Dynamic obstacle avoidance method and system based on global algorithm and local improved potential field method

Abstract

The invention discloses a dynamic obstacle avoidance method and system based on a global algorithm and a local improved potential field method, and relates to the technical field of dynamic obstacle avoidance and path planning algorithm optimization, and the method comprises the steps: carrying out the path planning based on an A * algorithm; obstacle avoidance is carried out according to a local improved potential field method; and switching a global algorithm mechanism and a local algorithm mechanism to perform cooperative work according to judgment. According to the dynamic obstacle avoidance method based on the global algorithm and the local improved potential field method, the A * algorithm is used for executing a global path planning task, and the shortest path from the starting point to the target point is accurately positioned by comprehensively considering the actual cost of the path and the estimated cost of the path to the target point. According to the method, the global path optimization advantage of the A * algorithm and the local flexible planning capability of the dynamic window method are ingeniously fused, perfect connection of a global view and local fine adjustment is achieved, and the overall efficiency and the instant response speed of path planning are remarkably improved. And path planning and dynamic obstacle avoidance can be effectively carried out.

Description

Technical Field

[0001] The present invention relates to the technical field of dynamic obstacle avoidance and path planning algorithm optimization, and in particular to a dynamic obstacle avoidance method and system based on a global algorithm and a local improved potential field method. Background Art

[0002] With the rapid development of robotics, dynamic obstacle avoidance technology has become a core research direction in the fields of automated inspection, logistics and transportation, and service robots. At present, dynamic obstacle avoidance methods usually combine global path planning and local obstacle avoidance strategies to improve the robot's navigation ability in complex dynamic environments. Among them, the A* algorithm is widely used in global path planning due to its high path search efficiency and easy implementation, while the potential field method is valued in dynamic environments due to its real-time and local obstacle avoidance capabilities. In recent years, in order to cope with changing environments and real-time requirements, researchers have tried to improve the efficiency and safety of dynamic obstacle avoidance by improving traditional algorithms or introducing intelligent optimization mechanisms, further promoting the application and development of related technologies.

[0003] Although the above methods have achieved certain results, the existing technologies still face many challenges and shortcomings in practical applications. First, the paths generated by the traditional A* algorithm in global path planning may be difficult to cope with the rapidly changing obstacle distribution in a dynamic environment, resulting in path failure or low planning efficiency; secondly, although the classical potential field method is suitable for local obstacle avoidance, there is a problem of local optimal solution, which may cause the robot to fall into a deadlock area or a loop trajectory, making it difficult to achieve the global goal. In addition, the existing technologies generally lack an effective global and local algorithm coordination mechanism. When the robot faces densely distributed obstacles or dynamic obstacles during the execution of tasks, the algorithm switching often lacks real-time and robustness, making it difficult to ensure the obstacle avoidance effect. On the other hand, many existing methods are highly dependent on environmental data and lack the versatility to adapt to different scenarios. Especially in the case of high-speed movement of dynamic obstacles or complex scenarios, it is difficult to achieve efficient path adjustment and obstacle avoidance. Summary of the invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problems solved by the present invention are: the existing dynamic obstacle avoidance and path planning methods have the following problems: the global path planning cannot adapt to the dynamic environment, the local obstacle avoidance algorithm is prone to fall into the local optimum, the coordination between the global and local algorithms lacks real-time and robustness, and how to achieve efficient obstacle avoidance and path optimization in a dynamic environment.

[0006] To solve the above technical problems, the present invention provides the following technical solutions: a dynamic obstacle avoidance method based on a global algorithm and a local improved potential field method, including path planning based on an A* algorithm; obstacle avoidance based on a local improved potential field method; and switching between global and local algorithm mechanisms based on judgment to work in coordination.

[0007] As a preferred solution of the dynamic obstacle avoidance method based on the global algorithm and the local improved potential field method described in the present invention, wherein: the path planning based on the A* algorithm includes abstracting the working scene of the inspection robot into a grid map model, and setting the fineness and overall coverage of the grid map according to the on-site information;

[0008] Mark the location and outline shape of static obstacles on the grid map;

[0009] The inspection robot operation scene is modeled as a grid map M, where each grid represents an area;

[0010] The value of M[i][j] is used to distinguish the passable area and the obstacle area and is expressed as:

[0011]

[0012] Among them, M[i][j] represents the distinction between the passable area and the obstacle area;

[0013] Based on the on-site environmental data, the outlines and positions of static obstacles are marked. The degree of detail is determined by the resolution of the grid, which needs to balance accuracy and computational cost.

[0014] As a preferred solution of the dynamic obstacle avoidance method based on the global algorithm and the local improved potential field method described in the present invention, wherein: the path planning based on the A* algorithm includes clarifying the starting point and destination of the inspection robot, using the A* algorithm to search for a path on the constructed grid map, and outputting the optimal route;

[0015] The path cost function is expressed as:

[0016] f(n)=g(n)+h(n)

[0017] Among them, g(n) represents the actual cost from the starting point to the current node, and h(n) represents the estimated cost from the current node to the target point;

[0018] Manhattan distance is used to calculate h(n) as:

[0019]

[0020] Among them, x n ,y n Indicates the horizontal and vertical coordinates of the current node, x t ,y t Indicates the horizontal and vertical coordinates of the target node;

[0021] Based on the minimum search of the cost function, a global optimal path P is generated from the starting point S to the target point T as follows:

[0022] P={p1,p2,…,p k}

[0023] Among them, p1, p2, …, p k Indicates a waypoint;

[0024] Visibility graph method and Dijkstra algorithm are used to optimize the path.

[0025] As a preferred solution of the dynamic obstacle avoidance method based on the global algorithm and the local improved potential field method described in the present invention, the obstacle avoidance according to the local improved potential field method includes outputting a repulsive potential field according to the relative spatial position and relative movement speed between the obstacle and the robot when the inspection robot encounters a dynamic obstacle, which is expressed as:

[0026]

[0027] in, represents the repulsive potential field, k rep represents the repulsive force constant, d(x r ,y r ,x o ,y o ) represents the distance between the robot's current position and the obstacle, d0 represents the maximum distance threshold that affects the repulsive force, represents the unit vector pointing from the robot to the obstacle, indicating the direction of the repulsive force;

[0028] The strength of the repulsive potential field changes dynamically as the distance between the obstacle and the robot changes;

[0029] When the distance of the obstacle is inversely proportional to the strength of the repulsive force;

[0030] The strength of the gravitational potential field is inversely proportional to the distance from the robot's current position to the target point.

[0031] As a preferred solution of the dynamic obstacle avoidance method based on the global algorithm and the local improved potential field method described in the present invention, wherein: the obstacle avoidance based on the local improved potential field method includes adopting a strategy of adaptive sub-target point setting when the inspection robot encounters a local optimal solution problem during the path planning process, resulting in the inability to move directly toward the final target;

[0032] The obstacle information around the robot is collected through sensors, and the obstacle set o is established.

[0033] {o1,o2,…,on}, the position of each obstacle i is (x oi ,y oi );

[0034] The obstacle area is defined as the spatial range that the robot cannot pass through. The distance function d(R,0i) is used to represent the shortest distance between the robot and the obstacle 0i. According to the current position of the robot and the distribution of surrounding obstacles, a candidate set of sub-target points is generated.

[0035] Ensure that the candidate sub-target point is no more than the specified range d from the machine R max , to ensure the rationality of the sub-target points:

[0036] d(R,T i )≤d max

[0037] Where, d(R,T i ) represents the robot and the obstacle T i The closest distance, d max Indicates the scope of regulations;

[0038] Ensure that the sub-target point T i Not inside the obstacle area:

[0039]

[0040] Among them, r safe represents the safety distance threshold, d(T i ,O j ) indicates obstacle T i With obstacles O j The closest distance, O j Indicates an obstacle;

[0041] The candidate points are limited by the direction vector angle θ and are expressed as:

[0042]

[0043] in, represents the robot motion direction vector, Represents the robot's target direction vector.

[0044] As a preferred solution of the dynamic obstacle avoidance method based on the global algorithm and the local improved potential field method described in the present invention, wherein: the obstacle avoidance based on the local improved potential field method includes applying a fuzzy control mechanism to output a danger assessment index according to the relative distance, relative speed and relative direction between the robot and the obstacle;

[0045] Fuzzy control adjusts the parameters of the repulsive potential field and the gravitational potential field according to the changes in the risk assessment index;

[0046] The input variable is relative distance, and the output variable is repulsion adjustment coefficient;

[0047] Fuzzify input and output variables, define language variables and membership functions, fuzzify input variables, and fuzzify output variables;

[0048] The membership function is expressed as:

[0049]

[0050] Where d represents the relative distance, d near_max Indicates the maximum relative distance, d near_min Represents the minimum relative distance, μ Near (d) represents the membership function;

[0051] Establish a fuzzy rule base based on experience or scenario analysis;

[0052] Based on fuzzy rules, the fuzzy reasoning method is used to calculate the output membership:

[0053] Output the premise membership of each rule, determine the applicability of the rule based on the premise membership, integrate all the rules, and output the fuzzy set.

[0054] As a preferred solution of the dynamic obstacle avoidance method based on the global algorithm and the local improved potential field method described in the present invention, wherein: the switching of the global and local algorithm mechanisms for collaborative work based on the judgment includes the system automatically switching from the global A* algorithm to the local improved potential field method due to the obstacle situation;

[0055] Based on real-time environmental perception and analysis, determine whether it is necessary to switch from the global A* algorithm to the local improved potential field method, or vice versa;

[0056] When the obstacle density exceeds the set threshold, it indicates that the robot is in an obstacle-dense area and needs to switch to the local improved potential field method;

[0057] When the shortest distance from the robot to the obstacle is less than the safety distance, it triggers the switch to the local improved potential field method;

[0058] If the obstacle is a dynamic obstacle and its relative speed exceeds the threshold, it is also necessary to switch to the local improved potential field method;

[0059] If the robot is close to the target area and the remaining path length of the global A* plan is less than the threshold, switch to the local improved potential field method to complete the final adjustment;

[0060] The comprehensive state index S is defined to quantify the complexity of the robot's current environment, expressed as:

[0061]

[0062] Among them, α1, α2, α3, α4 represent weight parameters, D obsrepresents the density of obstacles, d min represents the shortest distance from the robot to the obstacle, v obs represents the obstacle speed, P progress Indicates the stage completion degree of path planning;

[0063] When S>St, switch to the local improved potential field method;

[0064] When S≤St and obstacles are sparse, switch back to the global A* algorithm.

[0065] Another object of the present invention is to provide a dynamic obstacle avoidance system based on a global algorithm and a local improved potential field method, which can avoid obstacles through the local improved potential field method, solving the problem that the current dynamic obstacle avoidance and path planning containing local obstacle avoidance algorithms are prone to fall into local optimality.

[0066] As a preferred solution of the dynamic obstacle avoidance system based on global algorithm and local improved potential field method described in the present invention, it includes: a path planning module, a field method obstacle avoidance module, and a collaborative work module; the path planning module is used to adopt the global A* algorithm path planning; the field method obstacle avoidance module is used to avoid obstacles based on the local improved potential field method; the collaborative work module is used to work collaboratively according to the global and local algorithms.

[0067] A computer device comprises a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of a dynamic obstacle avoidance method based on a global algorithm and a local improved potential field method.

[0068] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of a dynamic obstacle avoidance method based on a global algorithm and a local improved potential field method.

[0069] Beneficial effects of the present invention: The dynamic obstacle avoidance method based on the global algorithm and the local improved potential field method provided by the present invention uses the A* algorithm to perform the global path planning task, and by comprehensively considering the actual cost of the path and the estimated cost to the target point, accurately locates the shortest path from the starting point to the target point, thereby ensuring the optimization of the global path planning. The A* algorithm shows excellent adaptability and flexibly copes with map environments of various sizes and complexities. Whether facing an open area without obstacles or a complex indoor or outdoor space, the A* algorithm can efficiently complete the path planning work, showing its wide applicability and powerful processing capabilities. At the local path planning level, an optimized artificial potential field method is adopted. This method cleverly constructs a gravitational field and a repulsive field as a driving force to guide the robot to move to the target point, ensuring that it can effectively avoid obstacles. Through this improvement, the problem that the traditional artificial potential field method is prone to fall into a local minimum is successfully overcome. When the robot accidentally falls into a local optimal solution, the improved method intelligently guides the robot to escape from the current position and go to the preset sub-target point, thereby restarting the path planning process until it successfully reaches the destination. In dealing with dynamic obstacles, the fuzzy control concept is incorporated, and a risk assessment mechanism for moving obstacles is designed, based on which the real-time path planning strategy is dynamically adjusted. This adjustment mechanism gives the robot the ability to respond quickly when encountering dynamic obstacles, ensuring that it can avoid collision risks in real time and effectively prevent planning failures caused by obstacle movement. This method cleverly combines the global path optimization advantages of the A* algorithm with the local flexible planning capabilities of the dynamic window method, achieving a perfect connection between global vision and local fine adjustment, and significantly improving the overall efficiency and instant response speed of path planning. The optimized artificial potential field method has the characteristics of fast calculation and efficient execution, and can determine the local optimal route in a very short time, improving the efficiency of path planning. This method can handle map environments of different sizes and complexities, whether it is a simple open environment or a complex indoor or outdoor scene, and can effectively perform path planning and dynamic obstacle avoidance. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0071] Figure 1 An overall flow chart of a dynamic obstacle avoidance method based on a global algorithm and a local improved potential field method is provided for the first embodiment of the present invention.

[0072] Figure 2A global A* algorithm path planning diagram of a dynamic obstacle avoidance method based on a global algorithm and a local improved potential field method is provided for the first embodiment of the present invention.

[0073] Figure 3 A flow chart of inspection robot path planning for a dynamic obstacle avoidance method based on a global algorithm and a local improved potential field method is provided for the first embodiment of the present invention.

[0074] Figure 4 A local improved potential field method obstacle avoidance diagram of a dynamic obstacle avoidance method based on a global algorithm and a local improved potential field method is provided as a first embodiment of the present invention.

[0075] Figure 5 A first embodiment of the present invention provides a dynamic obstacle avoidance method based on a global algorithm and a local improved potential field method, and a path planning diagram for an inspection robot based on a global A* algorithm and a local improved potential field method.

[0076] Figure 6 An overall flow chart of a dynamic obstacle avoidance system based on a global algorithm and a local improved potential field method is provided for the third embodiment of the present invention. DETAILED DESCRIPTION

[0077] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0078] Example 1, reference Figure 1-Figure 5 , is an embodiment of the present invention, and provides a dynamic obstacle avoidance method based on a global algorithm and a local improved potential field method, comprising:

[0079] S1: Path planning based on A* algorithm.

[0080] Furthermore, the inspection robot's operation scene is abstracted into a grid map model, in which each grid unit represents a specific area, which is clearly divided into two categories: passable and impassable. In order to accurately reflect the characteristics of the actual environment, it is necessary to set the fineness (i.e., resolution) and overall coverage (i.e., size) of the grid map based on the specific information on the site.

[0081] At the same time, the location and outline shape of static obstacles should be accurately marked on the grid map. These obstacles include walls, fixed equipment or obstacles, etc., which constitute physical limitations in the robot's movement. By accurately reflecting the location and shape of these obstacles on the grid map, it is possible to clearly identify which areas the robot cannot cross, thus providing strong support for subsequent path planning.

[0082] Grid modeling: The inspection robot operation scene is modeled as a grid map M, where each grid represents an area;

[0083] The value of M[i][j] is used to distinguish the passable area and the obstacle area and is expressed as:

[0084]

[0085] Among them, M[i][j] represents the distinction between the passable area and the obstacle area;

[0086] Static obstacle marking: Based on the on-site environmental data, the outline and position of static obstacles are clearly marked. The degree of detail is determined by the resolution of the grid, which needs to balance accuracy and computational cost.

[0087] It should be noted that the starting point and destination of the inspection robot must be clearly defined first. The A* algorithm is used to search for paths on the constructed grid map to find the optimal route from the starting point to the destination. The A* algorithm is a heuristic search method that gradually explores the possible path space. Each time, it estimates the shortest path cost to the destination based on the current state and selects the next moving direction accordingly. This process continues until the algorithm successfully finds a path to the destination or determines that the destination cannot be reached under the current environment.

[0088] (1) Cost function design, the path cost function is defined as:

[0089] f(n)=g(n)+h(n)

[0090] Among them, g(n) represents the actual cost from the starting point to the current node, and h(n) represents the estimated cost from the current node to the target point;

[0091] Manhattan distance is used to calculate h(n) as:

[0092]

[0093] Among them, x n ,y n Indicates the horizontal and vertical coordinates of the current node, x t ,y t Indicates the horizontal and vertical coordinates of the target node;

[0094] Based on the minimum search of the cost function, a global optimal path P is generated from the starting point S to the target point T as follows:

[0095] P={p1,p2,…,p k}

[0096] Among them, p1, p2, …, p k Indicates a waypoint;

[0097] Visibility graph method and Dijkstra algorithm are used to optimize the path.

[0098] Furthermore, the path given by the A* algorithm was carefully polished, and the path was further streamlined and optimized using advanced path optimization methods, such as the visibility graph method and the Dijkstra algorithm. The purpose is to allow the inspection robot to reach the target location more smoothly with fewer turns and shorter path length when performing inspection tasks.

[0099] It should be noted that the visibility graph method is a classic method for optimizing paths in irregular environments, often used to eliminate unnecessary turns and redundant paths. The core idea is to use the visible obstacle boundaries in the environment to build a graph to generate a more direct path. The following is the optimization process:

[0100] 1. Environmental modeling: In the grid map, by determining the locations of the inspection robot’s starting and target points and the locations of all obstacles, a visibility graph is calculated, that is, whether the robot’s “line of sight” from one point to another is blocked by obstacles.

[0101] 2. Build a visible graph: For each point pi in the environment, check whether there are obstacles blocking it from all other points. If there are no obstacles, connect these points and form an edge, indicating that there is a feasible direct path between these points.

[0102] 3. Path optimization: In the visible graph, use the vertices and edges in the graph to build a new path that will avoid detouring around obstacles and minimize the path length and number of turns. Optimize the path in the visible graph using the shortest path algorithm to generate a smoother and more direct path.

[0103] Dijkstra algorithm optimization: Dijkstra algorithm is a classic shortest path algorithm, which is applicable to the case where the weights of all edges in the graph are non-negative. In path optimization, Dijkstra algorithm is used to further streamline and optimize the path generated by A* algorithm, with the goal of reducing path length and avoiding redundant paths. The following is the optimization process:

[0104] 1. Path optimization: Use the Dijkstra algorithm to optimize the preliminary path generated by the A* algorithm, calculate the shortest path between each pair of adjacent points, and merge redundant points in the path. For example, if there are multiple consecutive points in the path, you can connect these points with straight lines to reduce the length of the path.

[0105] 2. Local path optimization: The Dijkstra algorithm can not only optimize the overall path length, but also play a role in local path optimization. For corners or obstacle-dense areas in some complex environments, the Dijkstra algorithm can avoid unnecessary turns and U-turns, reducing the robot's movement time.

[0106] 3. Streamlining the path: For redundant nodes in the A* path, the Dijkstra algorithm prioritizes the shortest path segment, simplifies the lengthy path segment, and ultimately generates a shorter and more direct path.

[0107] Combine the two for global and local optimization: Combine the visibility graph method and Dijkstra algorithm to further improve the accuracy and smoothness of the path.

[0108] 1. Global optimization: After initially planning the path using the A* algorithm, the visibility graph method is used to delete unnecessary turns in the path, and the Dijkstra algorithm is used to optimize the key nodes on the path. The final path can avoid unnecessary complexity while ensuring the shortest path.

[0109] 2. Local optimization: During the actual driving process of the robot, as the environmental information is updated (such as dynamic obstacles), the local path optimization strategy is applied. The Dijkstra algorithm can combine real-time sensor data to further avoid obstacles and maintain the smoothness of the path by re-optimizing the local path.

[0110] S2: Obstacle avoidance based on the local improved potential field method.

[0111] Furthermore, the repulsive potential field calculation: When the inspection robot encounters a dynamic obstacle, a force field called the repulsive potential field is calculated based on the relative spatial position and relative movement speed between the obstacle and the robot. The strength and direction of the repulsive potential field will be dynamically adjusted as the distance between the obstacle and the robot increases or decreases and the relative speed changes. The main function of the repulsive potential field is to provide the robot with an obstacle avoidance guide to ensure that the robot can flexibly bypass the obstacle and continue to move forward safely. It can be expressed as:

[0112]

[0113] in, represents the repulsive potential field, k rep represents the repulsive force constant, d(xr ,y r ,x o ,y o ) represents the distance between the robot's current position and the obstacle, d0 represents the maximum distance threshold that affects the repulsive force, represents the unit vector pointing from the robot to the obstacle, indicating the direction of the repulsive force;

[0114] The strength of the repulsive potential field changes dynamically with the distance between the obstacle and the robot. When the obstacle is far away, the repulsive force is weak; when the obstacle is close, the repulsive force increases. In addition, the influence of relative speed causes the repulsive force to increase rapidly when the obstacle approaches quickly, forcing the robot to quickly adjust its direction of movement to avoid the obstacle. The goal of the repulsive potential field is to maintain a safe distance between the robot and the obstacle and to avoid collisions between the robot and the obstacle. In this process, the robot will constantly adjust its path to avoid obstacles while keeping the travel route as short as possible.

[0115] Gravitational potential field calculation: The target point will release a force field called the gravitational potential field, which is mainly used to guide the inspection robot to move continuously toward the target point. The strength of the gravitational potential field is inversely proportional to the distance from the robot's current position to the target point. As the robot gradually approaches the target point, the gravitational potential field it feels will gradually weaken, but it will always maintain the force pulling toward the target point to ensure that the robot can continue to move in the right direction.

[0116] Adaptive sub-target point setting: When the inspection robot encounters a local optimal solution (i.e., local minimum) problem during path planning, which prevents it from moving directly toward the final goal, it adopts an adaptive sub-target point setting strategy. This strategy allows the robot to intelligently select a temporary sub-target point based on its current location, the layout of surrounding obstacles, and the orientation of the final goal point. By moving to this sub-target point first, the robot cleverly avoids the obstacles in front of it and replans a path to the final goal at the new location.

[0117] The obstacle information around the robot is collected through sensors, and an obstacle set O = {O1, O2, ..., On} is established. The position of each obstacle Oi is (xoi, yoi). The obstacle area is defined as the spatial range that the robot cannot pass through. The distance function d(R, Oi) is used to represent the closest distance between the robot and the obstacle Oi. Based on the current position of the robot and the distribution of surrounding obstacles, a sub-target point candidate set Tcandidates is generated. The generation rules of candidate sub-target points are as follows:

[0118] It should be noted that the distance constraint ensures that the candidate sub-target point is no more than the specified range d from the machine R. max , to ensure the rationality of the sub-target points:

[0119] d(R,T i )≤d max

[0120] Where, d(R,T i ) represents the robot and the obstacle T i The closest distance, d max Indicates the scope of regulations;

[0121] Accessibility check: Ensure that the sub-target point T i Not inside the obstacle area:

[0122]

[0123] Among them, r safe represents the safety distance threshold, d(T i ,O j ) indicates obstacle T i With obstacles O j The closest distance, O j Indicates an obstacle;

[0124] Directional constraint: The candidate points are restricted by the direction vector angle θ and are expressed as:

[0125]

[0126] in, represents the robot motion direction vector, Represents the robot's target direction vector.

[0127] Furthermore, the fuzzy control strategy:

[0128] In order to cope with the challenges brought by moving obstacles, a fuzzy control mechanism is designed, which first calculates a dynamic danger assessment index based on key information such as the relative distance, relative speed and relative direction between the robot and the obstacle. This index can reflect the degree of danger faced by the robot in real time.

[0129] Then, the fuzzy controller will adjust the parameters of the repulsive potential field and the gravitational potential field according to the changes in the risk assessment index. The adjustment of the repulsive potential field is intended to allow the robot to more keenly perceive and avoid obstacles, while the adjustment of the gravitational potential field ensures that the robot can avoid obstacles while maintaining the trend of moving toward the target point. Through such fine regulation, the robot can achieve smoother and safer obstacle avoidance operations.

[0130] The basic framework of fuzzy control strategy:

[0131] The core of the fuzzy control strategy is to design a fuzzy controller to integrate the information such as the relative distance, relative speed and relative direction between the robot and the obstacle, generate appropriate control output, and thus dynamically adjust the potential field parameters.

[0132] Input variables: Relative distance, the Euclidean distance between the robot and the obstacle, measures the proximity of the obstacle. Relative speed, the relative speed between the robot and the obstacle, indicates the obstacle approach rate. Relative direction: the angle between the robot's movement direction and the obstacle's direction, reflects the threatening direction of the obstacle.

[0133] Output variables: repulsion adjustment coefficient, which controls the parameters of the repulsion potential field strength and enhances the robot's ability to avoid obstacles. Gravity adjustment coefficient, which controls the parameters of the gravitational potential field strength and ensures that the robot remains goal-oriented while avoiding obstacles.

[0134] Design of fuzzy controller: fuzzify input and output variables, define linguistic variables and their membership functions

[0135] Fuzzification of input variables: Relative distance d: Near, Medium, Far. Relative speed vrel: Low, Medium, High. Relative direction θ: Front, Side, Rear.

[0136] Output variable fuzzification: repulsion adjustment coefficient krep: weak, medium, strong. Attraction adjustment coefficient kat: low, medium, high.

[0137] The membership function is expressed as:

[0138]

[0139] Where d represents the relative distance, d near_max Indicates the maximum relative distance, d near_min Represents the minimum relative distance, μ Near (d) represents the membership function;

[0140] Design of fuzzy rules: Based on experience or scenario analysis, a fuzzy rule base is established. Each rule describes the output control strategy under different input combinations. These rules can be expressed by fuzzy reasoning, covering the possibility of various input combinations, such as:

[0141] Rule 1: If d is near, and vrel is high, and θ is forward, then krep is strong and katt is low.

[0142] Rule 2: If d is far, and vrel is low, and θ is sideways, then krep is weak and katt is high.

[0143] Fuzzy reasoning: Based on fuzzy rules, fuzzy reasoning method is used to calculate the output membership:

[0144] Calculate the antecedent membership of each rule

[0145] The applicability of the rule is determined based on the degree of premise membership.

[0146] Combining all the rules, the output fuzzy set is calculated.

[0147] S3: Based on the judgment, switch the global and local algorithm mechanisms to work together.

[0148] Furthermore, the smooth switching mechanism:

[0149] In order to achieve seamless conversion between the global A* algorithm and the local improved potential field method, a flexible switching mechanism is designed, which can intelligently select the applicable path planning method according to the current environmental conditions of the robot.

[0150] Specifically, when the robot approaches an obstacle or is in an area with dense obstacles, in order to enhance the flexibility and safety of obstacle avoidance, the system automatically switches from the global A* algorithm to the local improved potential field method, because the local improved potential field method shows higher agility and adaptability when dealing with close obstacles and complex environments.

[0151] On the contrary, when the robot is far away from obstacles or in an area with sparse obstacles, in order to restore the efficiency and accuracy of global path planning, the system automatically switches from the local improved potential field method back to the global A* algorithm. The global A* algorithm shows stronger global optimization ability and stability in open areas and long-distance path planning.

[0152] The design principle of the smooth switching mechanism: The core of the switching mechanism is based on real-time environmental perception and analysis to determine whether it is necessary to switch from the global A* algorithm to the local improved potential field method, or vice versa. The following are the key switching conditions:

[0153] Obstacle density (Dobs): Dobs indicates the density of obstacles per unit area around the robot. When Dob exceeds the set threshold Dcrit, it indicates that the robot is in an obstacle-dense area and needs to switch to the local improved potential field method.

[0154] The shortest distance from the robot to the obstacle (dmind): dmin = minid(R,Oi), that is, the distance between the robot and the nearest obstacle. When dmin is less than the safe distance dsafe, it triggers the switch to the local improved potential field method.

[0155] Obstacle speed (vobs): If the obstacle is a dynamic obstacle and its relative speed vobs exceeds the threshold vcrit, it means that the obstacle poses a greater threat and it is also necessary to switch to the local improved potential field method.

[0156] Path planning stage completion (Pprogress): If the robot approaches the target area and the remaining path length lrem planned by the global A* is less than the threshold lcrit, it can switch to the local improved potential field method to complete the final adjustment.

[0157] Define a comprehensive state index S to quantify the complexity of the robot's current environment, combining the above multiple parameters:

[0158]

[0159] Among them, α1, α2, α3, α4 represent weight parameters, D obs represents the density of obstacles, d min represents the shortest distance from the robot to the obstacle, v obs represents the obstacle speed, P progress Indicates the stage completion degree of path planning;

[0160] It should be noted that when S>Sthreshold, the method switches to the local improved potential field method.

[0161] When S≤Sthreshold and obstacles are sparse, switch back to the global A* algorithm.

[0162] Information fusion: In the switching stage of path planning methods, an efficient information integration strategy is adopted, which deeply integrates the overall path blueprint generated by the global A* algorithm with the real-time obstacle avoidance details provided by the local improved potential field method. This comprehensive processing of global and local information enhances the robot's dynamic response and obstacle avoidance capabilities in complex and changing environments.

[0163] Incorporating redundant design concepts into the hardware and software architecture, such as configuring additional sensors and controllers as backup, greatly improves the system's fault tolerance. If the main sensor or controller fails, the system can quickly and automatically enable the backup component to ensure that the inspection robot can continue to perform its tasks stably.

[0164] The system continuously monitors its own operating status and activates the fault response mechanism once any signs of fault are detected. This mechanism covers multiple aspects such as accurate fault location, effective isolation and rapid repair, ensuring that the inspection robot maintains its stable operation in the face of any emergency.

[0165] Example 2, an embodiment of the present invention, provides a dynamic obstacle avoidance method based on a global algorithm and a local improved potential field method. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0166] First, simulate the real inspection environment and set up a 10m x 10m field, which contains static obstacles (such as fixed pillars and walls) and dynamic obstacles (such as moving people). The obstacle distribution needs to be complex, with both open areas and areas with dense obstacles. Set up multiple fixed reference points in the field to calibrate the robot position and path measurement. Install a lidar and camera system to record the experimental process.

[0167] Experimental procedures

[0168] 1. Environment initialization: Static obstacles are arranged, dynamic obstacles are simulated by mobile devices, and their movement trajectories are pre-planned.

[0169] 2. Global path planning: The robot starts the global A* algorithm, generates an initial path, and records the length of the global path and the planning time.

[0170] 3. Dynamic obstacle avoidance test: Introduce dynamic obstacles during the robot's driving process to test the obstacle avoidance ability of the locally improved potential field method, and run the comparison algorithm (pure A* and traditional potential field method) for obstacle avoidance at the same time.

[0171] 4. Repeated testing: Change the obstacle distribution density, dynamic obstacle speed and complexity, and conduct multiple repeated tests.

[0172] Experimental data and analysis

[0173] Data sample: The following are the experimental results (average values ​​of multiple groups):

[0174] Table 1 Experimental data table

[0175]

[0176] Path length and time: The proposed method generates the shortest path length and the total driving time is significantly lower than other algorithms, indicating that the combination of global A* and local potential field method can effectively reduce redundant paths and improve planning efficiency.

[0177] Obstacle avoidance times and success rate: The proposed method has the least obstacle avoidance times and the highest success rate, which shows that the local improved potential field method is more effective in dealing with dynamic obstacles.

[0178] Smoothness: The proposed method significantly reduces the number of turns and angle changes through fuzzy control and adaptive sub-target point setting, and improves the smoothness of the robot's operation.

[0179] Computation time: The computation time of the proposed method is slightly higher than that of the traditional potential field method, but lower than that of the global A*, and has better real-time performance in complex scenarios.

[0180] Experimental Conclusion

[0181] Experimental results show that the method of combining the global A* algorithm with the local improved potential field method has the following advantages in complex dynamic environments compared with a single algorithm:

[0182] Shorter path lengths and higher planning efficiency

[0183] Lower obstacle avoidance times and higher success rate

[0184] Smoother path and superior running stability

[0185] Example 3, reference Figure 6 , which is an embodiment of the present invention, provides a dynamic obstacle avoidance system based on a global algorithm and a local improved potential field method, including a path planning module, a field method obstacle avoidance module, and a collaborative work module.

[0186] The path planning module is used to adopt the global A* algorithm for path planning, the field method obstacle avoidance module is used to avoid obstacles based on the local improved potential field method, and the collaborative work module is used to work collaboratively according to the global and local algorithms.

[0187] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0188] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0189] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering, or processing in another suitable manner as necessary, and then stored in a computer memory.

[0190] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc. It should be noted that the above embodiments are only used to illustrate the technical solution of the present invention and are not limited. Although the present invention is described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention, which should be included in the scope of the claims of the present invention.

[0191] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A dynamic obstacle avoidance method based on a global algorithm and a local improved potential field method, characterized in that: include: Path planning based on A* algorithm; Obstacle avoidance based on the local improved potential field method; Based on the judgment, the global and local algorithm mechanisms are switched to work in collaboration.

2. The dynamic obstacle avoidance method based on the global algorithm and the local improved potential field method as claimed in claim 1, characterized in that: The path planning based on the A* algorithm includes abstracting the working scene of the inspection robot into a grid map model, and setting the precision and overall coverage of the grid map according to the on-site information; Mark the location and outline shape of static obstacles on the grid map; The inspection robot operation scene is modeled as a grid map M, where each grid represents an area; The value of M[i][j] is used to distinguish the passable area and the obstacle area and is expressed as: Among them, M[i][j] represents the distinction between the passable area and the obstacle area; Based on the on-site environmental data, the outlines and positions of static obstacles are marked. The degree of detail is determined by the resolution of the grid, which needs to balance accuracy and computational cost.

3. The dynamic obstacle avoidance method based on the global algorithm and the local improved potential field method as claimed in claim 2, characterized in that: The path planning based on the A* algorithm includes clarifying the starting point and destination of the inspection robot, using the A* algorithm to search for a path on the constructed grid map, and outputting the optimal route; The path cost function is expressed as: f(n)=g(n)+h(n) Among them, g(n) represents the actual cost from the starting point to the current node, and h(n) represents the estimated cost from the current node to the target point; Manhattan distance is used to calculate h(n) as: Among them, x n ,y n Indicates the horizontal and vertical coordinates of the current node, x t ,y t Indicates the horizontal and vertical coordinates of the target node; Based on the minimum search of the cost function, a global optimal path P is generated from the starting point S to the target point T as follows: P={p1,p2,…,p k } Among them, p1, p2, …, p k Indicates a waypoint; Visibility graph method and Dijkstra algorithm are used to optimize the path.

4. The dynamic obstacle avoidance method based on the global algorithm and the local improved potential field method as claimed in claim 3, characterized in that: The obstacle avoidance method according to the local improved potential field method includes outputting a repulsive potential field according to the relative spatial position and relative motion speed between the obstacle and the robot when the inspection robot encounters a dynamic obstacle, which is expressed as: in, represents the repulsive potential field, k rep represents the repulsive force constant, d(x r ,y r ,x o ,y o ) represents the distance between the robot's current position and the obstacle, d0 represents the maximum distance threshold that affects the repulsive force, represents the unit vector pointing from the robot to the obstacle, indicating the direction of the repulsive force; The strength of the repulsive potential field changes dynamically as the distance between the obstacle and the robot changes; When the distance of the obstacle is inversely proportional to the strength of the repulsive force; The strength of the gravitational potential field is inversely proportional to the distance from the robot's current position to the target point.

5. The dynamic obstacle avoidance method based on the global algorithm and the local improved potential field method as claimed in claim 4, characterized in that: The obstacle avoidance according to the local improved potential field method includes adopting a strategy of adaptive sub-target point setting when the inspection robot encounters a local optimal solution problem during the path planning process, resulting in the inability to move directly toward the final target; The obstacle information around the robot is collected through sensors, and an obstacle set o = {o1, o2, ..., on} is established. The position of each obstacle i is (x oi ,y oi ); The obstacle area is defined as the spatial range that the robot cannot pass through. The distance function d(R,0i) is used to represent the shortest distance between the robot and the obstacle 0i. According to the current position of the robot and the distribution of surrounding obstacles, a candidate set of sub-target points is generated. Ensure that the candidate sub-target point is no more than the specified range d from the machine R max , to ensure the rationality of the sub-target points: d(R,T i )≤d max Among them, d(R,T i ) represents the robot and the obstacle T i The closest distance, d max Indicates the scope of regulations; Ensure that the sub-target point T i Not inside the obstacle area: Among them, r safe represents the safety distance threshold, d(T i ,O j ) indicates obstacle T i With obstacles O j The closest distance, O j Indicates an obstacle; The candidate points are limited by the direction vector angle θ and are expressed as: in, represents the robot motion direction vector, Represents the robot's target direction vector.

6. The dynamic obstacle avoidance method based on the global algorithm and the local improved potential field method as claimed in claim 5, characterized in that: The obstacle avoidance according to the local improved potential field method includes applying a fuzzy control mechanism to output a danger assessment index according to the relative distance, relative speed and relative direction between the robot and the obstacle; Fuzzy control adjusts the parameters of the repulsive potential field and the gravitational potential field according to the changes in the risk assessment index; The input variable is relative distance, and the output variable is repulsion adjustment coefficient; Fuzzify input and output variables, define language variables and membership functions, fuzzify input variables, and fuzzify output variables; The membership function is expressed as: Where d represents the relative distance, d near_max Indicates the maximum relative distance, d near_min Represents the minimum relative distance, μ Near (d) represents the membership function; Establish a fuzzy rule base based on experience or scenario analysis; Based on fuzzy rules, the fuzzy reasoning method is used to calculate the output membership: Output the premise membership of each rule, determine the applicability of the rule based on the premise membership, integrate all the rules, and output the fuzzy set.

7. The dynamic obstacle avoidance method based on the global algorithm and the local improved potential field method as claimed in claim 6, characterized in that: The said judging and switching global and local algorithm mechanisms to work in coordination includes the system automatically switching from the global A* algorithm to the local improved potential field method due to obstacles; Based on real-time environmental perception and analysis, determine whether it is necessary to switch from the global A* algorithm to the local improved potential field method, or vice versa; When the obstacle density exceeds the set threshold, it indicates that the robot is in an obstacle-dense area and needs to switch to the local improved potential field method; When the shortest distance from the robot to the obstacle is less than the safety distance, it triggers the switch to the local improved potential field method; If the obstacle is a dynamic obstacle and its relative speed exceeds the threshold, it is also necessary to switch to the local improved potential field method; If the robot is close to the target area and the remaining path length of the global A* plan is less than the threshold, switch to the local improved potential field method to complete the final adjustment; The comprehensive state index S is defined to quantify the complexity of the robot's current environment, expressed as: Among them, α1, α2, α3, α4 represent weight parameters, D obs represents the density of obstacles, d min represents the shortest distance from the robot to the obstacle, v obs represents the obstacle speed, P progress Indicates the stage completion degree of path planning; When S>St, switch to the local improved potential field method; When S≤St and obstacles are sparse, switch back to the global A* algorithm.

8. A system using the dynamic obstacle avoidance method based on the global algorithm and the local improved potential field method as claimed in any one of claims 1 to 7, characterized in that: Including path planning module, field obstacle avoidance module, collaborative work module; The path planning module is used for path planning using a global A* algorithm; The field method obstacle avoidance module is used for obstacle avoidance based on the local improved potential field method; The collaborative working module is used for collaborative working according to global and local algorithms.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the dynamic obstacle avoidance method based on the global algorithm and the local improved potential field method described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the dynamic obstacle avoidance method based on a global algorithm and a local improved potential field method according to any one of claims 1 to 7 are implemented.

Images