Intelligent path planning and control system of automatic driving carrying robot

Abstract

The invention belongs to the technical field of data processing, and particularly discloses an intelligent path planning and control system for an automatic driving carrying robot, and the system is operated through the following method: collecting environment data through a multi-mode sensor, and processing a laser radar point cloud and a camera image; fusing multi-source information to generate a real-time map, and obtaining a three-dimensional environment model containing dynamic obstacle positions; according to the three-dimensional environment model, obstacle features are extracted, an improved neural network model is adopted to identify a plurality of current obstacles, obstacle types and movement tracks are judged, and a classified obstacle data set is obtained; through the classified obstacle data set, the distance and the relative speed between the obstacle and the current position of the robot are calculated, and if the distance is smaller than a preset threshold value and the speed is larger than zero, high-priority interference is marked; the objective of the invention is to solve the problem that accurate sensing, real-time decision making and efficient obstacle avoidance are difficult to realize in a complex dynamic environment in the prior art.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, specifically to an intelligent path planning and control system for an autonomous driving transport robot. Background Technology

[0002] In the application scenarios of autonomous transport robots, the robots need to achieve autonomous navigation and obstacle avoidance in complex and dynamic environments to ensure the efficiency and safety of transportation tasks. However, the core technical challenge lies in how to construct a real-time updated 3D environment model in a dynamic and unpredictable environment, and then achieve accurate path planning and dynamic obstacle avoidance based on this model. This major problem encompasses the challenges of the entire process from data acquisition to decision execution. Especially in scenarios such as logistics centers or open roads, the environment often includes moving vehicles, pedestrians, and other unpredictable dynamic obstacles. The robot needs to accurately identify the type, location, and trajectory of these obstacles under limited computing resources and real-time requirements, while also assessing their potential threat level to its own path.

[0003] Focusing further on the details, the LiDAR point cloud data and camera image data collected by sensors may contain noise or be incomplete. How to fuse these multi-source heterogeneous data to generate a reliable real-time map is the primary technical challenge. Furthermore, when identifying obstacles, the characteristics of pedestrians and vehicles differ significantly, and deep learning models may face the risk of misjudgment or omission, especially under conditions of drastic lighting changes or severe occlusion. More critically, even after generating path planning, predicting potential collision points in dynamic environments and adjusting path parameters in a timely manner to cope with sudden changes remains a complex problem. These seemingly minor issues all revolve around the core problem: how to achieve accurate environmental perception, rapid decision-making, and safe navigation for robots in dynamic environments. Especially in autonomous driving scenarios, any delay in perception or decision-making can lead to mission failure or safety accidents. Therefore, solving this problem is of paramount importance for improving robot autonomy. Summary of the Invention

[0004] This invention provides an intelligent path planning and control system for an autonomous driving transport robot, aiming to solve the problem that existing technologies struggle to achieve accurate perception, real-time decision-making, and efficient obstacle avoidance in complex dynamic environments.

[0005] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: An intelligent path planning and control method for an autonomous driving transport robot includes: collecting environmental data through multimodal sensors, processing lidar point clouds and camera images, fusing multi-source information to generate a real-time map, and obtaining a 3D environment model containing the locations of dynamic obstacles; extracting obstacle features based on the 3D environment model, identifying several current obstacles using an improved neural network model, determining the obstacle type and trajectory, and obtaining a classified obstacle dataset; calculating the distance and relative speed between the obstacle and the robot's current position using the classified obstacle dataset, marking obstacles as high-priority interference if the distance is less than a preset threshold and the speed is greater than zero, and obtaining a priority-ranked obstacle list; and inputting the data according to the priority-ranked obstacle list. The algorithm incorporates the current map to generate alternative paths, evaluates the length and obstacle avoidance probability of each path, and obtains an optimal path sequence. Using this optimal path sequence, the robot's trajectory is simulated, and an adaptive filtering method is used to predict potential collision points. If a predicted point overlaps with an obstacle, the sequence parameters are adjusted to obtain a corrected path sequence. Based on the corrected path sequence, a control command sequence is generated and transmitted to the robot's actuator. Execution deviations are monitored in real time, and it is determined whether the deviation exceeds a threshold. If it does, the commands are recalculated to obtain updated control commands. Using the updated control commands and sensor feedback data, the 3D environment model is continuously optimized, and path planning is iterated to obtain a final inspection path adapted to the dynamic environment.

[0006] In one aspect of the invention, environmental data is collected by a multimodal sensor, and LiDAR point clouds and camera images are processed using a spatial mapping algorithm. Multi-source information is then fused to generate a real-time map, resulting in a three-dimensional environmental model that includes the locations of dynamic obstacles. Data is collected from the field environment using multimodal sensors to obtain raw information including lidar point clouds and camera images, thus obtaining preliminary environmental perception data. Based on the collected lidar point cloud, a spatial registration algorithm is used to perform coordinate correction on the point cloud data, determine the position information in the point cloud, and form preliminary spatial structure data. By using camera images and combining visual mapping algorithms to extract features and perform geometric projection on the image content, visual mapping data corresponding to the LiDAR point cloud is obtained, and the basic content of the real-time map is constructed. By employing multi-source data fusion technology, spatial structure data and visual mapping data are integrated and processed to obtain fused environmental data that includes the location of dynamic obstacles; If dynamic obstacles are detected in the fused environmental data, the obstacle locations are filtered according to preset filtering rules to determine whether the obstacles are in critical areas and generate filtered obstacle distribution data. Based on the filtered obstacle distribution data, a three-dimensional environment model containing the dynamic obstacle locations is constructed to obtain complete environmental description information; By continuously updating the 3D environment model and adjusting the movement trajectory of dynamic obstacles in real time, the final dynamic environmental information is determined.

[0007] In one aspect of the invention, obstacle features are extracted based on a three-dimensional environment model, and an improved neural network model is used to identify pedestrians and vehicles, determine the obstacle type and movement trajectory, and obtain a classified obstacle dataset, including: By acquiring 3D environmental data, an initial environmental model is constructed, and multi-source data fusion technology is used to integrate multi-source information to obtain preliminary environmental model data. Based on preliminary environmental model data, obstacle features are extracted, and a pre-trained improved neural network model is applied to perform preliminary identification of objects in the environment to determine the set of potential obstacle objects. For a set of potential obstacle objects, analyze the distinguishing features between pedestrians and vehicles. If an object is detected to have dynamic displacement, it is classified as a pedestrian, and a subset of dynamic objects is obtained after classification. Based on the categorized subset of dynamic objects and combined with historical location data, the continuity characteristics of the motion trajectory are calculated, and the trajectory analysis method is used to determine the motion trajectory pattern of the dynamic objects. By using motion trajectory patterns, static objects are reconfirmed. If the position of a static object does not change and it meets the characteristics of a vehicle, it is classified as a vehicle, and a subset of static obstacles is determined. Based on the dynamic object subset and the static obstacle subset, obstacle type information is integrated, and classification and organization techniques are applied to construct a complete obstacle dataset. By using a complete obstacle dataset, the 3D environment model is updated, and data storage technology is used to save the classification results, thereby obtaining real-time updated environmental information.

[0008] In one aspect of the invention, using a classified obstacle dataset, the distance and relative speed between the obstacle and the robot's current position are calculated. If the distance is less than a preset threshold and the speed is greater than zero, it is marked as a high-priority obstacle, thus obtaining a priority-ranked list of obstacles, including: By extracting the coordinate and motion information of each obstacle from the obstacle dataset and combining it with real-time data on the robot's position, the spatial distance and relative velocity between each obstacle and the robot are calculated to obtain preliminary distance and velocity results. Based on the calculated distance and speed results, condition judgment is made for each obstacle. If the distance is lower than a preset threshold and the relative speed is greater than zero, it is marked as a high-priority interference, and the marked obstacle classification set is obtained. Using the labeled obstacle classification set, all obstacles are sorted according to the identifier of high-priority interference to generate an initial priority sorting list; By combining the initial priority ranking list with dynamic updates of position and velocity data, the distance and velocity status of each obstacle are re-verified to determine whether the high priority conditions are still met, and the updated priority ranking results are obtained. Based on the updated priority ranking results, detailed motion trajectory data of obstacles with high priority interference are obtained, potential collision risks are analyzed, and the final priority ranking list is determined. By combining the final priority ranking list with the changes in the robot's position and obstacle dataset acquired in real time, the fluctuations in distance and speed are continuously monitored to determine whether the priority needs to be adjusted, and a dynamically adjusted obstacle priority list is obtained. A dynamically adjusted obstacle priority list is used to generate structured data output, which is then called by the subsequent path planning module to determine the execution order of the real-time obstacle avoidance strategy.

[0009] In one aspect of the invention, an obstacle list ordered by priority is input into a path planning algorithm, candidate paths are generated in conjunction with the current map, the length and obstacle avoidance probability of each path are evaluated, and a preferred path sequence is obtained, including: An initial obstacle distribution map is generated by retrieving sorted data from the obstacle priority list and combining it with map data sources in the current map library. Based on the obstacle distribution map and path planning method, a set of alternative paths is constructed using a pre-established path generation model; For each path in the candidate path set, a path evaluation method is used to calculate the path length value and record the length data of each path; If the path length exceeds the preset threshold range, the obstacle avoidance probability value is further analyzed by probability calculation to obtain the obstacle avoidance probability data for each path. Based on the obstacle avoidance probability value and path length value, the candidate path set is comprehensively sorted using the sequence sorting method to determine the preliminary preferred path order; By performing secondary verification on the path data in the preferred path sequence and combining it with the dynamic information after the obstacle distribution map is updated, it is determined whether the path sequence needs to be adjusted, and the final preferred path sequence is obtained. If the obstacle avoidance probability value in the final preferred path sequence is lower than the preset threshold, then the candidate path set is regenerated by combining the latest map data source in the current map library, and the path evaluation method and sequence sorting method are executed cyclically to determine a new preferred path sequence.

[0010] In one aspect of the invention, by optimizing the path sequence and simulating the robot's motion trajectory, an adaptive filtering method is used to predict potential collision points. If the predicted point overlaps with an obstacle, the sequence parameters are adjusted to obtain a corrected path sequence, including: By optimizing the path sequence, the initial simulation data of the robot's motion trajectory is obtained. An adaptive filtering method is used to estimate the distribution of trajectory points and obtain the prediction results of potential collision points. Based on the prediction results of potential collision points, determine whether the predicted points overlap with obstacles. If the predicted points overlap with obstacles, extract the key control points in the path parameters and determine the range of parameters that need to be adjusted. By extracting the range of path parameters, the parameters are optimized using preset adjustment rules to generate the first corrected path sequence; Based on the first corrected path sequence, the robot's motion trajectory is re-simulated to obtain the updated trajectory point distribution and determine whether there are any new potential collision points. If potential collision points still exist in the updated trajectory point distribution, the local weights of the path parameters are adjusted for the new collision point locations to obtain the second corrected path sequence. The final robot motion trajectory data is obtained through the second corrected path sequence. An adaptive filtering method is used to verify the final trajectory and determine whether the collision risk has been eliminated. If no collision risk is found during the final trajectory verification, the second corrected path sequence will be used as the final path output to generate control commands for the robot's movement.

[0011] In one aspect of the invention, a control command sequence is generated based on the corrected path sequence, transmitted to the robot actuator, and the execution deviation is monitored in real time to determine whether the deviation exceeds a threshold. If it does, the commands are recalculated to obtain updated control commands, including: Initial data is obtained from the path sequence, and the path sequence is transformed into a control command sequence using a pre-established mapping model to obtain a preliminary set of control commands. The control command set is sent to the robot execution module through the transmission channel, and the status data during the execution process is collected in real time to determine the dynamic changes during the execution process; For dynamically changing data, a real-time monitoring mechanism is used to compare the data with preset execution standards, calculate the execution deviation value, and determine whether the deviation exceeds the preset threshold. If the deviation exceeds the preset threshold, the instruction update process is triggered, and the adjusted control instruction sequence is regenerated by combining the execution feedback data and the path correction logic. By adjusting the control command sequence, the command input of the robot execution module is updated, new execution status data is obtained, and the execution effect is continuously monitored. Based on continuously monitored data, the deviation judgment process is executed cyclically. If the deviation still exceeds the preset threshold, the instruction update and adjustment steps are repeated to obtain stable execution results. After the execution results stabilize, the final path sequence and the correspondence between the control commands are recorded and stored in the historical database for reference in subsequent tasks.

[0012] In one aspect of the invention, by updating control commands, fusing sensor feedback data, continuously optimizing the three-dimensional environment model, and employing a path planning algorithm for iterative path planning, a final inspection path adapted to the dynamic environment is obtained, including: By collecting real-time environmental information through sensors and fusing feedback data from sensors and historical records, an initial dataset is constructed to obtain preliminary environmental perception results. Based on the preliminary environmental perception results and combined with the pre-established 3D model, a Kalman filter is used to perform data alignment and correction operations to determine the basis of the updated environmental model. Based on the updated environmental model, and integrating real-time adjustments of control commands, the A* algorithm is applied to perform preliminary path calculations and obtain an initial inspection path scheme. For the initial inspection path plan, combined with the dynamic environment change information, OpenCV is used to perform path feasibility analysis. If obstacles or environmental changes exceed the preset threshold, the path parameters are readjusted to obtain an optimized path plan. By continuously collecting sensor feedback data through the optimized path scheme, updating the dynamic elements in the environmental model, and determining whether the path execution conditions are met; If the path execution conditions are met, the final inspection path data will be generated based on the updated environment model and the optimized path scheme. If the conditions are not met, return to the OpenCV-based path feasibility analysis stage and iteratively calculate the path scheme. After obtaining the final inspection path data, the execution details of the inspection path are dynamically adjusted in conjunction with real-time control commands to determine the real-time adaptive solution for the inspection task.

[0013] In another aspect, the present invention also relates to an intelligent path planning and control system for an autonomous driving transport robot, the system comprising: The data acquisition and modeling module is configured to acquire field data through multimodal sensors, process lidar point clouds and camera images using spatial mapping algorithms, fuse multi-source information to generate a real-time map, and obtain a three-dimensional environment model containing the location of dynamic obstacles. The obstacle recognition and classification module is configured to extract obstacle features based on a 3D environment model, identify pedestrians and vehicles using an improved neural network model, determine the obstacle type and movement trajectory, and obtain a classified obstacle dataset. The priority evaluation module is configured to calculate the distance and relative speed between the obstacle and the robot's current position using the classified obstacle dataset. If the distance is less than a preset threshold and the speed is greater than zero, it is marked as a high-priority interference, thus obtaining a priority-ranked list of obstacles. The path planning module is configured to take a priority-sorted list of obstacles, which is then fed into the path planning algorithm. This algorithm generates alternative paths based on the current map, evaluates the length and obstacle avoidance probability of each path, and obtains the optimal path sequence. The collision prediction and correction module is configured to simulate the robot's motion trajectory by optimizing the path sequence, predict potential collision points using an adaptive filtering method, and adjust the sequence parameters to obtain a corrected path sequence if the predicted point overlaps with an obstacle. The control command generation and update module is configured to generate a control command sequence based on the corrected path sequence, transmit it to the robot actuator, monitor the execution deviation in real time, determine whether the deviation exceeds the threshold, and if it does, recalculate the command to obtain the updated control command. The environmental model optimization module is configured to continuously optimize the 3D environmental model by integrating sensor feedback data with updated control commands, and to obtain the final inspection path that adapts to the dynamic environment by iteratively planning the path using a path planning algorithm.

[0014] Compared with the prior art, the present invention has the following beneficial effects: This invention first addresses the problems of inaccurate path planning and poor real-time obstacle avoidance in existing autonomous driving robots in complex dynamic environments. It utilizes multimodal sensor fusion to collect environmental data and employs a spatial mapping algorithm to process point cloud and image information, constructing a 3D environmental model that includes the locations of dynamic obstacles. This significantly improves the accuracy and real-time performance of environmental perception. This enables the robot to more accurately identify surrounding obstacles, especially dynamic objects, laying a reliable foundation for subsequent path planning. Second, an improved neural network model is used for intelligent obstacle identification and classification. Combined with a priority evaluation mechanism, high-priority interference is dynamically marked based on the distance and relative speed between the obstacle and the robot, generating a sorted obstacle list. This design effectively solves the deficiency of traditional path planning in ignoring the movement trends of dynamic obstacles, allowing the system to prioritize potential collision risks and improve the initiative and safety of obstacle avoidance. Furthermore, this invention uses a path planning algorithm to generate alternative paths and comprehensively evaluates path length and obstacle avoidance probability. An adaptive filtering method is used to predict potential collision points, and the path sequence is corrected in real time. This dynamic adjustment mechanism ensures the feasibility and optimality of the path, reducing path failures caused by environmental changes. Meanwhile, by monitoring execution deviations in real time and updating control commands, the system can quickly respond to external disturbances, ensuring the stability and accuracy of robot movement. Finally, by fusing sensor feedback data to continuously optimize the 3D environment model and iteratively executing path planning, this invention achieves a high degree of adaptability to dynamic environments, generating inspection paths that ultimately adapt to complex scenarios. Overall, this system not only improves the intelligence and robustness of path planning but also significantly enhances the robot's operating efficiency and safety in dynamic environments, making it suitable for various application scenarios such as warehousing and logistics. Attached Figure Description

[0015] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained from these drawings without creative effort.

[0016] Figure 1 This is one of the flowcharts for an intelligent path planning and control system for an autonomous driving transport robot according to the present invention.

[0017] Figure 2 This is the second flowchart of an intelligent path planning and control system for an autonomous driving transport robot according to the present invention.

[0018] Figure 3 This is the third flowchart of an intelligent path planning and control system for an autonomous driving transport robot according to the present invention. Detailed Implementation

[0019] The present invention will be further described below with reference to embodiments. These embodiments are merely some, not all, of the embodiments of the present invention. Other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are all within the protection scope of the present invention.

[0020] Please see Figures 1-3 As shown in the figure, this embodiment discloses an intelligent path planning and control system for an autonomous driving robot provided by the present invention. Its core lies in acquiring environmental data through multimodal sensors and combining it with an improved neural network model and path planning algorithm to achieve accurate perception, real-time decision-making, and efficient obstacle avoidance in dynamic environments. The specific implementation of this system will be described in detail below.

[0021] The system comprises modules for data acquisition and modeling, obstacle recognition and classification, priority evaluation, path planning, collision prediction and correction, control command generation and updating, and environmental model optimization. These modules cooperate with each other through data flow and signal transmission to complete the entire path planning and control process. Starting with the data acquisition and modeling module, this module collects data from the surrounding environment using multimodal sensors, acquiring raw information including LiDAR point clouds and camera images. The LiDAR point clouds are used to obtain spatial structure information of the environment, while the camera images are used to extract visual features. The LiDAR point clouds are processed by a spatial registration algorithm to form preliminary spatial structure data, while the camera images are processed using a visual mapping algorithm to extract features and perform geometric projection to generate visual mapping data. These two sets of data are integrated into fused environmental data using multi-source data fusion technology. The locations of dynamic obstacles are filtered using preset rules to generate filtered obstacle distribution data. Finally, a 3D environmental model containing the locations of dynamic obstacles is constructed, and the model content is adjusted in real time according to the movement trajectory of the dynamic obstacles.

[0022] Next is the obstacle recognition and classification module. This module extracts obstacle features based on a 3D environment model and identifies and classifies obstacles using an improved neural network model. The improved neural network model is pre-trained to distinguish between different types of obstacles, such as pedestrians and vehicles. During the recognition process, the dynamic displacement features of each object in the potential obstacle object set are first extracted. If dynamic displacement is detected, it is classified as a pedestrian and forms a dynamic object subset. For static objects, the continuity features of their movement trajectory are calculated based on historical location data. If the position does not change and meets the characteristics of a vehicle, it is classified as a vehicle and forms a static obstacle subset. The dynamic object subset and the static obstacle subset are integrated to form a complete obstacle dataset. This dataset is used to update the 3D environment model and save the classification results for subsequent use.

[0023] The priority evaluation module is responsible for calculating the distance and relative velocity between obstacles and the robot's current position based on the categorized obstacle dataset. The distance and velocity calculations are based on obstacle coordinates and the robot's real-time position data. If the distance to an obstacle is less than a preset threshold and its relative velocity is greater than zero, it is marked as a high-priority obstacle. All obstacles are sorted according to their high-priority identifiers to generate an initial priority ranking list. Subsequently, the distance and velocity status of each obstacle are re-verified using dynamic updates to position and velocity data to determine if the high-priority conditions are still met, generating an updated priority ranking result. The final priority ranking list contains detailed motion trajectory data, used to analyze potential collision risks and guide subsequent path planning.

[0024] The path planning module receives a priority ranking list and generates a set of candidate paths based on map data in the current map library. The path generation model constructs the candidate path set based on an obstacle distribution map and path planning methods, and evaluates each path. The evaluation includes path length and obstacle avoidance probability; if the path length exceeds a preset threshold, the obstacle avoidance probability is further analyzed. The candidate path set is sorted using a sequence ranking method, combining the path length and obstacle avoidance probability, to generate a preliminary preferred path sequence. Subsequently, the preferred path sequence is validated a second time using updated dynamic information from the obstacle distribution map to determine if adjustments are needed and to generate a final preferred path sequence. If the obstacle avoidance probability in the final preferred path sequence is lower than a preset threshold, the candidate path set is regenerated, and path evaluation and sequence ranking are repeated until the requirements are met.

[0025] The collision prediction and correction module simulates the robot's trajectory based on a preferred path sequence and uses an adaptive filtering method to predict potential collision points. If a predicted point overlaps with an obstacle, key control points in the path parameters are extracted, and the range of parameters requiring adjustment is determined. Adjustment rules are used to optimize the parameters, generating a first corrected path sequence, and the robot's trajectory is re-simulated. If potential collision points still exist in the updated trajectory point distribution, the local weights of the path parameters are adjusted for the new collision point locations to generate a second corrected path sequence. Finally, the collision risk of the second corrected path sequence is verified using an adaptive filtering method. If no collision risk is found, the second corrected path sequence is output as the final path.

[0026] The control command generation and update module obtains initial data from the path sequence and transforms it into a control command sequence using a pre-established mapping model. This control command sequence is sent to the robot execution module via a transmission channel, and status data during execution is collected in real time. A real-time monitoring mechanism compares the execution with a preset execution standard, calculates the execution deviation, and determines whether the deviation exceeds a preset threshold. If the deviation exceeds the threshold, a command update process is triggered, combining execution feedback data and path correction logic to regenerate an adjusted control command sequence. The adjusted control command sequence updates the command input to the robot execution module and continuously monitors the execution effect until the deviation stabilizes.

[0027] The environmental model optimization module collects real-time environmental information through sensors and fuses feedback data from sensors and historical records to construct an initial dataset. A Kalman filter performs data alignment and correction operations to determine the updated environmental model basis. The A* algorithm performs preliminary path calculations to generate an initial inspection path plan and uses OpenCV to perform path feasibility analysis based on dynamic environmental change information. If obstacles are detected or environmental changes exceed preset thresholds, path parameters are readjusted to generate an optimized path plan. Finally, real-time control commands are combined to dynamically adjust the execution details of the inspection path to determine the real-time adaptive scheme for the inspection task.

[0028] The collaboration between these modules is achieved through data flow and signal transmission. For example, the 3D environment model generated by the data acquisition and modeling module is directly input into the obstacle recognition and classification module for obstacle feature extraction and classification. The classified obstacle dataset is then passed to the priority evaluation module to generate a priority ranking list for the path planning module to use. The optimal path sequence generated by the path planning module is passed to the collision prediction and correction module for trajectory simulation and collision prediction. The corrected path sequence is then passed to the control command generation and update module to generate a control command sequence and monitor execution deviations in real time. Finally, the environment model optimization module continuously optimizes the 3D environment model and iterates path planning based on sensor feedback data to adapt to dynamic environmental changes.

[0029] In a real-world application scenario, an autonomous transport robot is deployed in a large logistics warehouse to perform cargo transportation tasks. The warehouse contains various dynamic obstacles, such as moving forklifts and pedestrians, as well as static obstacles such as shelves and pillars. After the robot starts, the data acquisition and modeling module collects warehouse environmental data using LiDAR and cameras and generates a 3D environment model. The obstacle recognition and classification module identifies and classifies obstacles based on the 3D environment model, classifying moving forklifts as vehicles and walking workers as pedestrians. The priority evaluation module generates a priority ranking list based on the distance and relative speed between obstacles and the robot, and transmits information about high-priority obstacles to the path planning module. The path planning module combines the warehouse map to generate a set of alternative paths and evaluates the length and obstacle avoidance probability of each path to generate an optimal path sequence. The collision prediction and correction module simulates the robot's trajectory, predicts potential collision points, and adjusts path parameters to generate a corrected path sequence. The control command generation and update module converts the corrected path sequence into a control command sequence and monitors execution deviations in real time to ensure the robot travels along the planned path. The environment model optimization module continuously optimizes the 3D environment model and iterates path planning based on sensor feedback data to cope with changes in the dynamic environment within the warehouse.

[0030] The above embodiments describe in detail the specific operating principle and process of the present invention, covering the complete process from data acquisition to path planning and control command generation. The collaborative relationships between the modules are clear and explicit, effectively solving the technical problems of achieving accurate perception, real-time decision-making, and efficient obstacle avoidance in complex dynamic environments.

[0031] To enable those skilled in the art to fully understand and implement this invention, the specific implementation principles of this invention are further supplemented below with a specific application scenario.

[0032] When deploying autonomous transport robots to perform cargo transportation tasks in large logistics warehouses, the data acquisition and modeling module is activated first. This module collects environmental information through LiDAR and cameras. The LiDAR point cloud is processed by a spatial registration algorithm to generate preliminary spatial structure data, while the camera images are processed by a visual mapping algorithm to extract features and perform geometric projection to generate visual mapping data. Subsequently, these two sets of data are integrated into fused environmental data through multi-source data fusion technology, and filtered obstacle distribution data is generated based on the location of dynamic obstacles. Finally, a 3D environmental model containing the locations of dynamic obstacles is constructed. This model can reflect the distribution of static obstacles such as shelves and pillars, as well as dynamic obstacles such as moving forklifts and pedestrians in the warehouse in real time.

[0033] Next, the obstacle recognition and classification module extracts obstacle features based on the 3D environment model and identifies and classifies obstacles using an improved neural network model. The improved neural network model is pre-trained to distinguish between different types of obstacles, such as pedestrians and vehicles. During actual operation, the dynamic displacement features of each object in the potential obstacle object set are extracted. If dynamic displacement is detected, it is classified as a pedestrian and forms a dynamic object subset. For static objects, the continuity features of their movement trajectory are calculated based on historical location data. If the position remains unchanged and meets the characteristics of a vehicle, it is classified as a vehicle and forms a static obstacle subset. The dynamic object subset and the static obstacle subset are integrated to form a complete obstacle dataset, which is used to update the 3D environment model and save the classification results for subsequent use.

[0034] The priority evaluation module calculates the distance and relative velocity between obstacles and the robot's current position based on the categorized obstacle dataset. The distance and velocity calculations are based on obstacle coordinates and the robot's real-time position data. If the distance to an obstacle is less than a preset threshold and its relative velocity is greater than zero, it is marked as a high-priority disturbance. All obstacles are sorted according to their high-priority identifiers to generate an initial priority ranking list. The module then re-verifies the distance and velocity status of each obstacle by dynamically updating the position and velocity data to determine if the high-priority conditions are still met and generates an updated priority ranking result. The final priority ranking list contains detailed motion trajectory data, used to analyze potential collision risks and guide subsequent path planning.

[0035] The path planning module receives a priority ranking list and generates a set of candidate paths based on map data in the current map library. The path generation model constructs this set of candidate paths based on an obstacle distribution map and path planning methods, and evaluates each path. The evaluation includes path length and obstacle avoidance probability; if the path length exceeds a preset threshold, the obstacle avoidance probability is further analyzed. The candidate path set is then sorted using a sequence ranking method, combining the path length and obstacle avoidance probability, to generate a preliminary preferred path sequence. Subsequently, the preferred path sequence is validated a second time using updated dynamic information from the obstacle distribution map to determine if adjustments are needed and to generate a final preferred path sequence. If the obstacle avoidance probability in the final preferred path sequence is lower than a preset threshold, the candidate path set is regenerated, and path evaluation and sequence ranking are repeated until the requirements are met.

[0036] The collision prediction and correction module simulates the robot's trajectory based on a preferred path sequence and uses an adaptive filtering method to predict potential collision points. If a predicted point overlaps with an obstacle, key control points in the path parameters are extracted, and the range of parameters requiring adjustment is determined. Adjustment rules optimize the parameters to generate a first corrected path sequence, and the robot's trajectory is re-simulated. If potential collision points still exist in the updated trajectory point distribution, the local weights of the path parameters are adjusted for the new collision point locations to generate a second corrected path sequence. Finally, the collision risk of the second corrected path sequence is verified using an adaptive filtering method. If no collision risk is found, the second corrected path sequence is output as the final path.

[0037] The control command generation and update module obtains initial data from the path sequence and transforms it into a control command sequence using a pre-established mapping model. The control command sequence is sent to the robot execution module via a transmission channel, and status data during execution is collected in real time. A real-time monitoring mechanism compares the execution deviation with a preset execution standard, calculates the deviation value, and determines whether the deviation exceeds a preset threshold. If the deviation exceeds the threshold, a command update process is triggered, combining execution feedback data and path correction logic to regenerate an adjusted control command sequence. The adjusted control command sequence updates the command input to the robot execution module and continuously monitors the execution effect until the deviation stabilizes.

[0038] The environmental model optimization module collects real-time environmental information through sensors and fuses feedback data from sensors and historical records to construct an initial dataset. A Kalman filter performs data alignment and correction operations to determine the updated environmental model basis. The A* algorithm performs preliminary path calculations to generate an initial inspection path plan and uses OpenCV to perform path feasibility analysis based on dynamic environmental change information. If obstacles are detected or environmental changes exceed preset thresholds, path parameters are readjusted to generate an optimized path plan. Finally, real-time control commands are combined to dynamically adjust the execution details of the inspection path to determine the real-time adaptive scheme for the inspection task.

[0039] The collaboration between these modules is achieved through data flow and signal transmission. For example, the 3D environment model generated by the data acquisition and modeling module is directly input into the obstacle recognition and classification module for obstacle feature extraction and classification. The classified obstacle dataset is then passed to the priority evaluation module to generate a priority ranking list for the path planning module to use. The optimal path sequence generated by the path planning module is passed to the collision prediction and correction module for trajectory simulation and collision prediction. The corrected path sequence is then passed to the control command generation and update module to generate a control command sequence and monitor execution deviations in real time. Finally, the environment model optimization module continuously optimizes the 3D environment model and iterates path planning based on sensor feedback data to adapt to dynamic environmental changes.

[0040] In some embodiments, the system operates by the following method, and the specific implementation of the system will be described in detail step by step below; S101. Collect environmental data through multimodal sensors, process lidar point clouds and camera images, fuse multi-source information to generate a real-time map, and obtain a three-dimensional environmental model containing the location of dynamic obstacles.

[0041] Data is collected from the site environment using multimodal sensors (such as LiDAR and cameras) to obtain raw information including LiDAR point clouds and camera images. The LiDAR point clouds are used to acquire spatial structure information of the environment, while the camera images are used to extract visual features. The LiDAR point clouds are processed by a spatial registration algorithm to form preliminary spatial structure data, while the camera images undergo feature extraction and geometric projection using a visual mapping algorithm to generate visual mapping data. These two sets of data are integrated into fused environmental data using multi-source data fusion technology. The locations of dynamic obstacles are filtered using preset rules to generate filtered obstacle distribution data. Finally, a 3D environmental model containing the locations of dynamic obstacles is constructed, and the model content is adjusted in real time according to the movement trajectory of the dynamic obstacles.

[0042] For example, suppose an autonomous transport robot is deployed in a large logistics warehouse. The LiDAR uses a Velodyne VLP-16 sensor to collect point cloud data with a density of 300,000 points per second and a coverage range of 100 meters. The camera uses 1080p resolution at a frame rate of 30fps. The LiDAR point cloud is spatially registered using the ICP (Iterative Closest Point) algorithm to correct coordinate deviations and generate preliminary spatial structure data. The camera images are used for target detection using the YOLOv5 model to extract feature points and generate visual mapping data through stereo vision geometric projection. Multi-source data fusion uses a Kalman filter algorithm to integrate point cloud and image data to generate fused environmental data. If a dynamic obstacle (such as a moving forklift) is detected, it is filtered according to preset rules (such as the obstacle being less than 5 meters from the robot) to generate obstacle distribution data. The final 3D environment model is stored in the form of a point cloud mesh with a mesh resolution of 0.1 meters and the dynamic obstacle trajectory is updated in real time.

[0043] S102. Based on the three-dimensional environment model, extract obstacle features, use an improved neural network model to identify several current obstacles, determine the obstacle type and movement trajectory, and obtain a classified obstacle dataset.

[0044] Obstacle features are extracted based on a 3D environment model, and an improved neural network model is used to identify and classify obstacles. The improved neural network model is pre-trained to distinguish between different types of obstacles, such as pedestrians and vehicles. During the identification process, the dynamic displacement features of each object in the potential obstacle object set are first extracted. If dynamic displacement is detected, it is classified as a pedestrian and forms a dynamic object subset. For static objects, the continuity features of their movement trajectory are calculated based on historical location data. If the position remains unchanged and meets the characteristics of a vehicle, it is classified as a vehicle and forms a static obstacle subset. The dynamic object subset and the static obstacle subset are integrated to form a complete obstacle dataset. This dataset is used to update the 3D environment model and save the classification results for subsequent use.

[0045] For example, in a logistics warehouse scenario, the 3D environment model includes shelves (static obstacles) and moving forklifts (dynamic obstacles). The improved neural network model uses the PointNet++ architecture, taking point cloud data as input and outputting obstacle categories and bounding boxes. The model is trained using the COCO dataset for transfer learning, achieving a recognition accuracy of 95%. Feature extraction includes the obstacle's size, shape, and speed. For example, a typical forklift is 2m × 1m × 1.5m in size and moves at a speed less than 2m / s; a typical pedestrian is 0.5m × 0.5m × 1.7m in size and moves at a speed less than 1.5m / s. If an object is detected with continuous displacement (e.g., displacement change greater than 0.1 m / s), it is classified as a pedestrian; if the object's position is stable and its size matches vehicle characteristics, it is classified as a forklift. Historical location data is analyzed over time series to calculate the acceleration and direction of the movement trajectory, forming an obstacle dataset. The dataset is stored in JSON format and includes obstacle ID, type, location, speed, and trajectory information.

[0046] S103. Using the classified obstacle dataset, calculate the distance and relative speed between the obstacle and the robot's current position. If the distance is less than a preset threshold and the speed is greater than zero, mark it as a high-priority interference and obtain a priority-ranked list of obstacles.

[0047] Based on the categorized obstacle dataset, the Euclidean distance and relative velocity between each obstacle and the robot's current position are calculated. The distance and velocity calculations are based on obstacle coordinates and the robot's real-time position data. If an obstacle's distance is less than a preset threshold and its relative velocity is greater than zero, it is marked as a high-priority obstacle. All obstacles are sorted according to their high-priority identifiers to generate an initial priority ranking list. Subsequently, the distance and velocity status of each obstacle are re-verified using dynamic updates to position and velocity data to determine if the high-priority condition is still met, generating an updated priority ranking result. The final priority ranking list contains detailed motion trajectory data, used to analyze potential collision risks and guide subsequent path planning.

[0048] For example, assume the robot's current position is (0,0,0), obstacle A (pedestrian) is located at (3,2,0) with a velocity vector of (0.5,0,0) m / s, and obstacle B (forklift) is located at (5,5,0) with a velocity vector of (1,0,0) m / s. Distance is calculated using the Euclidean algorithm, with the distance to obstacle A being... ≈3.6 meters, the distance to obstacle B is ≈7.1 meters. The preset distance threshold is 5 meters, and the speed threshold is 0. Obstacle A, with a distance less than 5 meters and a speed greater than 0, is marked as high priority; obstacle B, with a distance greater than 5 meters, is marked as low priority. The priority ranking list is sorted in ascending order of distance: Obstacle A (high priority), Obstacle B (low priority). After real-time updates, if the distance of obstacle A changes to 4 meters and the speed remains greater than 0, it retains its high priority; if the distance of obstacle B changes to 4.5 meters, it is upgraded to high priority. The list is maintained using a priority queue data structure for use by the path planning module.

[0049] S104. The obstacle list sorted by priority is input into the path planning algorithm. The algorithm generates alternative paths in combination with the current map, evaluates the length and obstacle avoidance probability of each path, and obtains the optimal path sequence.

[0050] The path planning module receives a priority ranking list and generates a set of candidate paths based on map data in the current map library. The path generation model constructs the candidate path set based on an obstacle distribution map and a path planning algorithm (such as the A* algorithm) and evaluates each path. The evaluation includes path length and obstacle avoidance probability; if the path length exceeds a preset threshold, the obstacle avoidance probability is further analyzed. The candidate path set is sorted using a sequence sorting method, combining the path length and obstacle avoidance probability, to generate a preliminary preferred path sequence. Subsequently, the preferred path sequence is validated a second time using updated obstacle distribution map information to determine if adjustments are needed and to generate a final preferred path sequence. If the obstacle avoidance probability in the final preferred path sequence is lower than a preset threshold, the candidate path set is regenerated, and path evaluation and sequence sorting are repeated until the requirements are met.

[0051] For example, in a logistics warehouse map, the robot travels from the starting point (0,0) to the ending point (10,10). The A* algorithm generates three candidate paths: Path 1 is 12 meters long with an obstacle avoidance probability of 0.9; Path 2 is 10 meters long with an obstacle avoidance probability of 0.7; and Path 3 is 15 meters long with an obstacle avoidance probability of 0.95. The preset path length threshold is 20 meters, and the obstacle avoidance probability threshold is 0.8. Path evaluation uses a weighted score: score = 0.6 × obstacle avoidance probability + 0.4 × (1 / path length). Path 1 score = 0.6 × 0.9 + 0.4 × (1 / 12) ≈ 0.593; Path 2 score = 0.6 × 0.7 + 0.4 × (1 / 10) = 0.46; Path 3 score = 0.6 × 0.95 + 0.4 × (1 / 15) ≈ 0.597. The preliminary preferred sequence is Path 3, Path 1, and Path 2. During the second verification, if a new obstacle appears on path 3 and the obstacle avoidance probability drops to 0.6, the order is reordered, and path 1 becomes the preferred choice. The final preferred sequence is path 1, path 3, and path 2.

[0052] S105. By optimizing the path sequence, the robot's motion trajectory is simulated, and an adaptive filtering method is used to predict potential collision points. If the predicted point overlaps with the obstacle, the sequence parameters are adjusted to obtain a corrected path sequence.

[0053] The robot's trajectory is simulated based on an optimized path sequence, and potential collision points are predicted using an adaptive filtering method (such as Kalman filtering). If a predicted point overlaps with an obstacle, key control points in the path parameters are extracted, and the range of parameters requiring adjustment is determined. Adjustment rules are used to optimize the parameters, generating a first corrected path sequence, and the robot's trajectory is re-simulated. If potential collision points still exist in the updated trajectory point distribution, the local weights of the path parameters are adjusted for the new collision point locations to generate a second corrected path sequence. Finally, the collision risk of the second corrected path sequence is verified using an adaptive filtering method. If no collision risk is found, the second corrected path sequence is output as the final path.

[0054] For example, the robot's motion trajectory simulation uses a bicycle model with a prediction time step of 0.1 seconds. Adaptive Kalman filtering is used to estimate the trajectory point positions, and the covariance matrix is ​​initialized to the identity matrix. Assuming that trajectory point P(5,5) on path 1 overlaps with the predicted position of obstacle A, control point C(4.5,4.8) is extracted, and the parameter is adjusted to the path curvature from 0.1 to 0.05, generating the first corrected path. After resimulation, trajectory point P(5.2,5.1) still overlaps with obstacle A, so the local weights are adjusted, increasing the obstacle avoidance weight of obstacle A from 0.5 to 0.7, generating the second corrected path. In the final verification, the trajectory points do not overlap with the obstacles, and the collision risk is eliminated. The corrected path sequence is output in the form of a B-spline curve with 10 control points.

[0055] S106. Based on the corrected path sequence, generate a control command sequence, transmit it to the robot actuator, monitor the execution deviation in real time, determine whether the deviation exceeds the threshold, and if it does, recalculate the command to obtain the updated control command.

[0056] Initial data is obtained from the corrected path sequence, and the path sequence is transformed into a control command sequence using a pre-established mapping model (such as differential flatness mapping). The control command sequence is sent to the robot execution module via a transmission channel (such as a CAN bus), and status data during execution is collected in real time. A real-time monitoring mechanism compares the execution deviation value with a preset execution standard and determines whether the deviation exceeds a preset threshold. If the deviation exceeds the threshold, an instruction update process is triggered, and the adjusted control command sequence is regenerated by combining the execution feedback data and the path correction logic. The adjusted control command sequence updates the instruction input of the robot execution module and continuously monitors the execution effect until the deviation stabilizes. For example, the path sequence is converted into a control command sequence, including linear velocity v and angular velocity ω. The mapping model is v = path tangent direction, ω = path curvature × v. Control commands are sent via the CAN bus at a baud rate of 500kbps. Real-time monitoring uses IMU and encoder data to calculate position deviation. The preset deviation threshold is 0.1 meters. If the actual position (5.1, 5.1) deviates from the expected position (5, 5) by 0.14 meters, exceeding the threshold, the command is recalculated, adjusting v from 1 m / s to 0.8 m / s and ω from 0.1 rad / s to 0.12 rad / s. After the updated command is sent, the deviation drops to 0.05 meters, reaching stability. The command update cycle is 100 ms.

[0057] S107. By updating the control commands and integrating sensor feedback data, the three-dimensional environment model is continuously optimized, and the path planning algorithm is used to iterate the path planning to obtain the final inspection path that adapts to the dynamic environment.

[0058] Real-time environmental information is collected by sensors (such as LiDAR and cameras), and an initial dataset is constructed by fusing sensor feedback data with historical data. A Kalman filter performs data alignment and correction operations to determine the basis of the updated environmental model. The A* algorithm performs preliminary path calculations to generate an initial inspection path scheme, and OpenCV is used to perform path feasibility analysis based on dynamic environmental change information. If obstacles are detected or environmental changes exceed preset thresholds, path parameters are readjusted to generate an optimized path scheme. Finally, the execution details of the inspection path are dynamically adjusted in conjunction with real-time control commands to determine the real-time adaptive scheme for the inspection task.

[0059] For example, sensor feedback data includes LiDAR point clouds and camera images, updated at a frequency of 10Hz. A Kalman filter is used to fuse multi-source data, with a state vector of [position, velocity] and an identity matrix. The A* algorithm calculates the initial path, using the Manhattan distance as the heuristic function. OpenCV is used for path feasibility analysis and obstacle contour detection. If a new obstacle is detected within 0.5 meters of the path, the path parameters are readjusted, such as increasing the safety distance from 0.3 meters to 0.5 meters. The optimized path is executed via control commands, and the robot ultimately completes the inspection task along the path, adapting to dynamic environmental changes.

[0060] The system module collaboration relationships are as follows: The above steps correspond to the system's seven modules: data acquisition and modeling, obstacle identification and classification, priority evaluation, path planning, collision prediction and correction, control command generation and updating, and environment model optimization. These modules collaborate through data flow and signal transmission. For example, the 3D environment model generated by the data acquisition and modeling module is directly input into the obstacle identification and classification module for obstacle feature extraction and classification. The classified obstacle dataset is then passed to the priority evaluation module to generate a priority ranking list for the path planning module to use. The optimal path sequence generated by the path planning module is passed to the collision prediction and correction module for trajectory simulation and collision prediction. The corrected path sequence is then passed to the control command generation and updating module to generate a control command sequence and monitor execution deviations in real time. Finally, the environment model optimization module continuously optimizes the 3D environment model and iterates path planning based on sensor feedback data to adapt to dynamic environmental changes. In some alternative embodiments, in practical application scenarios, it is assumed that an autonomous transport robot is deployed in a large logistics warehouse to perform cargo transportation tasks. The warehouse contains various dynamic obstacles such as moving forklifts and pedestrians, as well as static obstacles such as shelves and pillars. After startup, the robot executes the complete process from S101 to S107: collecting environmental data through multimodal sensors and generating a 3D environmental model; identifying and classifying obstacles; calculating obstacle priorities; generating and optimizing paths; predicting and correcting collision risks; generating and controlling commands; and ultimately achieving safe inspection in a dynamic environment. The entire process is achieved through automated algorithms, requiring no human intervention, ensuring the efficiency and safety of the transportation task.

[0061] The above embodiments describe in detail the specific operating principle and process of the present invention, covering the complete process from data acquisition to path planning and control command generation. The logical relationships between each step are clear and explicit, effectively solving the technical problems of achieving accurate perception, real-time decision-making, and efficient obstacle avoidance in complex dynamic environments.

[0062] All content not described in detail in this specification is prior art known to those skilled in the art, and the model parameters of each electrical appliance are not specifically limited; conventional equipment can be used. Electrical control components not mentioned in this technical solution are not shown in the figures because they are prior art, and will not be described further here.

[0063] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An intelligent path planning and control system for an autonomous driving transport robot, characterized in that, include: By collecting environmental data through multimodal sensors, processing lidar point clouds and camera images, and fusing multi-source information to generate a real-time map, a three-dimensional environmental model containing the location of dynamic obstacles is obtained. Based on the 3D environment model, obstacle features are extracted, and an improved neural network model is used to identify several current obstacles, determine the obstacle type and movement trajectory, and obtain a classified obstacle dataset. Using the categorized obstacle dataset, the distance and relative speed between the obstacle and the robot's current position are calculated. If the distance is less than a preset threshold and the speed is greater than zero, it is marked as a high-priority obstacle, and a priority-ranked list of obstacles is obtained. The obstacle list, sorted by priority, is input into the path planning algorithm. The algorithm generates alternative paths based on the current map, evaluates the length and obstacle avoidance probability of each path, and obtains the optimal path sequence. By optimizing the path sequence, simulating the robot's motion trajectory, and using an adaptive filtering method to predict potential collision points, if the predicted point overlaps with an obstacle, the sequence parameters are adjusted to obtain a corrected path sequence. Based on the corrected path sequence, a control command sequence is generated and transmitted to the robot actuator. The execution deviation is monitored in real time to determine whether the deviation exceeds the threshold. If it does, the command is recalculated to obtain the updated control command. By updating control commands, integrating sensor feedback data, continuously optimizing the 3D environment model, and iterating path planning, an inspection path that adapts to the dynamic environment is finally obtained.

2. The intelligent path planning and control system for an autonomous driving transport robot according to claim 1, characterized in that: Environmental data is collected using multimodal sensors, and LiDAR point clouds and camera images are processed using spatial mapping algorithms. Multi-source information is fused to generate a real-time map, resulting in a 3D environmental model that includes the locations of dynamic obstacles. Data is collected from the field environment using multimodal sensors to obtain raw information including lidar point clouds and camera images, thus obtaining preliminary environmental perception data. Based on the collected lidar point cloud, a spatial registration algorithm is used to perform coordinate correction on the point cloud data, determine the position information in the point cloud, and form preliminary spatial structure data. By using camera images and combining visual mapping algorithms to extract features and perform geometric projection on the image content, visual mapping data corresponding to the LiDAR point cloud is obtained, and the basic content of the real-time map is constructed. By employing multi-source data fusion technology, spatial structure data and visual mapping data are integrated and processed to obtain fused environmental data that includes the location of dynamic obstacles; If dynamic obstacles are detected in the fused environmental data, the obstacle locations are filtered according to preset filtering rules to determine whether the obstacles are in critical areas and generate filtered obstacle distribution data. Based on the filtered obstacle distribution data, a three-dimensional environment model containing the dynamic obstacle locations is constructed to obtain complete environmental description information; By continuously updating the 3D environment model and adjusting the movement trajectory of dynamic obstacles in real time, the final dynamic environmental information is determined.

3. The intelligent path planning and control system for an autonomous driving transport robot according to claim 1, characterized in that: Based on the 3D environment model, obstacle features are extracted, and an improved neural network model is used to identify pedestrians and vehicles, determine obstacle types and movement trajectories, and obtain a classified obstacle dataset, including: By acquiring 3D environmental data, an initial environmental model is constructed, and multi-source data fusion technology is used to integrate multi-source information to obtain preliminary environmental model data. Based on preliminary environmental model data, obstacle features are extracted, and a pre-trained improved neural network model is applied to perform preliminary identification of objects in the environment to determine the set of potential obstacle objects. For a set of potential obstacle objects, analyze the distinguishing features between pedestrians and vehicles. If an object is detected to have dynamic displacement, it is classified as a pedestrian, and a subset of dynamic objects is obtained after classification. Based on the categorized subset of dynamic objects and combined with historical location data, the continuity characteristics of the motion trajectory are calculated, and the trajectory analysis method is used to determine the motion trajectory pattern of the dynamic objects. By using motion trajectory patterns, static objects are reconfirmed. If the position of a static object does not change and it meets the characteristics of a vehicle, it is classified as a vehicle, and a subset of static obstacles is determined. Based on the dynamic object subset and the static obstacle subset, obstacle type information is integrated, and classification and organization techniques are applied to construct a complete obstacle dataset. By using a complete obstacle dataset, the 3D environment model is updated, and data storage technology is used to save the classification results, thereby obtaining real-time updated environmental information.

4. The intelligent path planning and control system for an autonomous driving transport robot according to claim 1, characterized in that: Using the categorized obstacle dataset, the distance and relative velocity between the obstacle and the robot's current position are calculated. If the distance is less than a preset threshold and the velocity is greater than zero, it is marked as a high-priority obstacle, resulting in a priority-ranked list of obstacles, including: By extracting the coordinate and motion information of each obstacle from the obstacle dataset and combining it with real-time data on the robot's position, the spatial distance and relative velocity between each obstacle and the robot are calculated to obtain preliminary distance and velocity results. Based on the calculated distance and speed results, condition judgment is performed for each obstacle. If the distance is lower than a preset threshold and the relative speed is greater than zero, it is marked as a high-priority interference, and a set of marked obstacle classifications is obtained. Using the labeled obstacle classification set, all obstacles are sorted according to the identifier of high-priority interference to generate an initial priority sorting list; By combining the initial priority ranking list with dynamic updates of position and velocity data, the distance and velocity status of each obstacle are re-verified to determine whether the high priority conditions are still met, and the updated priority ranking results are obtained. Based on the updated priority ranking results, detailed motion trajectory data of obstacles with high priority interference are obtained, potential collision risks are analyzed, and the final priority ranking list is determined. By combining the final priority ranking list with the changes in the robot's position and obstacle dataset acquired in real time, the fluctuations in distance and speed are continuously monitored to determine whether the priority needs to be adjusted, and a dynamically adjusted obstacle priority list is obtained. A dynamically adjusted obstacle priority list is used to generate structured data output, which is then called by the subsequent path planning module to determine the execution order of the real-time obstacle avoidance strategy.

5. The intelligent path planning and control system for an autonomous driving transport robot according to claim 1, characterized in that: The priority list of obstacles is input into the path planning algorithm. Combined with the current map, alternative paths are generated. The length and obstacle avoidance probability of each path are evaluated to obtain the optimal path sequence, including: An initial obstacle distribution map is generated by retrieving sorted data from the obstacle priority list and combining it with map data sources in the current map library. Based on the obstacle distribution map and path planning method, a set of alternative paths is constructed using a pre-established path generation model; For each path in the candidate path set, a path evaluation method is used to calculate the path length value and record the length data of each path; If the path length exceeds the preset threshold range, the obstacle avoidance probability value is further analyzed by probability calculation to obtain the obstacle avoidance probability data for each path. Based on the obstacle avoidance probability value and path length value, the candidate path set is comprehensively sorted using the sequence sorting method to determine the preliminary preferred path order; By performing secondary verification on the path data in the preferred path sequence and combining it with the dynamic information after the obstacle distribution map is updated, it is determined whether the path sequence needs to be adjusted, and the final preferred path sequence is obtained. If the obstacle avoidance probability value in the final preferred path sequence is lower than the preset threshold, then the candidate path set is regenerated by combining the latest map data source in the current map library, and the path evaluation method and sequence sorting method are executed cyclically to determine a new preferred path sequence.

6. The intelligent path planning and control system for an autonomous driving transport robot according to claim 1, characterized in that: By optimizing the path sequence and simulating the robot's trajectory, an adaptive filtering method is used to predict potential collision points. If the predicted point overlaps with an obstacle, the sequence parameters are adjusted to obtain a corrected path sequence, including: By optimizing the path sequence, the initial simulation data of the robot's motion trajectory is obtained. An adaptive filtering method is used to estimate the distribution of trajectory points and obtain the prediction results of potential collision points. Based on the prediction results of potential collision points, determine whether the predicted points overlap with obstacles. If the predicted points overlap with obstacles, extract the key control points in the path parameters and determine the range of parameters that need to be adjusted. By extracting the range of path parameters, the parameters are optimized using preset adjustment rules to generate the first corrected path sequence; Based on the first corrected path sequence, the robot's motion trajectory is re-simulated to obtain the updated trajectory point distribution and determine whether there are any new potential collision points. If potential collision points still exist in the updated trajectory point distribution, the local weights of the path parameters are adjusted for the new collision point locations to obtain the second corrected path sequence. The final robot motion trajectory data is obtained through the second corrected path sequence. An adaptive filtering method is used to verify the final trajectory and determine whether the collision risk has been eliminated. If no collision risk is found during the final trajectory verification, the second corrected path sequence will be used as the final path output to generate control commands for the robot's movement.

7. The intelligent path planning and control system for an autonomous driving transport robot according to claim 1, characterized in that: Based on the corrected path sequence, a control command sequence is generated and transmitted to the robot actuator. Execution deviations are monitored in real time, and it is determined whether the deviation exceeds a threshold. If it does, the commands are recalculated to obtain updated control commands, including: Initial data is obtained from the path sequence, and the path sequence is transformed into a control command sequence using a pre-established mapping model to obtain a preliminary set of control commands. The control command set is sent to the robot execution module through the transmission channel, and the status data during the execution process is collected in real time to determine the dynamic changes during the execution process; For dynamically changing data, a real-time monitoring mechanism is used to compare the data with preset execution standards, calculate the execution deviation value, and determine whether the deviation exceeds the preset threshold. If the deviation exceeds the preset threshold, the instruction update process is triggered, and the adjusted control instruction sequence is regenerated by combining the execution feedback data and the path correction logic. By adjusting the control command sequence, the command input of the robot execution module is updated, new execution status data is obtained, and the execution effect is continuously monitored. Based on continuously monitored data, the deviation judgment process is executed cyclically. If the deviation still exceeds the preset threshold, the instruction update and adjustment steps are repeated to obtain stable execution results. After the execution results stabilize, the final path sequence and the correspondence between the control commands are recorded and stored in the historical database for reference in subsequent tasks.

8. The intelligent path planning and control system for an autonomous driving transport robot according to claim 1, characterized in that: By updating control commands and fusing sensor feedback data, the 3D environment model is continuously optimized. A path planning algorithm is used for iterative path planning to obtain the final inspection path adapted to the dynamic environment, including: By collecting real-time environmental information through sensors and fusing feedback data from sensors and historical records, an initial dataset is constructed to obtain preliminary environmental perception results. Based on the preliminary environmental perception results and combined with the pre-established 3D model, a Kalman filter is used to perform data alignment and correction operations to determine the basis of the updated environmental model. Based on the updated environmental model, and incorporating real-time adjustments to control commands, the A* algorithm is applied to perform preliminary path calculations and obtain an initial inspection path scheme. For the initial inspection path plan, combined with the dynamic environment change information, OpenCV is used to perform path feasibility analysis. If obstacles or environmental changes exceed the preset threshold, the path parameters are readjusted to obtain an optimized path plan. By continuously collecting sensor feedback data through the optimized path scheme, updating the dynamic elements in the environmental model, and determining whether the path execution conditions are met; If the path execution conditions are met, the final inspection path data will be generated based on the updated environment model and the optimized path scheme. If the conditions are not met, return to the OpenCV-based path feasibility analysis stage and iteratively calculate the path scheme. After obtaining the final inspection path data, the execution details of the inspection path are dynamically adjusted in conjunction with real-time control commands to determine the real-time adaptive solution for the inspection task.

9. The intelligent path planning and control system for an autonomous driving transport robot according to claim 1, characterized in that: The system includes: The data acquisition and modeling module is configured to acquire field data through multimodal sensors, process lidar point clouds and camera images using spatial mapping algorithms, fuse multi-source information to generate a real-time map, and obtain a three-dimensional environment model containing the location of dynamic obstacles. The obstacle recognition and classification module is configured to extract obstacle features based on a 3D environment model, use an improved neural network model to identify pedestrians and vehicles, determine the obstacle type and movement trajectory, and obtain a classified obstacle dataset. The priority evaluation module is configured to calculate the distance and relative speed between the obstacle and the robot's current position using the classified obstacle dataset. If the distance is less than a preset threshold and the speed is greater than zero, it is marked as a high-priority interference, and a priority-ranked list of obstacles is obtained. The path planning module is configured to take a priority list of obstacles and input it into the path planning algorithm. It then combines the current map to generate alternative paths, evaluates the length and obstacle avoidance probability of each path, and obtains the optimal path sequence. The collision prediction and correction module is configured to simulate the robot's motion trajectory by optimizing the path sequence, predict potential collision points using an adaptive filtering method, and adjust the sequence parameters to obtain a corrected path sequence if the predicted point overlaps with an obstacle. The control command generation and update module is configured to generate a control command sequence based on the corrected path sequence, transmit it to the robot actuator, monitor the execution deviation in real time, determine whether the deviation exceeds the threshold, and if it does, recalculate the command to obtain the updated control command. The environmental model optimization module is configured to continuously optimize the 3D environmental model by integrating sensor feedback data with updated control commands, and to obtain the final inspection path that adapts to the dynamic environment by iteratively planning the path using a path planning algorithm.

Images