Robot control instruction analysis system based on multi-algorithm fusion

Abstract

The invention relates to the technical field of robot control instruction analysis, in particular to a robot control instruction analysis system based on multi-algorithm fusion, which comprises a data collection module, an environment map analysis module, a global path analysis module and a moving trajectory analysis module, wherein the data collection module collects state data of a robot, surrounding environment data of the robot and a target position; the environment map analysis module determines an environment occupation grid map according to the environment data; the global path analysis module determines a global path according to an environment occupation grid map, a target position, self state data and robot motion constraint parameters; the running track analysis module determines a smooth track according to the global path and the state data of the running track analysis module and detects an environment occupation grid map, if a dynamic obstacle avoidance object exists in the environment occupation grid map, the environment occupation grid map is updated, the smooth track is modified according to the environment occupation grid map updated in real time, and the dynamic obstacle avoidance object is obtained. And the final robot moving trajectory is obtained.

Description

Technical Field

[0001] The present invention relates to the technical field of robot control instruction analysis, and in particular to a robot control instruction analysis system based on multi-algorithm fusion. Background Art

[0002] Traditional robot control command analysis systems have obvious shortcomings in dealing with complex dynamic environments. First, they often cannot effectively integrate static path planning and dynamic obstacle avoidance functions, resulting in slow response or lack of flexibility when facing environmental changes.

[0003] Secondly, traditional systems find it difficult to strike a good balance between global path planning and local dynamic adjustment. They either focus too much on global optimality and ignore real-time performance, or over-pursue real-time response and sacrifice overall efficiency.

[0004] In addition, traditional systems usually lack effective mechanisms to dynamically identify and handle changes in the environment, especially in distinguishing between static and dynamic obstacles. This results in the system being unable to adjust its operating strategy in a timely manner, reducing the robot's adaptability and safety in complex environments.

[0005] Finally, the coordination between path smoothness and dynamic obstacle avoidance in traditional systems is not ideal, and it is often difficult to maintain the smoothness and efficiency of the trajectory while ensuring safety. These problems combined have severely limited the application effect and scope of traditional robot control systems in complex and dynamic environments. Summary of the invention

[0006] The purpose of the present invention is to provide a robot control instruction analysis system based on multi-algorithm fusion to solve the problems raised in the above background technology.

[0007] To achieve the above object, the present invention provides the following technical solution: a robot control instruction analysis system based on multi-algorithm fusion, including a data collection module, an environment map analysis module, a global path analysis module and an operation trajectory analysis module, wherein:

[0008] The data collection module collects the robot's own state data; uses environmental sensors to collect the robot's surrounding environment data; and collects the target position; wherein the self-state data includes the current position, direction and speed and the robot's motion constraint parameters;

[0009] The environmental map analysis module determines an environmental occupancy grid map based on the environmental data;

[0010] The global path analysis module determines the global path according to the environment occupation grid map, the target position, the current position in the self-state data and the robot motion constraint parameters;

[0011] The operation trajectory analysis module determines a smooth trajectory according to the global path and its own state data, and detects the environment occupancy grid map in the environment map analysis module, and determines whether there are dynamic obstacle avoidance objects by dynamically updating the environment occupancy grid map and comparing maps at different time points, wherein: if there are dynamic obstacle avoidance objects in the environment occupancy grid map, the environment occupancy grid map is updated, and a locally optimized trajectory is determined according to the real-time updated environment occupancy grid map, and the smooth trajectory is modified according to the locally optimized trajectory as the final robot operation trajectory.

[0012] As a further improvement of the present technical solution, the environmental data collected by the data collection module is composed of data collected by different environmental sensors; the data collection module establishes a connection with the task planning system to obtain the target position coordinates, and the target position coordinates represent the final position that the robot needs to reach.

[0013] As a further improvement of the technical solution, the environment map analysis module determines the environment occupancy grid map according to the environment data, specifically including:

[0014] Median filtering is used to remove outliers and noise points in the environmental data, and the data from different environmental sensors in the environmental data are unified into the robot coordinate system;

[0015] Initialize the two-dimensional array as a grid map, where each element represents a grid, the initial state is set to unknown, and the occupancy probability is set to 0.5;

[0016] According to the data in the unified coordinate system, the random sampling consensus algorithm is applied to identify and remove ground points, and Euclidean clustering is performed on non-ground points to separate different obstacles. Based on the clustering results, the obstacles are preliminarily classified;

[0017] Use the results of obstacle detection and classification to update the grid map, trace rays from the sensor location to each detection point, and use the inverse sensor model to update the grid occupancy probability, increase the occupancy probability of the obstacle location, and reduce the occupancy probability of the idle area where the ray passes, by using the logarithmic probability ratio to represent the grid state;

[0018] Based on the updated grid map, optimization is performed, morphological operations are applied to remove small noise areas, obstacle boundaries are extracted and optimized, and grids are classified into occupied, idle, and unknown according to probability thresholds;

[0019] Use the optimized grid map to identify and extract environmental features, identify static obstacles, and analyze;

[0020] Based on the optimized and feature-extracted map data, the quadtree structure is used to compress the map data as the environment occupancy grid map.

[0021] As a further improvement of the technical solution, the global path analysis module determines the global path using a path planning algorithm based on the environment occupation grid map, the target position, the current position in the self-state data, and the robot motion constraint parameters, specifically including:

[0022] Use the Dijkstra light algorithm in the path planning algorithm to plan the path. The Dijkstra light algorithm searches backward from the target, sets the target position as the starting point, and the current position as the end point; use the Manhattan distance to initialize the heuristic function;

[0023] Determine the target position and the current position of the robot, and create a two-dimensional array of the same size as the environment occupancy grid map to store the cost value;

[0024] Calculation of the actual cost g(n): Start from the target position with an initial value of 0; for each adjacent grid, check the grid state in the optimized grid map; set the basic cost according to the grid state: set the occupied grid to infinity, the idle grid to 1, and the unknown grid to a value greater than 1; add additional costs to the grids close to the edge of static obstacles; adjust the cost according to the robot motion constraint parameters, and update the actual cost value of the grid to the minimum cumulative cost to reach the grid;

[0025] Heuristic estimate h(n) calculation: Use the heuristic function to calculate the estimated cost to the robot's current position;

[0026] Calculation of the total estimated cost f(n): For each traversable grid, calculate f(n) = g(n) + h(n); use a quadtree structure to quickly access and update the grid cost;

[0027] Path planning: Maintain a priority queue, sort by f(n) value, take out the grid with the smallest f(n) from the queue for expansion, and when reaching the current position of the robot, backtrack to generate a complete global path.

[0028] As a further improvement of the technical solution, the running trajectory analysis module includes a smooth trajectory analysis unit, and the smooth trajectory analysis unit determines the smooth trajectory according to the global path and its own state data, specifically including:

[0029] The global path is converted into a sequence of path points in a continuous coordinate system, and the path is initially smoothed using a moving average method to remove jagged trajectories;

[0030] Determine the current dynamic parameters based on its own state data; calculate the curvature of the path and identify the area where the curvature changes; adjust the path in the high curvature area based on the robot's minimum turning radius to plan the path length;

[0031] According to the path curvature and robot motion constraint parameters, a speed is assigned to each point on the path, the speed is reduced at the turn and increased at the straight section, which is used for speed planning;

[0032] Use spline interpolation to generate multiple intermediate points between path points;

[0033] According to the planned path length and speed, the time dimension is added to the trajectory to generate a complete smooth trajectory of time-position-speed.

[0034] As a further improvement of the technical solution, the running trajectory analysis module includes a detection unit, which detects the environment occupancy grid map in the environment map analysis module, dynamically updates the environment occupancy grid map, compares the maps at different time points, identifies the changed area, and determines the local optimized trajectory according to the changed area, specifically including:

[0035] Continuously receive real-time updates from the environment map analysis module and store snapshots of the environment occupancy grid map at multiple time points. Use image difference technology to compare the environment occupancy grid maps at consecutive time points, and determine the changed area by comparing the total estimated cost of the grids, where:

[0036] If the total estimated cost of the grid in the environment occupancy grid map at different time points is different, it is determined that there is a dynamic obstacle avoidance object, and the corresponding change area is extracted. Based on the difference results of the environment occupancy grid map at different time points, the connected region analysis algorithm is used to identify the change area, and the boundary of each connected region is extracted to obtain the outline of the change area;

[0037] Calculate the geometric features of each change area, analyze the change of grid values ​​in the change area, and estimate the type of obstacles; prioritize the change areas according to the distance and area size between the change area and the current planned trajectory, and select the change area with the highest priority;

[0038] Centered on the highest priority change area, the range is expanded as the re-planning influence area according to the current state of the robot; a point outside the influence area is selected on the original trajectory as the target point of local planning; a local path is generated within the influence area using a fast replanning algorithm.

[0039] As a further improvement of the technical solution, the running trajectory analysis module includes an optimization unit, and the optimization unit modifies the smooth trajectory according to the local path as the final robot running trajectory, specifically including:

[0040] The local path is inserted into the smooth trajectory, and cubic splines are used at the connection points of the trajectory to ensure the continuity of the trajectory. Specifically, the smooth trajectory is used, combined with the corresponding data of the local path, and a set of equations is established at the connection points. By solving the set of equations, a cubic polynomial function that can smoothly connect the global path and the local path is obtained;

[0041] The Bezier curve is applied to smooth the entire modified smooth trajectory. When using the Bezier curve to smooth the entire trajectory, the key points in the modified trajectory are used as control points. The control points include but are not limited to path points and connection points after cubic spline interpolation. The Bezier curve is constructed according to the control points. For each segment on the trajectory, several adjacent control points are used to generate a smooth curve segment, and these smooth curve segments are connected as the final robot operation trajectory.

[0042] Compared with the prior art, the present invention has the following beneficial effects:

[0043] 1. The environment occupancy grid map plays multiple key roles in the robot control command analysis system based on multi-algorithm fusion. It can not only determine the global path, but also determine whether there are dynamic obstacles to avoid by dynamically updating and comparing maps at different time points, thus producing significant beneficial effects. In terms of determining the global path, the environment occupancy grid map provides accurate environmental information for the path planning algorithm by accurately representing obstacles and traversable areas in the environment, so that the planned path can effectively avoid static obstacles. At the same time, through the dynamic update mechanism, the system can capture environmental changes in real time and compare maps at different time points to identify dynamic obstacles. This dynamic recognition capability enables the robot to adjust its operation strategy in time and effectively respond to sudden changes in the environment, greatly improving the adaptability and safety of the robot in complex dynamic environments. In addition, the multiple applications of the environment occupancy grid map also realize the seamless integration of static path planning and dynamic obstacle avoidance, so that the system can flexibly respond to local environmental changes while ensuring the optimality of the global path, thereby improving the overall operation efficiency and reliability of the robot.

[0044] 2. The total estimated cost in the robot control command analysis system based on multi-algorithm fusion is used in multiple applications in this system, which reflects its core role in path planning and dynamic obstacle avoidance, and produces significant beneficial effects. In global path planning, the system uses the total estimated cost f(n) = g(n) + h(n) to evaluate the cost of each grid, where g(n) reflects the actual cost from the starting point to the current grid, and h(n) estimates the expected cost from the current grid to the target. By minimizing the total estimated cost, the system can efficiently generate the optimal global path based on actual environmental constraints and goal orientation. At the same time, in the dynamic obstacle avoidance process, In the paper, by comparing the changes in the total estimated cost of the grids in the grid map occupied by the environment at different time points, we can determine whether there are dynamic obstacles and extract the corresponding change areas; this dynamic obstacle detection method based on the total estimated cost can not only accurately identify changes in the environment, but also reflect the potential impact of these changes on path planning; through this multiple application, the total estimated cost becomes the key bridge connecting static path planning and dynamic environment perception, enabling the system to quickly respond to environmental changes while ensuring the optimality of the global path, realizing the organic combination of static planning and dynamic adjustment, and greatly improving the robot's navigation performance in complex dynamic environments.

[0045] 3. The robot control instruction analysis system based on multi-algorithm fusion chooses to perform dynamic obstacle avoidance detection only after the smooth trajectory is determined, rather than directly in the global path planning stage. This design reflects the delicate balance between efficiency and real-time performance and produces significant beneficial effects. First, this method allows the system to quickly generate a global path and a smooth trajectory based on relatively stable environmental information, providing the robot with an initial, overall optimal motion plan. This can avoid excessive attention to instantaneous changes in the global planning stage, improving the efficiency and stability of path generation. Secondly, dynamic obstacle avoidance detection is performed after the smooth trajectory is determined, allowing the system to focus more on environmental changes related to the actual execution path. This method can effectively reduce unnecessary calculations and adjustments, because only those dynamic obstacles that may affect the current execution trajectory will be considered. Finally, this design allows the system to flexibly respond to local environmental changes while maintaining the integrity of the global planning. By performing local optimization based on the smooth trajectory, the system can adjust only the necessary parts while maintaining the general shape of the original trajectory, thereby maximizing the smoothness and efficiency of the original trajectory while ensuring safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 It is a schematic diagram of the overall module of the present invention;

[0047] Figure 2 It is a schematic diagram of the running trajectory analysis module unit of the present invention.

[0048] In the figure: 100, data collection module; 200, environment map analysis module; 300, global path analysis module; 400, operation trajectory analysis module; 401, smooth trajectory analysis unit; 402, detection unit; 403, optimization unit. DETAILED DESCRIPTION

[0049] The following will be combined with the accompanying drawings in the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0050] See also Figure 1-2 The present invention provides a technical solution: a robot control instruction analysis system based on multi-algorithm fusion, including a data collection module 100, an environment map analysis module 200, a global path analysis module 300 and an operation trajectory analysis module 400.

[0051] The data collection module 100 collects the robot's own state data, the robot's surrounding environment data and the target position; wherein the own state data includes the current position, direction and speed and the robot's motion constraint parameters; specifically includes:

[0052] Obtain the current position coordinates by using internal sensors (such as encoders, IMU); obtain the current direction (such as yaw angle, pitch angle, roll angle); measure the current speed (such as linear speed and angular velocity); record the robot motion constraint parameters (such as maximum speed, acceleration, turning radius, etc.);

[0053] Use different environmental sensors (such as lidar and cameras) to scan the surrounding environment and obtain the robot's surrounding environment data. The environmental data consists of data collected by different environmental sensors, such as point cloud data and image data;

[0054] The target position coordinates are obtained from the task planning system, where the task planning system is responsible for formulating and managing the overall tasks and goals of the robot, and the target position coordinates are obtained by establishing a connection between the data collection module 100 and the task planning system.

[0055] The environment map analysis module 200 determines the environment occupancy grid map according to the surrounding environment data, specifically including:

[0056] Apply median filtering to remove outliers and noise points in the environmental data, and unify the data of different environmental sensors in the environmental data into the robot coordinate system, such as converting point cloud data and image data;

[0057] Initialize the two-dimensional array as a grid map, where each element represents a grid, the initial state is set to unknown, and the occupancy probability is set to 0.5;

[0058] According to the data in the unified coordinate system, the random sampling consensus algorithm is applied to identify and remove ground points, and Euclidean clustering is performed on non-ground points to separate different obstacles. Based on the clustering results, obstacles (such as walls, objects, etc.) are preliminarily classified;

[0059] Use the results of obstacle detection and classification to update the grid map, perform ray tracing from the sensor position to each detection point, use the inverse sensor model to update the grid occupancy probability, increase the occupancy probability of the obstacle position, and reduce the occupancy probability of the idle area where the ray passes. Use the logarithmic probability ratio to represent the grid state to avoid probability saturation.

[0060] Based on the updated grid map, optimization is performed, morphological operations are applied to remove small noise areas, obstacle boundaries are extracted and optimized, and grids are classified into occupied, idle, and unknown according to probability thresholds;

[0061] Use the optimized raster map to identify and extract environmental features, identify static obstacles such as walls and fixed furniture, and analyze and identify large-scale spatial structures such as rooms and corridors;

[0062] Based on the map data after optimization and feature extraction, the quadtree structure is used to compress the map data as the environment occupancy grid map, which is convenient for subsequent modules to use.

[0063] The global path analysis module 300 determines the global path using a path planning algorithm based on the environment occupancy grid map, the target position, the current position in the robot's own state data, and the robot's motion constraint parameters, specifically including:

[0064] Use the Dijkstra light algorithm in the path planning algorithm to plan the path. The Dijkstra light algorithm searches backward from the target, sets the target position as the starting point, and the current position as the end point; use the Manhattan distance to initialize the heuristic function;

[0065] Determine the target position and the current position of the robot, and create a two-dimensional array of the same size as the environment occupancy grid map to store the cost value;

[0066] Calculation of the actual cost g(n): Start from the target position with an initial value of 0; for each adjacent grid, check the grid state in the optimized grid map; set the base cost according to the grid state: set the occupied grid to infinity, the idle grid to 1, and the unknown grid to a value greater than 1 to encourage exploration of known areas;

[0067] Consider environmental characteristics: grids close to the edge of static obstacles have additional costs, and grids in the center of the identified room or corridor have slightly reduced costs;

[0068] Apply robot motion constraints: adjust the cost based on turning radius, etc., and update the actual cost value of the grid to the minimum cumulative cost to reach the grid;

[0069] Heuristic estimate h(n) calculation: Use Euclidean distance or Manhattan distance to calculate the estimated cost to the robot's current position;

[0070] Calculation of the total estimated cost f(n): For each traversable grid, calculate f(n) = g(n) + h(n); use a quadtree structure to quickly access and update the grid cost;

[0071] Path planning: Maintain a priority queue, sort by f(n) value, take out the grid with the smallest f(n) from the queue for expansion, and when reaching the current position of the robot, backtrack to generate a complete global path.

[0072] The smooth trajectory analysis unit 401 in the running trajectory analysis module 400 determines the smooth trajectory according to the global path and its own state data, specifically including:

[0073] The global path is converted into a sequence of path points in a continuous coordinate system, and the path is initially smoothed using methods such as moving average or Bezier curve to remove jagged trajectories;

[0074] Determine the current dynamic parameters (maximum speed, acceleration, angular velocity, etc.) based on its own state data; calculate the curvature of the path and identify sharp turns or areas with drastic changes in curvature; adjust the path in high curvature areas based on the robot's minimum turning radius to plan the path length and ensure smoothness;

[0075] According to the path curvature and robot motion constraint parameters, a speed is assigned to each point on the path, the speed is reduced at the turn, and the speed is increased in the straight section, ensuring smooth speed changes and avoiding sudden acceleration or deceleration, which is used for speed planning;

[0076] Use spline interpolation (such as cubic spline) to generate multiple intermediate points between path points to ensure that the interpolated trajectory meets the continuity and smoothness requirements of the robot;

[0077] Time parameterization: Add a time dimension to the trajectory based on the planned path length and speed, and generate a complete smooth trajectory of time-position-speed.

[0078] The detection unit 402 in the running trajectory analysis module 400 detects the environment occupancy grid map in the environment map analysis module 200, dynamically updates the environment occupancy grid map, compares the maps at different time points, identifies the changed area, and determines the local optimized trajectory according to the changed area, specifically including:

[0079] Continuously receive real-time updates from the environment map analysis module 200, and store snapshots of the environment occupancy grid map at multiple time points, use image difference technology to compare the environment occupancy grid maps at consecutive time points, and determine the changed area by comparing the total estimated cost of the grids therein, where:

[0080] If the total estimated cost of the grid in the environment occupancy grid map at different time points is different, it is determined that there is a dynamic obstacle avoidance object, and the corresponding change area is extracted. Based on the difference results of the environment occupancy grid map at different time points, the connected region analysis algorithm is used to identify the change area, and the boundary of each connected region is extracted to obtain the outline of the change area;

[0081] Calculate the geometric features of each change area, such as area, centroid, and direction, analyze the change of grid values ​​in the change area, and estimate the type of obstacles; prioritize the change areas according to factors such as the distance between the change area and the current planned trajectory and the size of the area, and select the change area with the highest priority;

[0082] Taking the highest priority change area as the center, according to the current state of the robot, expand a certain range as the influence area that needs to be replanned; select a point outside the influence area on the original trajectory as the target point of local planning; use a fast replanning algorithm (such as the dynamic window method) to generate a local path within the influence area.

[0083] The optimization unit 403 in the running trajectory analysis module 400 modifies the smooth trajectory according to the local path as the final robot running trajectory, which specifically includes:

[0084] Insert the local path into the smooth trajectory to replace the affected part. Use spline interpolation method (such as cubic spline) at the connection point of the trajectory to ensure the continuity of the trajectory. Specifically, use the smooth trajectory and the corresponding data of the local path to establish a system of equations at the connection point. By solving the system of equations, a cubic polynomial function that can smoothly connect the global path and the local path is obtained, thereby achieving seamless connection of the trajectory.

[0085] A path smoothing algorithm (such as Bezier curve) is applied to smooth the entire modified smooth trajectory. When using the Bezier curve for overall trajectory smoothing, the key points in the modified trajectory (including the original smooth trajectory and the inserted local path) are used as control points. These control points include path points and connection points after cubic spline interpolation. Bezier curves are constructed based on the control points. For each segment on the trajectory, several adjacent control points (usually a cubic Bezier curve formed by four points) are used to generate a smooth curve segment, and these smooth curve segments are connected as the final robot operation trajectory.

[0086] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and descriptions are only preferred examples of the present invention and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A robot control instruction analysis system based on multi-algorithm fusion, characterized in that: The system comprises a data collection module (100), an environment map analysis module (200), a global path analysis module (300) and a running track analysis module (400), wherein: The data collection module (100) collects the robot's own state data; uses environmental sensors to collect the robot's surrounding environment data; and collects the target position; wherein the own state data includes the current position, direction and speed, and the robot's motion constraint parameters; The environmental map analysis module (200) determines an environmental occupancy grid map according to the environmental data; The global path analysis module (300) determines the global path according to the environment occupation grid map, the target position, the current position in the self-state data and the robot motion constraint parameters; The operation trajectory analysis module (400) determines a smooth trajectory based on the global path and its own state data, and detects the environment occupancy grid map in the environment map analysis module (200), and determines whether there are dynamic obstacle avoidance objects by dynamically updating the environment occupancy grid map and comparing maps at different time points, wherein: if there are dynamic obstacle avoidance objects in the environment occupancy grid map, the environment occupancy grid map is updated, and a local optimized trajectory is determined based on the real-time updated environment occupancy grid map, and the smooth trajectory is modified based on the local optimized trajectory to serve as the final robot operation trajectory.

2. The robot control instruction analysis system based on multi-algorithm fusion according to claim 1 is characterized in that: The environmental data collected by the data collection module (100) is composed of data collected by different environmental sensors; the data collection module (100) establishes a connection with the task planning system to obtain the target position coordinates, and the target position coordinates represent the final position that the robot needs to reach.

3. The robot control instruction analysis system based on multi-algorithm fusion according to claim 2 is characterized in that: The environment map analysis module (200) determines the environment occupancy grid map according to the environment data, specifically comprising: Median filtering is used to remove outliers and noise points in the environmental data, and the data from different environmental sensors in the environmental data are unified into the robot coordinate system; Initialize the two-dimensional array as a grid map, where each element represents a grid, the initial state is set to unknown, and the occupancy probability is set to 0.5; According to the data in the unified coordinate system, the random sampling consensus algorithm is applied to identify and remove ground points, and Euclidean clustering is performed on non-ground points to separate different obstacles. Based on the clustering results, the obstacles are preliminarily classified; Use the results of obstacle detection and classification to update the grid map, trace rays from the sensor location to each detection point, and use the inverse sensor model to update the grid occupancy probability, increase the occupancy probability of the obstacle location, and reduce the occupancy probability of the idle area where the ray passes, by using the logarithmic probability ratio to represent the grid state; Based on the updated grid map, optimization is performed, morphological operations are applied to remove small noise areas, obstacle boundaries are extracted and optimized, and grids are classified into occupied, idle, and unknown according to probability thresholds; Use the optimized grid map to identify and extract environmental features, identify static obstacles, and analyze; Based on the optimized and feature-extracted map data, the quadtree structure is used to compress the map data as the environment occupancy grid map.

4. The robot control instruction analysis system based on multi-algorithm fusion according to claim 3 is characterized in that: The global path analysis module (300) determines the global path using a path planning algorithm according to the environment occupation grid map, the target position, the current position in the robot's own state data, and the robot's motion constraint parameters, specifically including: Determine the target position and the current position of the robot, and create a two-dimensional array of the same size as the environment occupancy grid map to store the cost value; Calculation of the actual cost g(n): Start from the target position with an initial value of 0; for each adjacent grid, check the grid state in the optimized grid map; set the basic cost according to the grid state: set the occupied grid to infinity, the idle grid to 1, and the unknown grid to a value greater than 1; add additional costs to the grids close to the edge of static obstacles; adjust the cost according to the robot motion constraint parameters, and update the actual cost value of the grid to the minimum cumulative cost to reach the grid; Heuristic estimate h(n) calculation: Use the heuristic function to calculate the estimated cost to the robot's current position; Calculation of the total estimated cost f(n): For each traversable grid, calculate f(n) = g(n) + h(n); use a quadtree structure to quickly access and update the grid cost; Path planning: Maintain a priority queue, sort by f(n) value, take out the grid with the smallest f(n) from the queue for expansion, and when reaching the current position of the robot, backtrack to generate a complete global path.

5. The robot control instruction analysis system based on multi-algorithm fusion according to claim 4 is characterized in that: The running trajectory analysis module (400) comprises a smooth trajectory analysis unit (401), and the smooth trajectory analysis unit (401) determines a smooth trajectory according to a global path and its own state data, specifically comprising: The global path is converted into a sequence of path points in a continuous coordinate system, and the path is initially smoothed using a moving average method to remove jagged trajectories; Determine the current dynamic parameters based on its own state data; calculate the curvature of the path and identify the area where the curvature changes; adjust the path in the high curvature area based on the robot's minimum turning radius to plan the path length; According to the path curvature and robot motion constraint parameters, a speed is assigned to each point on the path, the speed is reduced at the turn and increased at the straight section, which is used for speed planning; Use spline interpolation to generate multiple intermediate points between path points; According to the planned path length and speed, the time dimension is added to the trajectory to generate a complete smooth trajectory of time-position-speed.

6. The robot control instruction analysis system based on multi-algorithm fusion according to claim 5 is characterized in that: The running trajectory analysis module (400) comprises a detection unit (402), wherein the detection unit (402) detects the environment occupancy grid map in the environment map analysis module (200), dynamically updates the environment occupancy grid map, compares maps at different time points, identifies the changed area, and determines the local optimized trajectory according to the changed area, specifically comprising: Continuously receiving real-time updates from the environment map analysis module (200), and storing snapshots of the environment occupancy grid map at multiple time points, using image difference technology to compare the environment occupancy grid maps at consecutive time points, and determining the changed area by comparing the total estimated cost of the grids therein, wherein: If the total estimated cost of the grid in the environment occupancy grid map at different time points is different, it is determined that there is a dynamic obstacle avoidance object, and the corresponding change area is extracted. Based on the difference results of the environment occupancy grid map at different time points, the connected region analysis algorithm is used to identify the change area, and the boundary of each connected region is extracted to obtain the outline of the change area; Calculate the geometric features of each change area, analyze the change of grid values ​​in the change area, and estimate the type of obstacles; prioritize the change areas according to the distance and area size between the change area and the current planned trajectory, and select the change area with the highest priority; Centered on the highest priority change area, the range is expanded as the re-planning influence area according to the current state of the robot; a point outside the influence area is selected on the original trajectory as the target point of local planning; a local path is generated within the influence area using a fast replanning algorithm.

7. The robot control instruction analysis system based on multi-algorithm fusion according to claim 6 is characterized in that: The running trajectory analysis module (400) includes an optimization unit (403), and the optimization unit (403) modifies the smooth trajectory according to the local path as the final robot running trajectory, specifically including: The local path is inserted into the smooth trajectory, and cubic splines are used at the connection points of the trajectory to ensure the continuity of the trajectory. Specifically, the smooth trajectory is used, combined with the corresponding data of the local path, and a set of equations is established at the connection points. By solving the set of equations, a cubic polynomial function that can smoothly connect the global path and the local path is obtained; The Bezier curve is applied to smooth the entire modified smooth trajectory. When using the Bezier curve to smooth the entire trajectory, the key points in the modified trajectory are used as control points. The control points include but are not limited to path points and connection points after cubic spline interpolation. The Bezier curve is constructed according to the control points. For each segment on the trajectory, several adjacent control points are used to generate a smooth curve segment, and these smooth curve segments are connected as the final robot operation trajectory.

Images