that the output trajectory always meets the obstacle avoidance requirements.

[0083] The iterative optimization algorithm uses iLQR (Iterative Linear Quadratic Regulator). The process is as follows: 1. Initialization: Generate an initial state sequence based on the reference trajectory. and control sequence ; 2. Linearization and Quadraticization: At the current trajectory point First-order linearization is performed on the nonlinear model:

[0084] in ; Performing a Taylor expansion on the cost function yields a quadratic term matrix. .

[0085] 3. Backward Riccati Iteration (Reverse Calculation): initialization:

[0086] Recursion:

[0087]

[0088] 4. Forward simulation update: Update Increment:

[0089] New trajectory:

[0090] in This is the step size factor.

[0091] The iLQR outputs the optimal control sequence. With state trajectory (Frenet indicates) that by converting back to Cartesian coordinates, the position and heading of the final time / sampling sequence can be obtained.

[0092] For example, collision detection is performed simultaneously during the forward trajectory calculation. If a trajectory segment is found to enter an obstacle area, the step size is appropriately reduced and the simulation is repeated to ensure that the output trajectory always meets the obstacle avoidance requirements.

[0093] In one optional implementation, the process of solving the problem using an iterative optimization algorithm includes: Calculate the difference between the state trajectory generated in this iteration and the state trajectory generated in the previous iteration.

[0094] Here, the state trajectory refers to a sequence of vehicle states at multiple time steps in the discrete time domain, with each state containing relevant variables such as position, orientation, and velocity. The difference is used to measure the degree of change between adjacent iterations, reflecting the convergence trend of the optimization process. For example, after each iteration, the system compares the state vectors of the current iteration with those of the previous iteration point by point, calculating a scalar difference value by measuring the overall deviation magnitude. This value can be used to determine whether the optimization is approaching stability. For instance, if the trajectories generated in two consecutive iterations show minimal changes in key indicators such as lateral deviation and heading error, it indicates that the optimal solution is close, and further iteration is unnecessary, thus saving computational resources and improving response efficiency.

[0095] If the difference is less than the preset convergence threshold, the iteration stops and the current state trajectory is used as the target driving trajectory. Otherwise, iterative optimization continues until the maximum number of iterations is reached, and the state trajectory generated in the last iteration is used as the target driving trajectory.

[0096] The convergence threshold is a pre-set small value used to determine whether the optimization process has entered a stable phase; the maximum number of iterations is the maximum number of optimization rounds allowed. For example, when the system detects that the current difference is below this threshold, it immediately terminates the optimization process and outputs the latest trajectory as the final result. If the difference consistently fails to meet the threshold, linearization and forward simulation updates continue until the maximum number of iterations is reached, at which point the last generated state trajectory is forcibly output. For instance, in complex scenarios with frequently moving dynamic obstacles, complete convergence may be difficult, but by setting the maximum number of iterations, the planning module can still output a usable trajectory within a limited time, meeting the real-time control requirements of the autonomous driving system.

[0097] In an optional embodiment, the driving trajectory planning method further includes the following steps: Collision detection is performed on the target vehicle's trajectory.

[0098] Collision detection refers to determining whether the target vehicle trajectory spatially intrudes into obstacles in the environmental perception information, thus verifying the safety of the planning results. For example, the system discretizes the generated target vehicle trajectory into a series of trajectory points and calculates the distance from each point to the nearest obstacle. If the distance between any trajectory point and an obstacle is less than a preset safe distance threshold, a collision risk is identified. For instance, during lane changing, if the trajectory generated in the previous optimization round is too close to a dynamic vehicle in the adjacent lane, an obstacle avoidance warning may be triggered due to insufficient lateral safety margin. In this case, a replanning mechanism needs to be activated to ensure driving safety.

[0099] If the detection results do not meet the preset safe distance requirements, the path reference information will be regenerated.

[0100] The path reference information forms the basis for subsequent continuous processing and trajectory optimization. For example, the system returns to the path search phase, combines the updated obstacle positions or adjusts the heuristic weights, and re-executes algorithms such as A* to generate a new path away from high-risk areas. For instance, when the original path crosses a static construction fence near a curve exit, the system can guide a new path to bypass it by increasing the obstacle's expansion radius, thus avoiding potential contact.

[0101] The reference trajectory information is updated based on the regenerated path reference information, and trajectory optimization is performed on the updated reference trajectory information to generate the target driving trajectory.

[0102] The reference trajectory information is obtained through continuous processing and serves as the benchmark for motion control modeling. Exemplarily, the system performs resampling and cubic spline interpolation on the new path to construct a smooth reference curve. Based on this, the state sequence and control variables are reinitialized, and iterative optimization algorithms such as iLQR are called again to solve for the new target vehicle trajectory. For example, after path replanning, the newly generated trajectory not only avoids obstacles but also maintains curvature continuity and smooth transitions, ensuring stable vehicle tracking without sudden braking or sharp steering. In other implementations, multiple safety thresholds can be set according to actual needs to achieve a graded obstacle avoidance response strategy.

[0103] Figure 4 A schematic diagram of a driving trajectory planning device according to an embodiment of this application is shown. Exemplarily, the device 100 includes: The area determination module 110 is used to acquire vehicle status information, target driving information and environmental perception information, and determine the driving area of ​​the vehicle based on the environmental perception information. The path search module 120 is used to perform path search based on the vehicle status information and target driving information within the vehicle's drivable area to generate path reference information. The path processing module 130 is used to perform continuous processing on the path reference information to construct reference trajectory information for vehicle motion control; The trajectory generation module 140 is used to establish a vehicle motion control model based on the reference trajectory information and perform trajectory optimization to generate a target driving trajectory.

[0104] It is understood that the apparatus of this embodiment corresponds to the method of the above embodiments, and the options in the above embodiments are also applicable to this embodiment, so they will not be described again here.

[0105] This application also provides a terminal device, exemplary of which includes a processor and a memory, wherein the memory stores a computer program, and the processor executes the computer program to enable the terminal device to perform the functions of the various modules in the above-described method or apparatus.

[0106] The processor can be an integrated circuit chip with signal processing capabilities. The processor can be a general-purpose processor, including at least one of a Central Processing Unit (CPU), Graphics Processing Unit (GPU), Network Processor (NP), Digital Signal Processor (DSP), Application-Specific Integrated Circuit (ASIC), Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. The general-purpose processor can be a microprocessor or any conventional processor, capable of implementing or executing the methods, steps, and logic block diagrams disclosed in the embodiments of this application.

[0107] The memory can be, but is not limited to, Random Access Memory (RAM), Read Only Memory (ROM), Programmable Read-Only Memory (PROM), Erasable Programmable Read-Only Memory (EPROM), Electrically Erasable Programmable Read-Only Memory (EEPROM), etc. The memory is used to store computer programs, and the processor can execute the computer programs accordingly after receiving execution instructions.

[0108] This application also provides a computer-readable storage medium for storing the computer program used in the aforementioned terminal device. For example, the computer-readable storage medium may include, but is not limited to, various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0109] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can also be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings show the architecture, functionality, and operation of possible implementations of apparatus, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that, in alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and combinations of blocks in the block diagram and / or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0110] In addition, the functional modules or units in the various embodiments of this application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0111] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a smartphone, personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application.

[0112] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A method for planning vehicle trajectory, characterized in that, The method includes: Acquire vehicle status information, target driving information, and environmental perception information, and determine the vehicle's drivable area based on the environmental perception information; Within the vehicle's drivable area, a path search is performed based on the vehicle's status information and the target driving information to generate path reference information; The path reference information is processed to be continuous to construct reference trajectory information for vehicle motion control; A vehicle motion control model is established based on the reference trajectory information, and trajectory optimization is performed to generate the target driving trajectory.

2. The driving trajectory planning method according to claim 1, characterized in that, After generating the target driving trajectory, the following is also included: Collision detection is performed on the target vehicle trajectory; If the detection results do not meet the preset safe distance requirements, the path reference information will be regenerated. The reference trajectory information is updated based on the regenerated path reference information, and trajectory optimization is performed on the updated reference trajectory information to generate the target driving trajectory.

3. The driving trajectory planning method according to claim 1, characterized in that, The process of acquiring vehicle status information, target driving information, and environmental perception information, and determining the drivable area of ​​the vehicle based on the environmental perception information, includes: The vehicle's current location, heading angle, and speed are obtained to form the vehicle's state information; Obtain the target location coordinates in the preset navigation path to construct the target driving information; Acquire laser point cloud or obstacle projection data containing obstacle distribution, and perform rasterization processing on the laser point cloud or obstacle projection data to generate an environmental raster map to constitute the environmental perception information. The connected regions in the environmental grid map that are not occupied by obstacles are defined as the drivable areas for the vehicle.

4. The driving trajectory planning method according to claim 1, characterized in that, Within the vehicle's drivable area, a path search is performed based on the vehicle's status information and the target driving information to generate path reference information, including: The current location information in the vehicle status information is used as the starting point for path search; The target location coordinates in the target driving information are determined as the endpoint of the path search; Within the drivable area of ​​the vehicle, a heuristic search algorithm is used to explore the path, with the starting point and the ending point as boundary conditions, to generate a discrete node sequence, which is then used as path reference information.

5. The driving trajectory planning method according to claim 1, characterized in that, The step of performing continuous processing on the path reference information to construct reference trajectory information for vehicle motion control includes: Resampling is performed on the discrete node sequence in the path reference information to generate a set of resampled points; Based on the resampled point set, a cubic spline curve is constructed to obtain the trajectory function continuously expressed under the path length parameter; The position, direction, and curvature information of the trajectory function are extracted to form a reference trajectory representation; The reference trajectory representation is converted to the coordinate system used by the vehicle control system to form the reference trajectory information.

6. The driving trajectory planning method according to claim 1, characterized in that, The step of establishing a vehicle motion control model based on the reference trajectory information and performing trajectory optimization to generate a target driving trajectory includes: Based on the reference trajectory information, determine the vehicle's motion state sequence in the discrete time domain; Determine the set of control variables used for vehicle control; Based on the motion state sequence and the set of control variables, a vehicle motion control model and a trajectory optimization objective function are established. Within the discrete time domain, an iterative optimization algorithm is used to solve the trajectory optimization objective function; The obtained state trajectory is used as the target driving trajectory.

7. The driving trajectory planning method according to claim 6, characterized in that, The process of solving the problem using the iterative optimization algorithm includes: Calculate the difference between the state trajectory generated in this iteration and the state trajectory generated in the previous iteration; If the difference is less than a preset convergence threshold, the iteration stops and the current state trajectory is taken as the target driving trajectory. Otherwise, continue iterative optimization until the maximum number of iterations is reached, and use the state trajectory generated in the last iteration as the target driving trajectory.

8. A vehicle trajectory planning device, characterized in that, include: The area determination module is used to acquire vehicle status information, target driving information and environmental perception information, and determine the driving area of ​​the vehicle based on the environmental perception information. The path search module is used to perform path search based on the vehicle status information and the target driving information within the vehicle's drivable area, so as to generate path reference information; The path processing module is used to perform continuous processing on the path reference information to construct reference trajectory information for vehicle motion control; The trajectory generation module is used to establish a vehicle motion control model based on the reference trajectory information and perform trajectory optimization to generate a target driving trajectory.

9. A terminal device, characterized in that, The terminal device includes a processor and a memory, the memory storing a computer program, and the processor executing the computer program to implement the driving trajectory planning method according to any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, It stores a computer program, which, when executed on a processor, implements the driving trajectory planning method according to any one of claims 1-7.

Images