Vehicle obstacle avoidance method based on digital twinning
Through the combination of digital twin technology and A* and D* algorithms, the problem that the roller cannot avoid obstacles in the working environment is solved, and the intelligent obstacle avoidance and path planning of the vehicle in complex environments is realized, and safety and response speed are improved. .
Patent Information
- Application Number
- CN202411192597.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-28
- Publication Date
- 2024-11-08
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
In the prior art, the roller cannot effectively avoid obstacles when encountering obstacles in the actual working environment, resulting in safety hazards.
Using a vehicle obstacle avoidance method based on digital twins, by constructing a virtual simulation model, simulating work scenarios, collecting obstacle information, predicting obstacle trajectory, evaluating collision risks, and generating instruction signals and path planning schemes based on obstacle avoidance strategies, using A* and D* algorithm dynamically generate and update the optimal path.
It realizes intelligent obstacle avoidance in complex environments, improves the accuracy and response speed of prediction of obstacle motion states, and ensures the stable operation of the vehicle in complex environments and the smooth execution of tasks.
Smart Images

Figure CN118915775A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of autonomous driving vehicles, and in particular to a vehicle obstacle avoidance method based on digital twins. Background Art
[0002] At present, most road rollers rely on manual driving to operate, but due to the complex construction site environment, safety accidents are prone to occur, posing safety hazards to drivers.
[0003] In order to improve the safety of roller operation, domestic and foreign scholars have begun to study the automatic driving technology of rollers. Most of them use deep learning-based neural network algorithms to identify obstacles, so as to realize the autonomous obstacle avoidance function of unmanned rollers. However, the method of using deep learning-based neural network algorithms to identify obstacles only considers how to identify obstacles, but does not consider how to plan a new route for the roller to follow the new path to avoid obstacles, resulting in the inability to truly solve the problem that the roller cannot avoid obstacles when encountering obstacles in the actual working environment. Summary of the invention
[0004] The purpose of the present invention is to provide a vehicle obstacle avoidance method based on digital twins, which is used to solve the problem in the prior art that vehicles are unable to avoid obstacles when encountering obstacles in an actual working environment.
[0005] The purpose of the present invention can be achieved through the following technical solutions: The present application provides a vehicle obstacle avoidance method based on digital twins, comprising the following steps: Constructing a virtual simulation model and simulating a real working scene, wherein the virtual simulation model is generated by a physical engine; An obstacle image file is generated by collecting obstacle information, and a motion state of the obstacle in the obstacle image file is obtained by using the virtual simulation model; According to the motion state of the obstacle, the trajectory information of the obstacle in the future period is calculated to obtain the predicted obstacle trajectory; According to the motion state of the obstacle and the predicted obstacle trajectory, the collision risk of the vehicle is evaluated and the obstacle avoidance strategy is obtained; Among them, when it is determined that an obstacle needs to be avoided based on the results of the obstacle avoidance strategy, a command signal and a path planning plan are generated.
[0006] Preferably, the virtual simulation model is constructed by acquiring historical behavior data of the vehicle, constructing a mapping relationship between the physical world and the digital world based on the historical behavior data, and utilizing the mapping relationship to transmit information in the real scene to the virtual simulation environment, thereby performing interaction between the virtual simulation environment and the real environment.
[0007] Preferably, the obstacle image file includes information of a plurality of obstacles, each obstacle corresponds to a position coordinate, and obtaining the motion state of the obstacle in the obstacle image file specifically includes: Determining a projection point of the obstacle on the vehicle according to the position coordinates of the obstacle, and then determining an attitude angle of the obstacle relative to the vehicle according to the projection point; The movement speed and acceleration of the obstacle are calculated according to the attitude angle and a preset algorithm, and the movement state of the obstacle in a future period of time is obtained.
[0008] Preferably, when the motion state of an obstacle is uniform motion along a straight line, a Gaussian filter is used to smooth the motion state of the obstacle; When the motion state of an obstacle is not uniform along a straight line, the Kalman filter is used to smooth the motion state of the obstacle.
[0009] Preferably, the predicted obstacle trajectory specifically includes: When there are two or more obstacles that are moving at a uniform speed along a straight line, a multivariate Gaussian mixture model is used to fuse the motion states of multiple obstacles to obtain the trajectory information of multiple obstacles in the future. When there are two or more obstacles whose motion states are not uniform along a straight line, the particle swarm optimization algorithm is used to fuse the motion states of multiple obstacles to obtain the trajectory information of multiple obstacles in the future. The predicted position of the obstacle is obtained by fusion processing based on the movement state of the obstacle.
[0010] Preferably, the evaluating the collision risk of the vehicle specifically includes: judging whether there is a collision risk between the vehicle and the obstacle according to the motion state of the obstacle and the predicted obstacle trajectory, and if there is a collision risk, judging whether the vehicle needs to avoid the obstacle.
[0011] Preferably, an obstacle avoidance strategy is obtained according to the collision risk assessment result of the vehicle, specifically including: When it is determined that there is no risk of collision between the vehicle and the obstacle, the original route is maintained and the vehicle continues to travel; or, when it is determined that there is a risk of collision between the vehicle and the obstacle and there is no need to avoid the obstacle, whether the obstacle needs to be avoided is determined according to the pre-stored driving rules; wherein the driving rules include: When the speed of the obstacle is less than a preset first threshold, the vehicle does not need to avoid the obstacle; Alternatively, when the speed of the obstacle is greater than a preset second threshold, the vehicle needs to avoid the obstacle; Alternatively, when an obstacle is within a certain distance ahead of the vehicle’s path, the vehicle needs to avoid the obstacle; Alternatively, when an obstacle is within a certain distance ahead of the vehicle's route, the vehicle needs to avoid the obstacle.
[0012] Preferably, when it is determined, based on the results of the obstacle avoidance strategy, that there is a risk of collision between the vehicle and an obstacle and the obstacle needs to be avoided, the current area where the obstacle is located and the area where the target is located are determined based on the current position information of the obstacle and the predicted position of the obstacle, and then, based on the pre-set map and navigation information, command signals and path planning plans are generated to control the vehicle to perform specific tasks and avoid obstacles.
[0013] Preferably, generating the command signal and the path planning scheme specifically includes: According to the movement state of the obstacle and the predicted obstacle trajectory, it is judged whether the vehicle needs to avoid the obstacle, and a first command signal is generated according to the judgment result. According to the pre-set work task parameters stored, when the first command signal is received, parameter information corresponding to the current work task is extracted from the work task parameters, and the type of task that the vehicle currently needs to perform is determined according to the parameter information, and a second command signal is generated. The vehicle is guided to perform a specific task according to the second command signal, and the second command signal carries identification information of the type of task that the vehicle currently needs to perform; after the first command signal and the second command signal are received, a path planning solution is generated according to the obstacle target position information in the first command signal, the current position information of the obstacle in the obstacle image file, and the identification information carried in the second command signal.
[0014] Preferably, the path planning is generated by combining the A* algorithm with the D* algorithm, specifically including: Initial path planning, from the starting point to the target location, use the A* algorithm to generate the initial path, add the target node to the open list, calculate g(n) and h(n) for each node, and select the node with the smallest f(n) as the current node until the starting point is found, where g(n) is the actual distance, h(n) is the estimated distance, and f(n) represents the total cost; When an obstacle is detected and a new obstacle is detected during driving, the obstacle area is marked as impassable, and the cost of the affected area is recalculated using the D* algorithm to update the path; Replan the path, re-search the path from the current node, avoid new obstacles, use the heuristic function to estimate the cost, and select the path with the minimum cost; Generate control instructions based on the updated path to guide the vehicle to move along the new path so that the vehicle avoids obstacles and reaches the target.
[0015] The beneficial effects of the present invention are: (1) By combining the motion state and predicted trajectory of the obstacle, it can intelligently generate an obstacle avoidance strategy and determine whether the vehicle's driving route needs to be changed. When it is determined that avoidance is necessary, the optimal path is dynamically generated and updated through the combination of the A* algorithm and the D* algorithm to guide the vehicle to avoid the obstacle. It can not only identify obstacles, but also dynamically generate and update the optimal path, solving the problem in the prior art that vehicles cannot avoid obstacles when encountering them in the actual working environment. This method can also respond to environmental changes in real time, guide the vehicle to avoid obstacles, and ensure the stable operation of the vehicle in a complex environment. (2) By building a virtual simulation model, the real working scene is accurately simulated, and obstacle information is collected and analyzed in real time to generate obstacle image files. By using the physical engine and different algorithms, the future trajectory of the obstacle can be predicted and the collision risk with the vehicle can be evaluated. This method significantly improves the prediction accuracy of the obstacle's motion state, speeds up the vehicle's response to emergencies, and effectively solves the problems of the traditional obstacle avoidance system in terms of real-time response and prediction accuracy. (3) Through digital twin technology, real-time interaction between the physical world and the digital world is achieved, enabling the vehicle to flexibly adjust its obstacle avoidance strategy according to real-time environmental changes. In addition, it can generate command signals based on preset work task parameters to guide the vehicle to effectively avoid obstacles while performing specific tasks. This intelligent obstacle avoidance method not only improves the environmental adaptability of the vehicle, but also ensures the smooth execution of vehicle tasks in complex environments, solving the obstacle avoidance problem of autonomous driving vehicles in changing environments. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] For better understanding and implementation, the technical solution of the present application is described in detail below with reference to the accompanying drawings.
[0017] Figure 1 A schematic diagram of a process flow of a vehicle obstacle avoidance method based on digital twins provided in Example 1 of the present application; Figure 2 A schematic diagram of a flow chart of obtaining the motion state of an obstacle by a vehicle obstacle avoidance method based on digital twins provided in Example 1 of the present application; Figure 3 A schematic diagram of a flow chart of generating command signals and path planning solutions for a vehicle obstacle avoidance method based on digital twins provided in Example 1 of the present application; Figure 4 A vehicle obstacle avoidance method based on digital twins provided in Example 1 of the present application generates a path planning process diagram by combining the A* algorithm with the D* algorithm. DETAILED DESCRIPTION
[0018] In order to further explain the technical means and effects taken by the present invention to achieve the predetermined invention purpose, exemplary embodiments will be described in detail here, and examples thereof are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are only examples of methods and systems consistent with some aspects of the present application as detailed in the attached claims.
[0019] The terms used in this application are for the purpose of describing specific embodiments only and are not intended to limit this application. The singular forms of "a", "said" and "the" used in this application and the appended claims are also intended to include plural forms unless the context clearly indicates other meanings. It should also be understood that the term "and / or" used herein refers to and includes any or all possible combinations of one or more associated listed items.
[0020] The specific implementation methods, features and effects of the present invention are described in detail below in conjunction with the accompanying drawings and preferred embodiments.
[0021] Example 1 See also Figure 1-Figure 4 , this embodiment provides a vehicle obstacle avoidance method based on digital twins, which is used to solve the problem in the prior art that vehicles cannot avoid obstacles when encountering obstacles in an actual working environment.
[0022] The present invention provides a vehicle obstacle avoidance method based on digital twin, comprising the following steps: A vehicle obstacle avoidance method based on digital twins comprises the following steps: S1. Build a virtual simulation model and simulate the real working scene, and add objects such as vehicles, obstacles and construction workers in the virtual simulation model; Wherein, the virtual simulation model is a virtual simulation model generated by a physical engine; S2. Generate an obstacle image file by collecting obstacle information, and obtain the motion state of the obstacle in the obstacle image file through a virtual simulation model; S3. Calculate the trajectory information of the obstacle in the future according to the movement state of the obstacle to obtain the predicted obstacle trajectory; S4. According to the motion state of the obstacle and the predicted obstacle trajectory, the collision risk of the vehicle is evaluated and an obstacle avoidance strategy is obtained.
[0023] Furthermore, a virtual simulation model is constructed by acquiring the historical behavior data of the vehicle, and a mapping relationship between the physical world and the digital world is constructed based on the historical behavior data. The mapping relationship is used to transmit information in the real scene to the virtual simulation environment, and interaction between the virtual simulation environment and the real environment is performed.
[0024] Furthermore, the obstacle image file includes information of multiple obstacles, each obstacle corresponds to a position coordinate, including the current position of the obstacle, and obtaining the motion state of the obstacle in the obstacle image file specifically includes: S21, determining a projection point of the obstacle on the vehicle according to the position coordinates of the obstacle, and then determining a posture angle of the obstacle relative to the vehicle according to the projection point; The specific vehicle attitude angle is expressed as: The actual coordinates of the obstacle are set as ( , ), the origin of the vehicle coordinate system is ( , ), the projection point is ( , ), the projection point represents the position coordinates of the obstacle relative to the vehicle, and the horizontal attitude angle between the obstacle and the vehicle It is expressed as: , the attitude angle between the obstacle and the vehicle It is expressed as: .
[0025] S22, calculating the speed and acceleration of the obstacle according to the attitude angle and the preset algorithm, and obtaining the movement state of the obstacle in a future period of time, such as uniform movement, accelerated movement, or decelerated movement along a straight line; The specific calculation results are the movement speed and acceleration of the obstacle, which are expressed as: The velocity vector is calculated by the rate of change of the obstacle position coordinates over time, expressed as ,in is the change in position coordinates, is the time interval; The acceleration vector is calculated by the rate of change of the velocity vector over time, using the formula: .
[0026] S23. When the motion state of an obstacle is uniform motion along a straight line, a Gaussian filter is directly used to smooth the motion state of the obstacle. The specific smoothing process using the Gaussian filter is expressed as: in, is the smoothed position, is the last smoothed position, is the current location, are the weights of the filter.
[0027] S24. When the motion state of an obstacle is not uniformly moving along a straight line, a Kalman filter is used to smooth the motion state of the obstacle. Specifically, the Kalman filter is used to effectively evaluate the state and smooth it, which is expressed as: Status Assessment: ;in is the predicted state at step k, is the state transition matrix, is the control matrix, is the control input, is the process noise covariance; Forecast error covariance: ;in, The forecast error covariance matrix before time step k; Observation residuals: ,in The observation error at time step k, The observation matrix at time step k maps the state to the observation space, The observations obtained at time step k; Status Update: ,in, The Kalman gain at time step k, which is used to determine the weight of the observed value relative to the predicted value; Update the estimated error covariance: .
[0028] The new position of the obstacle is predicted, and then this prediction is updated based on the new observations to achieve smoothing.
[0029] Furthermore, the predicted obstacle trajectory is obtained, which specifically includes: When there are two or more obstacles whose motion states are all uniformly moving along a straight line, a multivariate Gaussian mixture model is used to fuse the motion states of these multiple obstacles to obtain the trajectory information of these multiple obstacles in the future. When there are two or more obstacles whose motion states are not uniform along a straight line, the particle swarm optimization algorithm is used to fuse the motion states of these multiple obstacles to obtain the trajectory information of these multiple obstacles in the future period of time; The combination of the two algorithms enables the vehicle to adapt to obstacles with different motion characteristics and provide accurate trajectory prediction.
[0030] The predicted position of the obstacle is obtained by fusion processing based on the movement state of the obstacle.
[0031] Furthermore, the particle swarm optimization algorithm is used to predict the trajectory of obstacles, which is specifically expressed as: Initialize a group of particles, each particle represents a solution to the obstacle's motion state, including position and velocity; Calculate the fitness function for each particle. The fitness function is the error between the predicted trajectory and the actual observed data. For each particle, if the current solution is better than the best solution it has found before (individual extremum), the individual extremum is updated; if the individual extremum of the particle is better than the best solution known in the group (global extremum), the global extremum is updated; Update the position and velocity of each particle according to the individual extreme value and the global extreme value, expressed as: Update the velocity of each particle: ; Update the position of each particle: ; in, is the velocity of particle i at time t, is the inertia weight, which controls the degree to which the particle maintains its current state of motion. and are the cognitive coefficient and social coefficient, which control the degree to which particles approach individual extreme values and global extreme values. and is the random number of particle i in the current iteration, usually in the range [0,1], is the individual extreme position of particle i, is the global extreme position The new position of particle i at time step t+1 is based on the updated result of the current position and velocity. The current position of particle i at time step t. In the multidimensional search space, this position can be a multidimensional vector representing the value of the particle on each decision variable.
[0032] According to the above iterative process, until the stopping condition is met, the global extreme value position represents the optimal solution found by the entire particle swarm, that is, the predicted obstacle motion trajectory. The global extreme value is output as the optimal trajectory for obstacle avoidance decision.
[0033] Furthermore, before judging whether the vehicle needs to avoid the obstacle according to the movement state of the obstacle and the predicted obstacle trajectory, the method further includes: receiving a parameter input by a user from an external device, the parameter comprising one or more of a maximum speed, a minimum speed, an acceleration, a deceleration, a shortest distance threshold, a longest distance threshold, a shortest time threshold, and a longest time threshold; Based on the movement state of the obstacle and the predicted obstacle trajectory, combined with the parameters input by the user, it is determined whether the vehicle needs to avoid the obstacle.
[0034] Furthermore, evaluating the collision risk of the vehicle specifically includes: judging whether there is a collision risk between the vehicle and the obstacle according to the motion state of the obstacle and the predicted obstacle trajectory, and if there is a collision risk, judging whether the vehicle needs to avoid the obstacle.
[0035] The specific judgment of whether the vehicle needs to avoid obstacles includes: According to the predicted obstacle trajectory, the movement state of the obstacle is analyzed, including speed, direction, acceleration, etc., to determine its movement characteristics; Calculate the time when the vehicle and the obstacle may collide, and find the time point when the two trajectories of the vehicle and the obstacle intersect. If such a time point exists, a collision may occur.
[0036] Furthermore, an obstacle avoidance strategy is obtained based on the collision risk assessment result of the vehicle, including: When it is determined that there is no risk of collision between the vehicle and the obstacle, the vehicle continues to drive along the original route; or, when it is determined that there is a risk of collision between the vehicle and the obstacle and there is no need to avoid the obstacle, it is determined whether the obstacle needs to be avoided according to the pre-stored driving rules; wherein the driving rules include: When the speed of the obstacle is less than a preset first threshold, the vehicle does not need to avoid the obstacle; or, When the speed of the obstacle is greater than a preset second threshold, the vehicle needs to avoid the obstacle; or, When the obstacle is within a certain distance ahead of the vehicle's route, the vehicle needs to avoid the obstacle; or, when the obstacle is within a distance that can be reached within a certain time ahead of the vehicle's route, the vehicle needs to avoid the obstacle.
[0037] Furthermore, when it is determined that there is a risk of collision between the vehicle and an obstacle and the obstacle needs to be avoided, the current area where the obstacle is located and the area where the target is located are determined based on the current position information of the obstacle and the predicted position of the obstacle, and then, based on the pre-set map and navigation information, command signals and path planning plans are generated to control the vehicle to perform specific tasks and avoid obstacles.
[0038] Furthermore, generating a command signal and a path planning scheme specifically includes: S31, judging whether the vehicle needs to avoid the obstacle according to the movement state of the obstacle and the predicted obstacle trajectory, and generating a first command signal according to the judgment result; S32, according to the preset work task parameters stored, when the first command signal is received, extracting parameter information corresponding to the current work task from the work task parameters, and determining the task type that the vehicle currently needs to perform according to the parameter information, and generating a second command signal; The second command signal instructs the vehicle to perform a specific task, and the second command signal carries identification information of the type of task that the vehicle currently needs to perform; S33. After receiving the first command signal and the second command signal, a path planning scheme is generated according to the predicted position information of the obstacle in the first command signal, the current position information of the obstacle in the obstacle image file, and the identification information carried in the second command signal, so as to control the vehicle to move forward or stop according to the path planning scheme.
[0039] Furthermore, the path planning is generated by combining the A* algorithm with the D* algorithm, specifically including: S41, initial path planning, from the starting point to the target location, use the A* algorithm to generate the initial path, add the target node to the open list, calculate g(n) and h(n) for each node, and select the node with the smallest f(n) as the current node until the starting point is found, where g(n) is the actual distance, h(n) is the estimated distance, and f(n) represents the total cost; S42, an obstacle is detected. If a new obstacle is detected during driving, the obstacle area is marked as impassable, the cost of the affected area is recalculated using the D* algorithm, and the path is updated; S43, re-plan the path, re-search the path from the current node, avoid new obstacles, use the heuristic function to estimate the cost, and select the path with the minimum cost; S44, generating control instructions, generating control instructions according to the updated path, guiding the vehicle to move along the new path, and updating the path in real time by continuously monitoring environmental changes, so that the vehicle avoids obstacles and reaches the target.
[0040] It should be explained that D* is Dynamic A*. The D* algorithm enables the vehicle to adapt to changes in the environment by recalculating the path. The A* algorithm is used in static environments, while the D* algorithm is suitable for dynamic environments. Combining the two can achieve flexible and efficient path planning, which can not only quickly find paths in static environments, but also adapt to environmental changes in dynamic environments. For each node, f(n)=g(n)+h(n) is calculated.
[0041] The A* algorithm uses its high efficiency to quickly search for the optimal path from the starting point to the target in a static environment. When the D* algorithm detects environmental changes, such as new obstacles, it can quickly recalculate the cost of the affected area and update the path. This capability enables the path planning system to adapt to dynamically changing environments and avoid hitting new obstacles.
[0042] During driving, by constantly monitoring environmental changes, once an obstacle is detected, the D* algorithm is immediately activated and the path is updated in real time. This real-time performance ensures that the vehicle can respond quickly to emergencies and ensures driving safety. The D* algorithm can not only handle one-time path planning, but also continuously optimize the path. Even in the case of multiple environmental changes, it can continue to provide optimal or suboptimal paths. The specific path planning method that combines the A* algorithm and the D* algorithm provides a fast and reliable path search solution. When there is no environmental change, the A algorithm ensures the efficiency of path planning; when the environment changes dynamically, the D algorithm ensures the adaptability and reliability of path planning.
[0043] Since the D* algorithm only recalculates the path of the affected area instead of the entire map, it greatly reduces the consumption of computing resources and improves computing efficiency compared to global replanning. It also achieves efficient, flexible and reliable path search and update in static and dynamic environments.
[0044] This embodiment simulates the real working environment by constructing a virtual simulation model, and adds objects such as vehicles, obstacles and construction workers to it. The virtual simulation model is used to generate obstacle image files and analyze their motion states. Then, the future trajectories of obstacles in different motion states are predicted by multivariable Gaussian mixture models and particle swarm optimization algorithms, and collision risk assessment and obstacle avoidance strategies are generated based on these predictions. The mapping relationship between the physical and digital worlds is further constructed through historical behavior data to achieve interaction between virtual and reality. In addition, Gaussian filters and Kalman filters are used to smooth the motion state of obstacles, and multivariable Gaussian mixture models or particle swarm optimization algorithms may be used to process the trajectory prediction of multiple obstacles. The particle swarm optimization algorithm updates the particle position and velocity iteratively to find the optimal obstacle avoidance trajectory.
[0045] The implementation of this method significantly improves the efficiency and safety of vehicle obstacle avoidance. By accurately sensing and predicting the movement of obstacles, it can respond to environmental changes in real time and generate an adaptable obstacle avoidance strategy. Combining the path planning methods of algorithm A and algorithm D, the system can not only quickly find a path in a static environment, but also quickly adapt to the appearance and movement of obstacles in a dynamic environment, ensuring real-time updating and optimization of the path. This method effectively minimizes the risk of collision, while ensuring the safety of the vehicle when performing specific tasks, improving autonomous navigation capabilities, and providing strong technical support for the practical application of autonomous driving vehicles.
[0046] Example 2 This embodiment is different from the first embodiment in that this embodiment further evaluates the collision risk of the vehicle by calculating the risk factor of the obstacle and judging whether the vehicle needs to avoid it according to the risk factor. The specific process includes: Collect and monitor the current speed, acceleration, braking deceleration, etc. of the vehicle, and then collect the current speed, position and distance of the obstacle to predict the time point when the vehicle and the obstacle may collide. This is achieved by analyzing the relative motion equation of the vehicle and the obstacle. Calculate the relative speed of the vehicle and the obstacle: , is the current speed of the vehicle, is the current speed of the obstacle; Estimated time to collision point , when the obstacle is stationary, ,in is the distance between the obstacle and the vehicle; When the obstacle is moving, the vehicle or obstacle has acceleration. Considering the speed and direction of the obstacle, it is specifically expressed as: , is the vehicle's acceleration, is the acceleration of the obstacle, t is the time, is the distance between the vehicle and the obstacle; Determine the safe distance between the vehicle and the obstacle based on the vehicle's braking distance and reaction time; calculate the theoretical safe distance based on the vehicle's braking distance: ,in is the braking deceleration.
[0047] A risk factor formula is defined, taking into account the relative speed, distance, predicted collision time and obstacle size of the vehicle and the obstacle, which is specifically expressed as: in, is the maximum safe speed or speed limit for the vehicle, is the size of the obstacle, and a, b, c, and d are weight coefficients used to adjust the importance of different factors in the calculation of the hazard factor.
[0048] According to the above formula, the risk factor is calculated. The higher the relative speed, the closer the distance to the collision point, the shorter the collision time, and the larger the obstacle size, the higher the risk factor. is the relative speed term, which indicates the ratio of the relative speed between the vehicle and the obstacle to the maximum safe speed. The higher the ratio, the faster the relative speed and the greater the risk of collision. The distance difference item indicates the difference between the actual distance between the vehicle and the obstacle and the safe distance. If the actual distance is less than the safe distance, this value is a positive number, indicating a collision risk. It is the collision time term, which represents the inverse of the collision time. The shorter the collision time, the faster the collision occurs and the higher the risk factor. It is the obstacle size item, which indicates the size of the obstacle. The larger the size, the greater the impact or damage if a collision occurs.
[0049] Develop avoidance strategies based on the threshold of the risk factor.
[0050] In summary, this embodiment can achieve accurate obstacle avoidance decisions by calculating the risk coefficient of the obstacle and judging whether the vehicle needs to avoid it based on the coefficient. By comprehensively considering factors such as the relative speed, distance, predicted collision time and size of the obstacle between the vehicle and the obstacle, the system can take avoidance measures in time when the risk coefficient reaches a certain threshold, thereby effectively preventing collision accidents and improving the safety and intelligence level of vehicle driving.
[0051] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Although the present invention has been disclosed as a preferred embodiment as above, it is not used to limit the present invention. Any technical personnel in this field can make some changes or modify the technical contents disclosed above into equivalent embodiments without departing from the scope of the technical solution of the present invention. However, any brief modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention are still within the scope of the technical solution of the present invention.
Claims
1. A vehicle obstacle avoidance method based on digital twins, characterized by: The steps include: Constructing a virtual simulation model and simulating a real working scene, wherein the virtual simulation model is generated by a physical engine; An obstacle image file is generated by collecting obstacle information, and a motion state of the obstacle in the obstacle image file is obtained by using the virtual simulation model; According to the motion state of the obstacle, the trajectory information of the obstacle in the future period is calculated to obtain the predicted obstacle trajectory; According to the motion state of the obstacle and the predicted obstacle trajectory, the collision risk of the vehicle is evaluated and the obstacle avoidance strategy is obtained; in, The assessing of the collision risk of the vehicle specifically includes: judging whether there is a collision risk between the vehicle and the obstacle according to the motion state of the obstacle and the predicted obstacle trajectory, and if there is a collision risk, judging whether the vehicle needs to avoid the obstacle; The obstacle avoidance strategy specifically includes: When it is determined that there is no risk of collision between the vehicle and the obstacle, the original route is maintained and the vehicle continues to travel; or, when it is determined that there is a risk of collision between the vehicle and the obstacle and there is no need to avoid the obstacle, whether the obstacle needs to be avoided is determined according to the pre-stored driving rules; wherein the driving rules include: When the speed of the obstacle is less than a preset first threshold, the vehicle does not need to avoid the obstacle; Alternatively, when the speed of the obstacle is greater than a preset second threshold, the vehicle needs to avoid the obstacle; Alternatively, when an obstacle is within a certain distance ahead of the vehicle’s path, the vehicle needs to avoid the obstacle; Alternatively, when an obstacle is within a certain distance ahead of the vehicle's route, the vehicle needs to avoid the obstacle.
2. A vehicle obstacle avoidance method based on digital twin according to claim 1, characterized in that: The virtual simulation model is constructed by acquiring the historical behavior data of the vehicle, constructing a mapping relationship between the physical world and the digital world according to the historical behavior data, and using the mapping relationship to transmit information in the real scene to the virtual simulation environment, so as to interact the virtual simulation environment with the real environment.
3. The vehicle obstacle avoidance method based on digital twin according to claim 1, characterized in that: The obstacle image file includes information of multiple obstacles, each obstacle corresponds to a position coordinate, and obtaining the motion state of the obstacle in the obstacle image file specifically includes: Determining a projection point of the obstacle on the vehicle according to the position coordinates of the obstacle, and then determining an attitude angle of the obstacle relative to the vehicle according to the projection point; The movement speed and acceleration of the obstacle are calculated according to the attitude angle and a preset algorithm, and the movement state of the obstacle in a future period of time is obtained.
4. The vehicle obstacle avoidance method based on digital twin according to claim 3 is characterized in that: When the motion state of an obstacle is uniform motion along a straight line, a Gaussian filter is used to smooth the motion state of the obstacle; When the motion state of an obstacle is not uniform along a straight line, the Kalman filter is used to smooth the motion state of the obstacle.
5. The vehicle obstacle avoidance method based on digital twin according to claim 1, characterized in that: The predicted obstacle trajectory specifically includes: When there are two or more obstacles that are moving at a uniform speed along a straight line, a multivariate Gaussian mixture model is used to fuse the motion states of multiple obstacles to obtain the trajectory information of multiple obstacles in the future. When there are two or more obstacles whose motion states are not uniform along a straight line, the particle swarm optimization algorithm is used to fuse the motion states of multiple obstacles to obtain the trajectory information of multiple obstacles in the future. The predicted position of the obstacle is obtained by fusion processing based on the movement state of the obstacle.
6. The vehicle obstacle avoidance method based on digital twin according to claim 1, characterized in that: When it is determined that an obstacle needs to be avoided based on the results of the obstacle avoidance strategy, a command signal and a path planning plan are generated.
7. The vehicle obstacle avoidance method based on digital twin according to claim 6 is characterized in that: According to the results of the obstacle avoidance strategy, when it is determined that there is a risk of collision between the vehicle and an obstacle and the obstacle needs to be avoided, the current area where the obstacle is located and the area where the target is located are determined based on the current position information of the obstacle and the predicted position of the obstacle. Then, based on the pre-set map and navigation information, command signals and path planning plans are generated to control the vehicle to perform specific tasks and avoid obstacles.
8. The vehicle obstacle avoidance method based on digital twin according to claim 7 is characterized in that: The generating of the command signal and the path planning scheme comprises: Based on the movement state of the obstacle and the predicted obstacle trajectory, determine whether the vehicle needs to avoid the obstacle, generate a first command signal based on the judgment result, and based on the pre-set work task parameters stored, when the first command signal is received, extract parameter information corresponding to the current work task from the work task parameters, and determine the type of task that the vehicle currently needs to perform based on the parameter information, and generate a second command signal.
9. The vehicle obstacle avoidance method based on digital twin according to claim 7, characterized in that: The generating of the command signal and the path planning scheme also includes: guiding the vehicle to perform a specific task according to a second command signal, the second command signal carrying identification information of the type of task that the vehicle currently needs to perform; after receiving the first command signal and the second command signal, generating a path planning scheme according to the obstacle target position information in the first command signal, the current position information of the obstacle in the obstacle image file and the identification information carried in the second command signal.
10. The vehicle obstacle avoidance method based on digital twin according to claim 9, characterized in that: The generated path planning is generated by combining the A* algorithm with the D* algorithm, specifically including: Initial path planning, from the starting point to the target location, use the A* algorithm to generate the initial path, add the target node to the open list, calculate g(n) and h(n) for each node, and select the node with the smallest f(n) as the current node until the starting point is found, where g(n) is the actual distance, h(n) is the estimated distance, and f(n) represents the total cost; When an obstacle is detected and a new obstacle is detected during driving, the obstacle area is marked as impassable, and the cost of the affected area is recalculated using the D* algorithm to update the path; Replan the path, re-search the path from the current node, avoid new obstacles, use the heuristic function to estimate the cost, and select the path with the minimum cost; Generate control instructions based on the updated path to guide the vehicle to move along the new path so that the vehicle avoids obstacles and reaches the target.
Citation Information
Patent Citations
Unmanned transport ship
CN109795629A
Interactive simulation test system of automatic driving algorithm
CN112417756A
V2X-based intersection passage anti-collision device and anti-collision method
CN113232649A
Vehicle obstacle avoidance method and system, computer readable storage medium and electronic equipment
CN115454082A
Unmanned aerial vehicle control platform capable of three-dimensional holographic inspection
CN116520890A
Cited By
Mining area vehicle path planning method and device based on digital twinborn map
CN119290001A
AMR path planning method and device based on digital twinning and storage medium
CN119311012A
Multi-vertical fixed-wing unmanned aerial vehicle flight path planning and intelligent obstacle avoidance method
CN119690112A
Multi-heterogeneous robot underground space detection autonomous task planning method under dynamic networking
CN120122661A
Autonomous Task Planning Method for Underground Space Exploration of Multi-Heterogeneous Robots under Dynamic Networking
CN120122661B
