Intelligent vehicle obstacle avoidance method based on spatiotemporal environment safety scoring model

Abstract

This invention provides an intelligent vehicle obstacle avoidance method based on a spatiotemporal environmental safety scoring model, comprising: fusing multi-sensor data using extended Kalman filtering to obtain a rasterized environmental model, acquiring the vehicle's current position, obstacle positions, and target position; establishing a spatiotemporal environmental safety scoring model by introducing a nonlinear adjustment factor into the ant colony algorithm to obtain the final probability of each ant path being selected; constructing an enhanced pheromone update function using an improved genetic algorithm fitness formula to obtain an updated ant path; constructing a preliminary planned path; constructing improved gravitational and repulsive field functions, applying the improved gravitational and repulsive field functions to the preliminary planned path, and then combining the driving environment model to adjust and update the path, finally obtaining the optimal obstacle avoidance path. This invention can improve the safety and reliability of vehicle obstacle avoidance in complex traffic scenarios.

Description

Technical Field

[0001] This invention relates to the technical field of intelligent vehicles, and specifically to an obstacle avoidance method for intelligent vehicles based on a spatiotemporal environmental safety scoring model. Background Technology

[0002] Intelligent vehicle path planning is a series of cutting-edge technologies based on perception, decision-making, planning, and control. Path planning generates an optimal path for a driving vehicle to smoothly reach its destination from a starting point in a given road environment. The optimal path needs to meet the following conditions: shortest distance, shortest time, lowest energy consumption, and effective obstacle avoidance.

[0003] While traditional path planning algorithms can generate feasible paths from the starting point to the destination in the field of obstacle avoidance for intelligent vehicles, they do not adequately consider the smoothness and feasibility of the paths when faced with complex and dynamic traffic scenarios. The generated paths may pose certain safety hazards due to frequent turning or sudden braking. Summary of the Invention

[0004] In view of this, the present invention provides an intelligent vehicle obstacle avoidance method based on a spatiotemporal environment safety scoring model to improve the safety and reliability of vehicle obstacle avoidance in complex traffic scenarios.

[0005] An intelligent vehicle obstacle avoidance method based on a spatiotemporal environment safety scoring model includes:

[0006] Step S1: Use extended Kalman filter to fuse multi-sensor data to obtain a gridded environment model, and obtain the vehicle's current position, obstacle position and vehicle target position based on the gridded environment model;

[0007] Step S2: By introducing a nonlinear adjustment factor into the ant colony algorithm, a spatiotemporal environment safety scoring model is established to obtain the final probability of each ant path being selected.

[0008] Step S3: Based on the final probability of each ant path being selected, an enhanced pheromone update function is constructed using an improved genetic algorithm fitness formula to obtain the updated ant paths.

[0009] Step S4: Establish a driving environment model based on the vehicle's current position and target position, and calculate the target direction, obstacle direction, and comprehensive motion direction vector matrix; select feasible paths that meet the target point matching and collision-free conditions from all updated ant paths to obtain the preliminary planned path;

[0010] Step S5: The spatiotemporal environment safety scoring model established in step S2 is introduced into the artificial potential field to obtain improved gravitational and repulsive field functions. The improved gravitational and repulsive field functions are then applied to the preliminary planned path. Combined with the driving environment model, the path is adjusted and updated to finally obtain the optimal obstacle avoidance path.

[0011] The intelligent vehicle obstacle avoidance method based on a spatiotemporal environment safety scoring model provided by the present invention has the following beneficial effects:

[0012] 1. Based on the final probability of each ant path being selected, this invention uses an improved genetic algorithm fitness formula to construct an enhanced pheromone update function, so that the pheromone update no longer depends solely on the path length. This allows the path selection probability and fitness value to comprehensively reflect the quality of the path, and the pheromone is dynamically adjusted accordingly, thereby improving the adaptive ability of obstacle avoidance in complex traffic scenarios.

[0013] 2. This invention combines a spatiotemporal environmental safety scoring model to evaluate the vehicle's driving environment in real time. By dynamically adjusting the magnitude and direction of attraction and repulsion, it plans the optimal obstacle avoidance path, effectively avoiding collision risks, improving the safety and reliability of obstacle avoidance, and reducing the possibility of accidents.

[0014] 3. This invention utilizes a spatiotemporal environmental safety scoring model, enabling vehicles to flexibly adjust their driving strategies based on the spatiotemporal environmental safety scores of different road conditions and traffic scenarios. In environments with lower safety scores, vehicles can plan their routes more cautiously and increase the safe distance from surrounding obstacles; while in environments with higher safety scores, vehicles can appropriately increase their driving efficiency, achieving a balance between safety and efficiency. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating the intelligent vehicle obstacle avoidance method based on a spatiotemporal environment safety scoring model provided in an embodiment of the present invention.

[0016] Figure 2 This is a path simulation comparison diagram of the method of the present invention with the traditional ACO-GA method and the traditional APF method;

[0017] Figure 3 This is a comparison chart of the path length of the method of the present invention and the traditional APF method at different iteration numbers. Detailed Implementation

[0018] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain embodiments of the present invention, and should not be construed as limiting the present invention.

[0019] Please see Figure 1 The embodiments of the present invention provide an intelligent vehicle obstacle avoidance method based on a spatiotemporal environment safety scoring model, including steps S1-S5.

[0020] Step S1: Use extended Kalman filter to fuse multi-sensor data to obtain a gridded environment model, and obtain the vehicle's current position, obstacle position and vehicle target position based on the gridded environment model.

[0021] The state update equation for the extended Kalman filter is as follows:

[0022]

[0023] in, It is the fused state estimate, which is the updated state estimate based on the current measurement and prior estimate, and is also the posterior state estimate; It is in the The prior state estimate at each time step is obtained based on the state estimate at the previous time step and the prediction of the system model. It is Kalman gain. It is the measurement value at the current moment. It is a measurement matrix; This is a weighted coefficient matrix used to adjust the degree of influence of obstacle size on state estimation; Indicates the size parameters of the obstacle.

[0024] Step S2: By introducing a nonlinear adjustment factor into the ant colony algorithm, a spatiotemporal environment safety scoring model is established to obtain the final probability of each ant path being selected.

[0025] Specifically, step S2 includes:

[0026] Step S201: Set the vehicle speed and the distance between the vehicle and the obstacle as target parameters. Measure the convergence speed by comparing the changes in the historical best value of each target parameter within a preset period. Calculate the corresponding convergence degree for each target parameter and normalize the convergence degree of each target parameter to obtain the nonlinear adjustment factor for each target parameter.

[0027] Step S202: Introduce a nonlinear adjustment factor into the ant colony algorithm, construct a spatiotemporal environment safety scoring model, and calculate the spatiotemporal environment safety score of the ant path.

[0028] Step S203: Based on the spatiotemporal environment safety score, calculate the initial selection probability of the ant colony on the ant path, and combine the vehicle speed and the distance between the vehicle and the obstacle to obtain the real-time congestion of the ant colony.

[0029] Step S204: Introduce a spatiotemporal environment safety score into the ant colony algorithm to obtain the enhanced pheromone value;

[0030] Step S205: Based on the enhanced pheromone value and the real-time crowding of the ant population, the mutation probability of the ant population is obtained, and combined with the initial selection probability of the ant population on the ant path, a final selection probability function is constructed to obtain the final probability of each ant path being selected.

[0031] Wherein, step S201 satisfies the following condition:

[0032]

[0033]

[0034] in, For the first The goal is at any time The degree of convergence represents the convergence speed of the objective. The maximum convergence among all objectives. For the first The goal is at any time The optimal value; The observation window size represents the time interval for calculating convergence. For the first The goal is at a certain moment The optimal value, For the first The goal is at a certain moment The nonlinear adjustment factor, As a basic adjustment factor, in this embodiment, .

[0035] Step S202 satisfies the following condition:

[0036]

[0037]

[0038]

[0039] in, For the ant path at time The spatiotemporal environment safety score, This is the nonlinear adjustment factor for the speed regulation term. This is the nonlinear adjustment factor for the distance adjustment term. The normalized speed, The normalized distance between the vehicle and the obstacle. for The speed of the vehicle at any time for The distance between the vehicle and the obstacle at any given time. For the target speed, The safe distance between a vehicle and an obstacle. This represents the vehicle's maximum speed. This is the farthest distance that the vehicle's sensors can detect.

[0040] Step S203 satisfies the following condition:

[0041]

[0042]

[0043] in, It is the ant path at any moment The probability of being selected in the initial selection. For all candidate ant paths at time... Maximum spatiotemporal environmental safety score For the first The real-time crowding level of the ant colony during the next iteration. For the first The number of times an ant path appears in the ant population during each iteration. This represents the number of ants in an ant colony. The maximum number of iterations, The weighting coefficient for the urgency of obstacle avoidance. The weighting coefficients for speed sensitivity, The distance between the vehicle and the nearest obstacle. This is the speed adjustment coefficient.

[0044] Step S204 satisfies the following condition:

[0045]

[0046] in, for Enhanced pheromone value This represents the original pheromone value.

[0047] Step S205 satisfies the following condition:

[0048]

[0049]

[0050] in, for The first in an ant colony The mutation probability of an ant. The maximum pheromone value, The attenuation coefficient is... This is an adjustment factor for the initial selection probability. for The final probability that an ant path is selected.

[0051] At this time, it is in a simple non-interacting state, which represents the basic state without additional regulation. At this time, the system exhibits reinforcing interactions, and the probability is further amplified as its value increases, forming a mechanism of preferential reinforcement. When the system exhibits inhibitory interactions, the probability is inhibited by itself. As the probability increases, the amplified effect weakens, thus stabilizing or balancing the system.

[0052] Step S3: Based on the final probability of each ant path being selected, an enhanced pheromone update function is constructed using an improved genetic algorithm fitness formula to obtain the updated ant paths.

[0053] The improved genetic algorithm fitness formula is expressed as follows:

[0054]

[0055]

[0056]

[0057] in, For the fitness of the genetic algorithm, The weighting coefficient is the path length. The weighting coefficients for comfort indicators. The weighting coefficient for driving safety. This represents the normalized ant path length. For the normalized vehicle acceleration, This represents the current ant path length. and These are the maximum and minimum values ​​of the ant path length, respectively. For the current vehicle acceleration, and These are the maximum and minimum values ​​of the vehicle's acceleration, respectively.

[0058] The expression for the enhanced pheromone update function is:

[0059]

[0060]

[0061] in, For the first An ant on the edge of an obstacle The increase in pheromones released from the upper body The total amount of pheromones, For the first The length of the path traveled by an ant; For indicator functions, if the first... Only one ant passed through the edge of the obstacle in this iteration. ,but ,otherwise, ; The pheromone evaporation coefficient; , These are weighting coefficients used to balance the fitness of the genetic algorithm with the urgency of obstacle avoidance; for Pheromones value after time-lapse enhancement.

[0062] Step S4: Establish a driving environment model based on the vehicle's current position and target position, and calculate the target direction, obstacle direction, and comprehensive motion direction vector matrix; select feasible paths that meet the target point matching and collision-free conditions from all updated ant paths to obtain the preliminary planned path.

[0063] Wherein, the target direction vector matrix The direction from the vehicle's current position to the target position is expressed as:

[0064]

[0065] in, This is the vehicle's current location. The target location of the vehicle. For heading angle, For time intervals;

[0066] obstacle direction vector matrix The direction from the vehicle's current position to each obstacle is represented by the expression:

[0067]

[0068] in, The vehicle's steering angle relative to the nearest obstacle;

[0069] Combined motion direction vector matrix It is the weighted sum of the target direction vector and the direction vectors of all obstacles, expressed as:

[0070]

[0071]

[0072]

[0073] in, As the target factor, For obstacle factors, This represents the number of obstacles.

[0074] Step S5: The spatiotemporal environment safety scoring model established in step S2 is introduced into the artificial potential field to obtain improved gravitational and repulsive field functions. The improved gravitational and repulsive field functions are then applied to the preliminary planned path. Combined with the driving environment model, the path is adjusted and updated to finally obtain the optimal obstacle avoidance path.

[0075] Traditional potential field methods are prone to getting stuck in extreme values ​​and optima in complex obstacle scenarios, thus failing to reach the target point. In this invention, a spatiotemporal environmental safety scoring model is introduced for improvement. In a gravitational field, the spatiotemporal environmental safety scoring model dynamically adjusts the target attraction, enabling the vehicle to reach the target while taking environmental safety into account, avoiding potential dangers caused by excessive pursuit of speed. In a repulsive field, the spatiotemporal environmental safety scoring model strengthens or weakens the repulsive force in real time according to the safety of the environment, enabling the vehicle to cope more effectively with obstacles in dynamic and complex environments, improving the safety and reliability of obstacle avoidance.

[0076] Specifically, in this embodiment, the expression for the improved gravitational field is:

[0077]

[0078] in, for The gravitational field strength at time t. The gravitational intensity coefficient, for The distance between the vehicle and the target location at any given time;

[0079] The expression for the improved repulsive field function is:

[0080]

[0081] in, for The magnitude of the repulsive field at any given moment. The repulsive force intensity coefficient, This refers to the range of influence of the potential field.

[0082] Figure 2 This diagram compares the path of the method of this invention with the traditional ACO-GA method (traditional ant colony-genetic fusion method) and the traditional APF method (traditional artificial potential field method). The simulation experiment used Matlab R2024a. In the simulation, the black solid squares represent obstacles, with some obstacles having coordinates of (5, 7), (5.5, 0.5), (2, 14), and (12, 14); the circle represents the starting point with coordinates of (0.5, 0.5); and the pentagram represents the target point with coordinates of (19.5, 19.5). Figure 2 As can be seen, compared with the traditional APF method, the method of the present invention can effectively solve the problem of target unreachability, and the planned obstacle avoidance path is smoother and more stable. Compared with the traditional ACO-GA method, the present invention takes into account factors such as the actual size of the vehicle and the turning radius of the vehicle, which is more in line with the physical constraints of vehicle motion and has higher feasibility in practical applications.

[0083] Figure 3 This is a comparison chart of the path length of the present invention and the traditional APF method (traditional artificial potential field method) at different iteration numbers. Figure 3 As can be seen, the method of this invention can quickly find a high-quality path and maintain stability, effectively avoiding the problem of getting trapped in local optima, and has a faster convergence speed and better stability.

[0084] comprehensive Figure 2 and Figure 3 As can be seen, compared with traditional methods, this invention not only effectively solves the problem of algorithms easily getting trapped in local optima, but also obtains obstacle avoidance paths with shorter distances and higher smoothness, enabling safe and efficient obstacle avoidance.

[0085] In summary, the intelligent vehicle obstacle avoidance method based on the spatiotemporal environment safety scoring model according to the above embodiments has the following beneficial effects:

[0086] 1. Based on the final probability of each ant path being selected, this invention uses an improved genetic algorithm fitness formula to construct an enhanced pheromone update function, so that the pheromone update no longer depends solely on the path length. This allows the path selection probability and fitness value to comprehensively reflect the quality of the path, and the pheromone is dynamically adjusted accordingly, thereby improving the adaptive ability of obstacle avoidance in complex traffic scenarios.

[0087] 2. This invention combines a spatiotemporal environmental safety scoring model to evaluate the vehicle's driving environment in real time. By dynamically adjusting the magnitude and direction of attraction and repulsion, it plans the optimal obstacle avoidance path, effectively avoiding collision risks, improving the safety and reliability of obstacle avoidance, and reducing the possibility of accidents.

[0088] 3. This invention utilizes a spatiotemporal environmental safety scoring model, enabling vehicles to flexibly adjust their driving strategies based on the spatiotemporal environmental safety scores of different road conditions and traffic scenarios. In environments with lower safety scores, vehicles can plan their routes more cautiously and increase the safe distance from surrounding obstacles; while in environments with higher safety scores, vehicles can appropriately increase their driving efficiency, achieving a balance between safety and efficiency.

[0089] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention. Therefore, the protection scope of this invention patent should be determined by the appended claims.

Claims

1. An intelligent vehicle obstacle avoidance method based on a spatiotemporal environmental safety scoring model, characterized in that, include: Step S1: Use extended Kalman filter to fuse multi-sensor data to obtain a gridded environment model, and obtain the vehicle's current position, obstacle position and vehicle target position based on the gridded environment model; Step S2: By introducing a nonlinear adjustment factor into the ant colony algorithm, a spatiotemporal environment safety scoring model is established to obtain the final probability of each ant path being selected. Step S3: Based on the final probability of each ant path being selected, an enhanced pheromone update function is constructed using an improved genetic algorithm fitness formula to obtain the updated ant paths. Step S4: Establish a driving environment model based on the vehicle's current position and target position, and calculate the target direction, obstacle direction, and comprehensive motion direction vector matrix; select feasible paths that meet the target point matching and collision-free conditions from all updated ant paths to obtain the preliminary planned path; Step S5: The spatiotemporal environment safety scoring model established in step S2 is introduced into the artificial potential field to obtain improved gravitational and repulsive field functions. The improved gravitational and repulsive field functions are then applied to the preliminary planned path. Combined with the driving environment model, the path is adjusted and updated to finally obtain the optimal obstacle avoidance path.

2. The intelligent vehicle obstacle avoidance method based on a spatiotemporal environment safety scoring model according to claim 1, characterized in that, Step S2 specifically includes: Step S201: Set the vehicle speed and the distance between the vehicle and the obstacle as target parameters. Measure the convergence speed by comparing the changes in the historical best value of each target parameter within a preset period. Calculate the corresponding convergence degree for each target parameter and normalize the convergence degree of each target parameter to obtain the nonlinear adjustment factor for each target parameter. Step S202: Introduce a nonlinear adjustment factor into the ant colony algorithm, construct a spatiotemporal environment safety scoring model, and calculate the spatiotemporal environment safety score of the ant path. Step S203: Based on the spatiotemporal environment safety score, calculate the initial selection probability of the ant colony on the ant path, and combine the vehicle speed and the distance between the vehicle and the obstacle to obtain the real-time congestion of the ant colony. Step S204: Introduce a spatiotemporal environment safety score into the ant colony algorithm to obtain the enhanced pheromone value; Step S205: Based on the enhanced pheromone value and the real-time crowding of the ant population, the mutation probability of the ant population is obtained, and combined with the initial selection probability of the ant population on the ant path, a final selection probability function is constructed to obtain the final probability of each ant path being selected.

3. The intelligent vehicle obstacle avoidance method based on a spatiotemporal environment safety scoring model according to claim 2, characterized in that, Step S201 satisfies the following condition: in, For the first The goal is at any time The convergence, The maximum convergence among all objectives. For the first The goal is at any time The optimal value, For the size of the observation window, For the first The goal is at a certain moment The optimal value, For the first The goal is at a certain moment The nonlinear adjustment factor, It is a basic regulatory factor.

4. The intelligent vehicle obstacle avoidance method based on a spatiotemporal environment safety scoring model according to claim 3, characterized in that, Step S202 satisfies the following condition: in, For the ant path at time The spatiotemporal environment safety score, This is the nonlinear adjustment factor for the speed regulation term. This is the nonlinear adjustment factor for the distance adjustment term. The normalized speed, The normalized distance between the vehicle and the obstacle. for The speed of the vehicle at any time for The distance between the vehicle and the obstacle at any given time. For the target speed, The safe distance between a vehicle and an obstacle. This represents the vehicle's maximum speed. This is the farthest distance that the vehicle's sensors can detect.

5. The intelligent vehicle obstacle avoidance method based on a spatiotemporal environment safety scoring model according to claim 4, characterized in that, Step S203 satisfies the following condition: in, It is the ant path at any time The probability of being selected in the initial selection. For all candidate ant paths at time... Maximum spatiotemporal environmental safety score For the first The real-time crowding level of the ant colony during the next iteration. For the first The number of times an ant path appears in the ant population during each iteration. This refers to the number of ants in an ant colony. The maximum number of iterations, The weighting coefficient for the urgency of obstacle avoidance. The weighting coefficients for speed sensitivity, The distance between the vehicle and the nearest obstacle. This is the speed adjustment coefficient.

6. The intelligent vehicle obstacle avoidance method based on a spatiotemporal environment safety scoring model according to claim 5, characterized in that, Step S204 satisfies the following condition: in, for Enhanced pheromone value This represents the original pheromone value.

7. The intelligent vehicle obstacle avoidance method based on a spatiotemporal environment safety scoring model according to claim 6, characterized in that, Step S205 satisfies the following condition: in, for The first in an ant colony The mutation probability of an ant. The maximum pheromone value, The attenuation coefficient is... This is an adjustment factor for the initial selection probability. for The final probability that an ant path is selected.

8. The intelligent vehicle obstacle avoidance method based on a spatiotemporal environment safety scoring model according to claim 7, characterized in that, In step S3, the expression for the improved genetic algorithm fitness formula is: in, For the fitness of the genetic algorithm, The weighting coefficient is the path length. The weighting coefficients for comfort indicators. The weighting coefficient for driving safety. This represents the normalized ant path length. For the normalized vehicle acceleration, This represents the current ant path length. and These are the maximum and minimum values ​​of the ant path length, respectively. For the current vehicle acceleration, and These are the maximum and minimum values ​​of the vehicle's acceleration, respectively. The expression for the enhanced pheromone update function is: in, For the first An ant on the edge of an obstacle The increase in pheromones released from the upper body The total amount of pheromones, For the first The length of the path traveled by an ant. For indicator functions, The pheromone evaporation coefficient, , These are the weighting coefficients. for Pheromones value after time-lapse enhancement.

9. The intelligent vehicle obstacle avoidance method based on a spatiotemporal environment safety scoring model according to claim 8, characterized in that, In step S4, the target direction vector matrix The expression is: in, This is the vehicle's current location. The target location of the vehicle. For heading angle, For time intervals; obstacle direction vector matrix The expression is: in, The vehicle's steering angle relative to the nearest obstacle; Combined motion direction vector matrix The expression is: in, As the target factor, For obstacle factors, This represents the number of obstacles.

10. The intelligent vehicle obstacle avoidance method based on a spatiotemporal environment safety scoring model according to claim 9, characterized in that, In step S5, the expression for the improved gravitational field is: in, for The gravitational field strength at time t. The gravitational intensity coefficient, for The distance between the vehicle and the target location at any given time; The expression for the improved repulsive field function is: in, for The magnitude of the repulsive field at any given moment. The repulsive force intensity coefficient, This refers to the range of influence of the potential field.

Images