Multi-robot path planning method based on fusion algorithm

Abstract

The invention relates to the technical field of path planning, and particularly discloses a multi-robot path planning method based on a fusion algorithm, and the method comprises the steps: 1, obtaining environment data through a sensor, analyzing and integrating the environment data, and constructing an environment model in combination with task features; 2, decomposing the whole task into sub-tasks, and dynamically allocating the tasks to different robots according to geographic positions and priorities; 3, performing path planning by adopting a multi-robot path planning algorithm based on task priority, and optimizing the path by using a machine learning algorithm; step 4, adopting a hierarchical multi-sensor information fusion algorithm to enable the robot to track a path center line and to rapidly and effectively avoid obstacles; according to the method, the moving safety and efficiency of the robot in a complex environment are improved, the dynamic adaptive capacity and cooperation capacity of the robot are enhanced, and a solid foundation is provided for popularization of a multi-robot system in practical application.

Description

Technical Field

[0001] The present invention belongs to the technical field of path planning, and particularly relates to a multi-robot path planning method based on a fusion algorithm. Background Art

[0002] The flexibility and intelligence of robots enable them to be widely applied in various fields, including intelligent inspection, healthcare, smart home, etc. The inspiration for multi-robot swarm control comes from the behaviors of biological groups in nature, such as social animals, birds, fish, and cell and microorganism swarms. Multi-robot systems have simple local interaction rules, such as maintaining similar speeds, keeping stable distances, and avoiding collisions, to achieve self-organized collective behaviors. Through the cooperation between robots, multi-robot systems can improve the intelligence of individual robot behaviors, demonstrating stronger fault tolerance, adaptability, and resource utilization to cope with environmental changes and task requirements. When multi-robots are in formation to perform tasks, obstacles, especially dynamic obstacles, pose a serious threat to the safety of robots. Therefore, multi-robot systems need to have the ability of autonomous navigation and obstacle avoidance to ensure the safe completion of tasks.

[0003] How to plan a safe, efficient, and optimal path in real time in a complex environment is a difficult problem that urgently needs to be overcome in the current field of mobile robot path planning. However, a major challenge in the path planning of multi-robot systems is to enable multiple robots to coordinate and cooperate with each other to avoid collisions with each other. In the path planning problem of mobile robots, the path planning of multi-robot systems is undoubtedly a major difficulty. Most previous path planning studies have focused on static environments, ignoring the situation where there are dynamic obstacles in the environment or multiple mobile robots are working simultaneously. Therefore, it is necessary to propose a multi-robot path planning method based on a fusion algorithm to at least partially solve the problems existing in the prior art.

[0004] The above information disclosed in this background art is only used to increase the understanding of the background art of the present invention. Therefore, it may include prior art that is not known to those of ordinary skill in the art. Summary of the Invention

[0005] The purpose of the present invention is to provide a multi-robot path planning method based on a fusion algorithm to solve the problems raised in the above background art.

[0006] To achieve the above purpose, the present invention provides the following technical solutions: A multi-robot path planning method based on a fusion algorithm, comprising: Step 1: Use sensors to obtain environmental data, analyze and integrate the environmental data, and construct an environmental model in combination with task characteristics; Step 2: Decompose the overall task into subtasks and dynamically allocate the tasks to different robots according to geographical location and priority; Step 3: Adopt a multi-robot path planning algorithm based on task priority for path planning and use machine learning algorithms to optimize the path; Step 4: Adopt an algorithm for hierarchical multi-sensor information fusion to enable the robot to track the center line of the path and quickly and effectively avoid obstacles; Step 5: Conduct tests in a simulation environment, simulate the behaviors of multiple robots in different scenarios, and collect data for further analysis; Step 6: Analyze the time consumption, path quality, and task completion rate of path planning, optimize the path planning and strategies according to the analysis results, and improve the overall performance.

[0007] Preferably, for the analysis and integration of environmental data: perform information fusion at the data layer on the acquired original ultrasonic data to improve the accuracy of ultrasonic detection, and then perform information fusion on the environmental information obtained by image processing and the ultrasonic fusion data to verify each other to improve the accuracy and integrity of environmental data.

[0008] Preferably, for the construction of the environmental model: use the grid method to establish a multi-robot path planning map model suitable for the current environment, describe the environmental elements in the map model, and stipulate the performance status and behavioral characteristics of the environmental elements in multi-robot path planning.

[0009] Preferably, for the dynamic allocation of tasks to different robots according to geographical location and priority: build an efficient communication network through the IoT platform and the MQTT protocol, realize information sharing and collaboration among robots through parallel processing, monitor the status and collaboration progress of each robot, and adjust the strategy in a timely manner.

[0010] Preferably, for the path planning using a multi-robot path planning algorithm based on task priority: First, determine the factors affecting the path planning priority in the multi-robot system according to the task requirements of the current scenario; then, perform global path planning for each robot executing the task through the IoT platform; next, each robot moves along the global path, and at the same time senses the surrounding environment through the accompanying sensors to judge whether there are unexpected obstacles on the current task route. If there are unexpected obstacles, perform local path optimization; finally, use the robot's movement to the task end point as the task completion flag and the failure information output by the robot as the task abort flag, and both the completed and aborted tasks are removed from the execution list.

[0011] Preferably, the application The algorithm completes the multi-robot path planning task, and the algorithm always accesses the top task of the task list to ensure that tasks with higher priorities are planned first and occupy the global reservation table. Information is saved and shared among multiple robots through the global reservation table. A machine learning algorithm is introduced to train a model using historical data to predict collision risks and optimize paths.

[0012] Analyze using the determined factors that affect path planning priorities, and then sort the tasks in the task list by priority. Then use the determined task priorities as the order for the algorithm to perform path planning. Among them, except for special tasks, the sorting criteria for ordinary tasks are determined by the task requirements of the current scenario.

[0013] Preferably, for the single-robot global path planning: The robot starts from the task starting point, continuously traverses the set of adjacent reachable nodes of the current node, judges whether the robot has waiting behavior, turning behavior, and diagonal straight-line movement at the adjacent node, and selects the one with the smallest cost as the next moving position. After obtaining the global path of the current task, output its path, record the positions occupied by this path at each time interval in the global reservation table in the form of keys, and then delete the current task from the task list.

[0014] Waiting behavior: When there is a scenario where multiple robots meet at a certain position in the workspace and affect each other's passage, the path-finding method is that the robot with a higher priority occupies the conflict point first, and the robot with a lower priority waits in place in a loop until the occupation of the conflict point is released again. Considering the time cost brought by the waiting behavior of robots in multi-robot path planning, the in-place waiting cost of one party when multi-robot path conflicts occur is described as follows:

[0015] Among them, represents the waiting cost that the robot spends at node to avoid dynamic obstacles; is the number of waiting loops; is the single waiting time; is the weight coefficient of the waiting cost.

[0016] Turning behavior: There are motion and time costs for the turning action of the robot during the execution of the path. For paths of the same length, the path cost including the robot's turning is significantly higher than that of a straight path. Based on the judgment of turning, take the turning coefficient for adjacent nodes with vertical turning, and take the turning coefficient for adjacent nodes with diagonal turning. The turning cost is defined as follows:

[0017] Among them, Represents the cost generated by the robot's turning action at the node ; is the time cost parameter for the robot to turn; is the weight coefficient of the turning cost.

[0018] Diagonal straight line: Based on the principle of self - protection during path execution, the robot often needs to decelerate for movements outside this recommended path. Therefore, when planning the path, the robot "prefers" to move along horizontal or vertical straight lines and "rejects" diagonal straight lines accordingly. This "rejection" is reflected in the cost function, and the cost of diagonal straight line is described as:

[0019] where Represents the cost generated by the robot moving along the diagonal straight line at the node ; is the time cost parameter for the robot to move diagonally straight; is the weight coefficient of the turning cost; is the flag indicating whether the robot has the behavior of moving diagonally straight. If it exists, it takes 1, otherwise it takes 0.

[0020] Preferably, the algorithm for avoiding obstacles by using hierarchical multi - sensor information fusion is as follows: A neural network is used to perform data - layer information fusion on the distance and azimuth information detected by multiple ultrasonic ranging sensors to eliminate the uncertainty of sensor data; then a fuzzy neural network is used to perform decision - layer information fusion on the result obtained from the data - layer information fusion and the path center deviation information, so that the robot can effectively avoid obstacles according to the obtained control signal and continue to move along the planned path.

[0021] The neural network for data - layer information fusion adopts a three - layer forward neural network structure. The input data of the network is the distance data measured by three sensors of an ultrasonic sensor group , and and the calculated distance data , where ; the output is the distance of the obstacle measured by the robot (when , , representing the distance of the obstacle on the left; when , , representing the distance of the obstacle in front; when , , representing the distance of the obstacle on the right). Thus, it can be seen that the input nodes of the neural network are 4, the output node is 1, and according to Kolmogorov's theorem and simulation tests, the number of hidden - layer neuron nodes is determined to be 15.

[0022] The information fusion at the decision-making level adopts a fuzzy neural network structure that uses a neural network to learn the parameters of the membership function, including an input layer, fuzzification, fuzzy inference, defuzzification, and output layer. Here, the defuzzification and output layer are combined and optimized to form a four-layer neural network structure.

[0023] A terminal device includes a processor and a computer-readable storage medium. The processor is used to implement each instruction; the computer-readable storage medium is used to store multiple instructions, and the instructions are adapted to be loaded and executed by the processor to perform the multi-robot path planning method based on a fusion algorithm described in any one of the above.

[0024] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the processor implements the multi-robot path planning method based on a fusion algorithm as described above.

[0025] Compared with the prior art, the beneficial effects of the present invention are as follows: The present invention uses a multi-robot path planning algorithm based on task priority to complete the global path planning of each robot, ensuring that tasks with higher priority are planned first, which can effectively avoid collisions between robots, improve the safety of overall task completion, optimize the path through a machine learning algorithm, enabling the robot to find shorter and more efficient paths, reducing energy consumption and time costs, and using a hierarchical multi-sensor information fusion method to process real-time data, which can accurately locate the position of obstacles, enabling the robot to quickly and effectively avoid obstacles, ensuring the continuity and stability of the path. The present invention not only improves the safety and efficiency of the robot moving in a complex environment but also enhances its dynamic adaptability and collaboration ability, providing a solid foundation for the popularization of multi-robot systems in practical applications.

[0026] The above summary is only for the purpose of the specification and is not intended to be limiting in any way. In addition to the illustrative aspects, embodiments, and features described above, further directions, embodiments, and features of the present invention will be readily apparent by referring to the drawings and the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 It is a flowchart of the multi-robot path planning method based on a fusion algorithm of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0028] Next, in combination with the accompanying drawings in the embodiments of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0029] Embodiment 1:

[0030] Please refer to Figure 1 As shown, a multi-robot path planning method based on a fusion algorithm includes: performing path planning using a multi-robot path planning algorithm based on task priority, completing the multi-robot path planning task through the algorithm, and the algorithm always accesses the top task in the task list to ensure that tasks with higher priorities are planned first and occupy the global reservation table. Information is saved and shared among multiple robots through the global reservation table. A machine learning algorithm is introduced to train a model using historical data to predict collision risks and optimize paths; an algorithm for hierarchical multi-sensor information fusion is used to enable the robot to track the path centerline and quickly and effectively avoid obstacles.

[0031] The global reservation table is implemented in the data structure of a hash table. The occupancy situation of any position on the global map at time by a robot is stored in the hash table in the form of a key . Once the key is occupied by a robot , its corresponding value is stored as . According to the characteristics of the data structure of the hash table, the keys in the hash table are unique, that is, there will be no same position-time keys in the hash table. Once a certain position-time key is written by a robot , subsequent single-robot planning will not consider this key as passable, thus preventing other robots from repeatedly occupying a specific position at a specific time and avoiding multi-robot collisions.

[0032] Path planning: First, determine the factors affecting path planning priority in the multi-robot system according to the task requirements of the current scenario; then, perform global path planning for each robot executing the task through the IoT platform: decompose the task into a series of single-robot path planning. For each single-robot planning, the robot starts from the task starting point, continuously traverses the adjacent reachable node set of the current node, judges whether the robot has waiting behavior, turning behavior, and diagonal straight-line movement at the adjacent node, selects the one with the smallest cost as the next moving position, outputs its path after obtaining the global path of the current task, records the positions occupied by this path at each time interval in the form of keys in the global reservation table, and then deletes the current task from the task list.

[0033] Next, each robot moves along the global path while sensing the surrounding environment through on-board sensors to determine whether there are unexpected obstacles on the current task route. If there are unexpected obstacles, local path optimization is performed. Finally, when the robot moves to the task end point, it is regarded as a task completion flag, and the failure information output by the robot is regarded as a task abortion flag. Both completed and aborted tasks are removed from the execution list.

[0034] Avoiding obstacles: The neural network is used to perform data-level information fusion on the distance and azimuth information detected by multiple ultrasonic ranging sensors to eliminate the uncertainty of sensor data. Then, the fuzzy neural network is used to perform decision-level information fusion on the results obtained from the data-level information fusion and the path center deviation information, so that the robot can effectively avoid obstacles according to the obtained control signal and continue to move along the planned path.

[0035] Embodiment 2:

[0036] A multi-robot path planning method based on a fusion algorithm further includes: using sensors to obtain environmental data, analyzing and integrating the environmental data, and constructing an environmental model in combination with task characteristics; decomposing the overall task into subtasks, and dynamically allocating tasks to different robots according to geographical location and priority; testing in a simulation environment, simulating the behaviors of multiple robots in different scenarios, and collecting data for further analysis; analyzing the time consumption, path quality, and task completion rate of path planning, and optimizing the path planning and strategies according to the analysis results to improve the overall performance.

[0037] Building an environmental model: Perform data-level information fusion on the acquired original ultrasonic data to improve the accuracy of ultrasonic detection. Then, use the environmental information obtained from image processing and the ultrasonic fusion data for information fusion to verify each other to improve the accuracy and integrity of environmental data. Use the grid method to establish a multi-robot path planning map model suitable for the current environment, describe environmental elements in the map model, and stipulate the performance states and behavioral characteristics of environmental elements in multi-robot path planning.

[0038] Corresponding to the ultrasonic data group, there is a corresponding angular interval in the robot's visual image. According to the corresponding ranging information of the ultrasonic data group, determine which group of information in the corresponding angular interval is used as a basis to correct the distance between the obstacle and the robot. The specific determination principle is: If an obstacle is detected by only one ultrasonic sensor, this group is used as the key distance data.

[0039] If an obstacle straddles two adjacent ultrasonic sensor groups, select the sensor closer to the robot's due front as the key distance data.

[0040] If an obstacle spans three adjacent ultrasonic sensor groups, select the middle sensor as the key distance data.

[0041] If there are multiple obstacles and there is no obstacle directly in front, calculate whether the distance between the two obstacles allows the robot to pass through the distance image knowledge. If it allows, the robot continues to move forward. If not, mark the area directly in front as having an obstacle, and the distance is the average value of the left and right sensors.

[0042] Task dynamic allocation: Build an efficient communication network through the IoT platform and the MQTT protocol, combine parallel processing to achieve information sharing and collaboration between robots, monitor the status and collaboration progress of each robot, and adjust the strategy in a timely manner.

[0043] As can be seen from the above, the present invention uses a multi-robot path planning algorithm based on task priority to complete the global path planning of each robot, ensuring that tasks with higher priorities are planned first, which can effectively avoid collisions between robots, improve the safety of overall task completion, optimize the path through machine learning algorithms, enabling the robot to find shorter and more efficient paths, reducing energy consumption and time costs, using a hierarchical multi-sensor information fusion method to process real-time data, being able to accurately locate the position of obstacles, enabling the robot to quickly and effectively avoid obstacles, ensuring the continuity and stability of the path. The present invention not only improves the safety and efficiency of the robot moving in a complex environment, but also enhances its dynamic adaptability and collaboration ability, providing a solid foundation for the popularization of multi-robot systems in practical applications.

[0044] Embodiment 3:

[0045] The embodiment of the present invention further provides a terminal device, including a processor and a computer-readable storage medium. The processor is used to implement each instruction; the computer-readable storage medium is used to store multiple instructions, and the instructions are suitable for being loaded and executed by the processor to perform the multi-robot path planning method described in any one of the above.

[0046] The embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by the processor, it realizes each process of the above path planning method embodiment and can achieve the same technical effect. To avoid repetition, it will not be elaborated here. Among them, the computer-readable storage medium, such as a read-only memory (ROM for short), a random access memory (RAM for short), a magnetic disk or an optical disc, etc.

[0047] In the description of this specification, the description referring to terms such as "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic descriptions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples. In addition, without contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0048] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowchart and / or block diagram can be implemented by computer program instructions, and the combination of the flows and / or blocks in the flowchart and / or block diagram can also be implemented. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a program, so that the instructions executed by the processor of the computer or other programmable data processing devices generate a method for implementing the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0049] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including an instruction device, and the instruction device implements the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0050] These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0051] In the attached drawings of the disclosed embodiments of the present invention, only the structures related to the disclosed embodiments of the present invention are involved. For other structures, reference can be made to the general design. Without conflict, the same embodiment and different embodiments of the present invention can be combined with each other.

[0052] The flowcharts shown in the accompanying drawings are merely illustrative examples and do not necessarily include all contents and operations / steps, nor are they necessarily executed in the order described. For example, some operations / steps can be decomposed, combined or partially merged, so the actual execution order may change according to the actual situation.

[0053] Although the embodiments of the present invention have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A multi-robot path planning method based on a fusion algorithm, characterized in that, it includes: Step 1: Use sensors to obtain environmental data, analyze and integrate the environmental data, and construct an environmental model in combination with task characteristics; Step 2: Decompose the overall task into subtasks, and dynamically allocate tasks to different robots according to geographical location and priority; Step 3: Adopt a multi-robot path planning algorithm based on task priority for path planning, and use a machine learning algorithm to optimize the path; Step 4: Adopt an algorithm for hierarchical multi-sensor information fusion to enable the robot to track the path centerline and quickly and effectively avoid obstacles; Step 5: Conduct tests in a simulation environment, simulate the behavior of multi-robots in different scenarios, and collect data for further analysis; Step 6: Analyze the time consumption, path quality and task completion rate of path planning, and optimize the path planning and strategy according to the analysis results to improve the overall performance.

2. A multi-robot path planning method based on a fusion algorithm according to claim 1, characterized in that: For the analysis and integration of environmental data: perform data-level information fusion on the acquired original ultrasonic data to improve the accuracy of ultrasonic detection, and then perform information fusion on the environmental information obtained by image processing and the ultrasonic fusion data to verify each other to improve the accuracy and integrity of environmental data.

3. A multi-robot path planning method based on a fusion algorithm according to claim 2, characterized in that: For the construction of the environmental model: use the grid method to establish a multi-robot path planning map model suitable for the current environment, describe environmental elements in the map model, and stipulate the performance status and behavior characteristics of environmental elements in multi-robot path planning.

4. A multi-robot path planning method based on a fusion algorithm according to claim 3, characterized in that: For dynamically allocating tasks to different robots according to geographical location and priority: build an efficient communication network through the IoT platform and the MQTT protocol, realize information sharing and cooperation between robots through parallel processing, monitor the status and cooperation progress of each robot, and adjust the strategy in a timely manner.

5. A multi-robot path planning method based on a fusion algorithm according to claim 4, characterized in that: For adopting a multi-robot path planning algorithm based on task priority for path planning: First, determine the factors affecting the path planning priority in the multi-robot system according to the task requirements of the current scenario; then, perform global path planning for each robot executing the task through the IoT platform; next, each robot moves along the global path, and at the same time senses the surrounding environment through the accompanying sensors to judge whether there are unexpected obstacles on the current task route. If there are unexpected obstacles, perform local path optimization; finally, use the robot's movement to the task end point as the task completion flag, and use the failure information output by the robot as the task abort flag. The completed and aborted tasks are both removed from the execution list.

6. A multi-robot path planning method based on a fusion algorithm according to claim 5, characterized in that: The application The algorithm completes the multi-robot path planning task, and the algorithm always accesses the top task of the task list to ensure that tasks with higher priorities are planned first and occupy the global reservation table. Information is saved and shared among multiple robots through the global reservation table. A machine learning algorithm is introduced to train a model using historical data to predict collision risks and optimize paths.

7. A multi-robot path planning method based on a fusion algorithm according to claim 6, wherein: The single-robot global path planning: The robot starts from the task starting point, continuously traverses the adjacent reachable node set of the current node, judges whether the robot has waiting behavior, turning behavior, and diagonal straight-line movement at the adjacent node, selects the one with the smallest cost as the next moving position, outputs the path after obtaining the global path of the current task, records the positions occupied within each time interval in the form of keys in the global reservation table, and then deletes the current task from the task list.

8. A multi-robot path planning method based on a fusion algorithm according to claim 7, wherein: The algorithm for avoiding obstacles by using a hierarchical multi-sensor information fusion: A neural network is used to perform data-level information fusion on the distance and azimuth information detected by multiple ultrasonic ranging sensors to eliminate the uncertainty of sensor data; then a fuzzy neural network is used to perform decision-level information fusion on the result obtained from the data-level information fusion and the path center deviation information, so that the robot can effectively avoid obstacles according to the obtained control signal and continue to move along the planned path.

9. A terminal device, characterized in that: It includes a processor and a computer-readable storage medium. The processor is used to implement each instruction; the computer-readable storage medium is used to store multiple instructions, and the instructions are suitable for being loaded and executed by the processor to perform a multi-robot path planning method based on a fusion algorithm according to any one of claims 1-8.

10. A computer-readable storage medium, characterized in that, A computer program is stored on the computer-readable storage medium, and when the computer program is executed by the processor, the processor is caused to implement a multi-robot path planning method based on a fusion algorithm as described in claims 1-8.

Images