Multi-robot cooperation method and system for intelligent transportation

Abstract

The invention provides a multi-robot cooperation method and system for intelligent traffic, and relates to the technical field of intelligent traffic, and the method comprises the steps: generating an initial task distribution scheme based on traffic demands and real-time traffic data; obtaining standard sensing data through multi-source sensors carried by a plurality of robots; acquiring a plurality of real-time states of the plurality of robots, and adjusting the initial task allocation scheme in combination with the standard sensing data to obtain a real-time task allocation scheme; and determining a plurality of task paths of the plurality of robots based on the real-time task allocation scheme, and performing collaborative management on the plurality of robots. According to the method and the device, the technical problem that the traffic management efficiency is affected due to inflexible distribution of multiple robots caused by the fact that traffic management cannot flexibly cope with dynamically changing traffic demands in a complex environment is solved, and the traffic management efficiency is improved by acquiring and analyzing the multi-source traffic data in real time and dynamically adjusting the task distribution of the robots. The execution efficiency of the whole task is ensured, and the traffic management efficiency is improved.

Description

Technical Field

[0001] The present application relates to the field of intelligent transportation technology, and in particular to a multi-robot collaboration method and system for intelligent transportation. Background Art

[0002] With the acceleration of urbanization, traffic management faces increasingly complex challenges. Traditional traffic management systems typically rely on technologies such as manual monitoring, fixed traffic light control, and camera monitoring. However, due to factors such as instantaneous changes in traffic flow, sudden traffic accidents, and weather changes, they often fail to make timely and effective adjustments, leading to traffic congestion, increased accident rates, and even waste of traffic resources and inefficiency. With the development of autonomous driving and robotics technologies, the concept of collaborative traffic management using multiple robots is gaining application and exploration. Multi-robot systems, equipped with sensors, computing platforms, and communication modules, can implement real-time monitoring, data collection and processing, traffic scheduling, and management. However, even with robot-based traffic management, task allocation may still rely on pre-set static schemes, making them difficult to adapt to rapidly changing traffic conditions. When traffic conditions change or a robot's status becomes abnormal, task allocation may not be adjusted in a timely manner, resulting in inefficient robot collaboration in complex environments and even waste of resources or mission failure.

[0003] In summary, the existing technology has a technical problem that traffic management cannot flexibly respond to dynamically changing traffic demands in complex environments, resulting in inflexible allocation of multiple robots, which in turn affects the efficiency of traffic management. Summary of the Invention

[0004] The purpose of this application is to provide a multi-robot collaboration method and system for intelligent transportation, so as to solve the technical problem in the existing technology that traffic management cannot flexibly respond to dynamically changing traffic demands in complex environments, resulting in inflexible allocation of multiple robots, which in turn affects the efficiency of traffic management.

[0005] In view of the above problems, the present application provides a multi-robot collaboration method and system for intelligent transportation.

[0006] In the first aspect, the present application provides a multi-robot collaboration method for intelligent transportation, which is implemented by a multi-robot collaboration system for intelligent transportation, wherein the multi-robot collaboration method for intelligent transportation includes: obtaining traffic demand in a target area, extracting real-time traffic data from a sensor monitoring module, and generating an initial task allocation plan based on the traffic demand and the real-time traffic data; obtaining traffic perception data through multi-source sensors carried by multiple robots, uploading the traffic perception data to an edge computing node for localized processing, and obtaining standard perception data; obtaining multiple real-time states of multiple robots, and adjusting the initial task allocation plan in combination with the standard perception data to obtain a real-time task allocation plan; determining multiple task paths of the multiple robots based on the real-time task allocation plan, and collaboratively managing the multiple robots.

[0007] Optionally, based on the historical traffic data set of the target area, multiple traffic flows of multiple sub-areas are obtained; the multiple traffic flows are sorted in descending order to determine the key traffic area; the historical traffic data set is traversed to obtain historical traffic event records of the key traffic area, and multiple sensing points are identified; multiple traffic sensing sensors are deployed in the key traffic area according to the multiple sensing points to obtain the sensor monitoring module.

[0008] Optionally, the edge computing node performs spatiotemporal alignment processing on the received traffic perception data to obtain first perception data; performs anomaly detection on the first perception data, performs data compensation based on the anomaly detection result to obtain second perception data; and performs standardization processing on the second perception data to obtain the standard perception data.

[0009] Optionally, the traffic perception data is time-aligned according to the acquisition timestamp and sampling frequency to obtain traffic time series data; the multi-source sensor is calibrated according to the edge computing node to obtain the spatial position relationship of the multi-source sensor; based on the spatial position relationship, combined with the acquisition data characteristics of the multi-source sensor, a spatial coordinate system is constructed; the traffic time series data is projected onto the spatial coordinate system to obtain first perception data.

[0010] Optionally, based on the traffic demand and the historical traffic data set, multiple traffic task types are determined; based on the importance and timeliness of the multiple traffic task types, task priorities are determined; the target area is divided according to the task priority to determine multiple sub-areas; based on task load balancing and multiple real-time status information of the multiple robots, multiple task allocation schemes of the multiple sub-areas are matched in the initial task allocation scheme; and the multiple task allocation schemes are integrated to obtain the real-time task allocation scheme.

[0011] Optionally, step a: randomly select a first sub-area from the multiple sub-areas; step b: extract multiple robot positions from the multiple real-time status information, determine the number and type of robots in the first sub-area in combination with the task priority, and obtain multiple zone one robots; step c: obtain the self-state data of the first robot among the multiple zone one robots, send it to at least one collaborative robot in the same area, and receive collaborative state data of at least one collaborative robot; step d: combine the self-state data and the collaborative state data, and under the first zone constraint conditions, perform path planning with the goal of optimal path and optimal task execution, and generate a first task allocation plan, wherein the first zone constraint conditions include robot load balancing and collaborative obstacle avoidance; traverse the multiple sub-areas, repeat steps a to d, and generate the multiple task allocation plans.

[0012] Optionally, the task execution status of the multiple robots is monitored in real time, and the task execution status is uploaded to the edge computing node for task evaluation to obtain task feedback information; and the multiple task paths are updated according to the task feedback information.

[0013] In the second aspect, the present application also provides a multi-robot collaborative system for intelligent transportation, which is used to execute a multi-robot collaborative method for intelligent transportation as described in the first aspect, wherein the multi-robot collaborative system for intelligent transportation includes: an initial plan determination module, which is used to obtain the traffic demand of the target area and extract real-time traffic data from the sensor monitoring module, and generate an initial task allocation plan based on the traffic demand and the real-time traffic data; a localization processing module, which is used to obtain traffic perception data through multi-source sensors carried by multiple robots, and upload the traffic perception data to an edge computing node for localization processing to obtain standard perception data; an allocation plan determination module, which is used to obtain multiple real-time states of multiple robots, and adjust the initial task allocation plan in combination with the standard perception data to obtain a real-time task allocation plan; a collaborative management module, which is used to determine multiple task paths of the multiple robots based on the real-time task allocation plan, and collaboratively manage the multiple robots.

[0014] One or more technical solutions provided in this application have at least the following beneficial effects:

[0015] By acquiring the traffic demand in the target area and extracting real-time traffic data from the sensor monitoring module, an initial task allocation plan is generated based on the traffic demand and the real-time traffic data. Traffic perception data is acquired through multi-source sensors carried by multiple robots, and the traffic perception data is uploaded to the edge computing node for local processing to obtain standard perception data. Multiple real-time states of the multiple robots are acquired and, combined with the standard perception data, the initial task allocation plan is adjusted to obtain a real-time task allocation plan. Multiple task paths for the multiple robots are determined based on the real-time task allocation plan, and the multiple robots are collaboratively managed. In other words, by acquiring traffic demand and real-time traffic data, an initial task allocation plan is generated accordingly, and the status information of the multiple robots and changes in the traffic environment are continuously monitored, and the task allocation plan is adjusted in real time, so that the robots can flexibly respond to changing traffic demand according to actual conditions, avoiding the problem of inflexible task allocation and improving overall traffic management efficiency.

[0016] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, which can be implemented in accordance with the contents of the description, and to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are specifically listed below. It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easy to understand through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in this application or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are merely exemplary, and a person of ordinary skill in the art can obtain other drawings based on the provided drawings without creative work.

[0018] Figure 1 A flowchart of a multi-robot collaborative method for intelligent transportation is provided for this application;

[0019] Figure 2 This is a structural diagram of a multi-robot collaborative system for intelligent transportation in this application.

[0020] Explanation of the reference numerals: initial plan determination module 11 , localization processing module 12 , allocation plan determination module 13 , collaborative management module 14 . DETAILED DESCRIPTION

[0021] This application provides a multi-robot collaborative method and system for intelligent transportation, addressing the existing technical issues of inflexible allocation of multiple robots, which in turn affects traffic management efficiency, due to the inability of traffic management to flexibly respond to dynamically changing traffic demands in complex environments. By acquiring traffic demand and real-time traffic data, generating an initial task allocation plan based on this, continuously monitoring the status information of multiple robots and changes in the traffic environment, and adjusting the task allocation plan in real time, the robots can flexibly respond to changing traffic demands based on actual conditions, avoiding the problem of inflexible task allocation and improving overall traffic management efficiency.

[0022] Below, the technical solutions in this application will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of this application, rather than all the embodiments of this application. It should be understood that this application is not limited to the example embodiments described herein. Based on the embodiments of this application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application. It should also be noted that, for the convenience of description, only the parts related to this application, rather than all of them, are shown in the accompanying drawings.

[0023] For example, see the attached Figure 1 The present application provides a multi-robot collaboration method for intelligent transportation, wherein the multi-robot collaboration method for intelligent transportation is performed by a multi-robot collaboration system for intelligent transportation, and the multi-robot collaboration method for intelligent transportation specifically includes the following steps:

[0024] S100: Obtaining traffic demand in a target area, extracting real-time traffic data from a sensor monitoring module, and generating an initial task allocation plan based on the traffic demand and the real-time traffic data.

[0025] Specifically, traffic demand in the target area can be predicted and inferred through long-term traffic monitoring data analysis, or dynamically acquired through real-time monitoring data. Traffic demand data includes traffic flow demand, traffic safety demand, special event demand, road capacity, and speed limits. Traffic flow demand is the intensity of traffic demand on each road section, determined based on the volume of traffic at different time periods and on different road sections. Traffic safety demand is a safety management requirement that requires priority attention, including high-risk intersections, accident-prone areas, and school areas. Special event demand includes traffic closures or congestion on certain road sections due to construction, accidents, or weather issues, requiring real-time adjustment of task allocation. Road capacity and speed limits need to take into account factors such as the road type, speed limit, and number of vehicles that can be accommodated on each road section.

[0026] For example, if the target area is a busy intersection in a city, traffic demand can be estimated by analyzing traffic volume, road load, traffic signals, and other information over the past few days or months. For example, if a city intersection experiences traffic of 2,200 vehicles per hour during the morning rush hour, the traffic demand is reflected in the high traffic load during that period.

[0027] Real-time traffic data is collected by sensing modules deployed in key traffic areas, such as cameras, radar sensors, infrared sensors, meteorological sensors, GPS, and vehicle detectors. This data includes traffic flow, congestion, weather conditions, and emergencies. Based on traffic demand and real-time traffic data, traffic pressure in each area at different time periods is analyzed. A congestion index is calculated based on real-time traffic data and road capacity, identifying areas with excessive traffic or bottlenecks. For example, traffic flow changes during peak hours, holidays, and periods of inclement weather can be analyzed to infer demand in a target area.

[0028] Real-time traffic data is obtained through sensor monitoring modules installed in traffic areas, including video surveillance cameras (which use image recognition to count vehicle flow), LiDAR (which monitors vehicle movement by measuring distance and speed), geomagnetic sensors (for detecting the presence of vehicles in lanes), and traffic signal control systems (for adjusting traffic light cycles). For example, video surveillance cameras can capture 100 frames of image per second and use image recognition technology to analyze the number and speed of vehicles in each frame, thereby obtaining real-time traffic flow information.

[0029] After acquiring traffic demand data and real-time traffic data, a preliminary task allocation plan is generated based on this information, including the number of robots required in each area, the tasks each robot will perform, how to assign robots to different lanes, and how to avoid path conflicts. For example, based on real-time traffic flow data, two robots can monitor the road in two directions with heavy traffic, while an unmanned vehicle can be dispatched to direct traffic. By analyzing real-time data and traffic demand, robots can be flexibly dispatched for traffic management. The task allocation plan can more accurately reflect the actual situation, improving the responsiveness and effectiveness of overall traffic management.

[0030] Furthermore, the present application S100 includes:

[0031] Based on the historical traffic data set of the target area, multiple traffic flows of multiple sub-areas are obtained; the multiple traffic flows are sorted in descending order to determine the key traffic area; the historical traffic data set is traversed to obtain historical traffic event records of the key traffic area, and multiple sensing points are identified; and multiple traffic sensing sensors are deployed in the key traffic area according to the multiple sensing points to obtain the sensing monitoring module.

[0032] Specifically, the system obtains the historical traffic data set of the target area, that is, the traffic flow, vehicle speed, traffic incidents and other data collected in the target area (such as a city, block or road section) over a period of time. The target area is divided into multiple sub-areas, and the historical traffic flow data in each sub-area is analyzed. For example, assuming that the target area of ​​a city is divided into 10 sub-areas (such as different blocks or road sections), by analyzing the historical traffic data set, the system can derive the traffic flow of each sub-area in the past week. The traffic flow of sub-area A is 5,000 vehicles per day, sub-area B is 7,000 vehicles per day, sub-area C is 9,000 vehicles per day, sub-area D is 4,000 vehicles per day, and so on.

[0033] Use a sorting algorithm to sort the traffic flow in each sub-area in descending order. After sorting, sub-areas with higher traffic flow are identified as critical traffic areas, which are typically prioritized for traffic management and monitoring. For example, if sub-areas A, B, C, and D are sorted in descending order, sub-areas B and C can be considered critical traffic areas.

[0034] The traffic event records in the historical traffic dataset are traversed to analyze traffic accidents, congestion, road closures, and other events in key traffic areas (such as sub-areas B and C). The locations of sensing points are determined based on the location and frequency of these events. For example, through the analysis of historical traffic event records, it is found that traffic accidents or congestion often occur at a certain intersection or road section. Therefore, a large number of traffic sensing devices need to be deployed in these areas. Multiple sensing points are determined, that is, the points where traffic sensing devices need to be deployed. Sensing points are usually places with high traffic volume or where traffic events often occur, such as intersections, exits, traffic bottlenecks, and traffic rush hours.

[0035] Based on the identified sensing points, multiple traffic sensing sensors, such as video surveillance cameras, radar, geomagnetic sensors, and lidar, are deployed in key traffic areas to collect traffic data such as vehicle volume, speed, road conditions, and traffic signals. By deploying multiple sensing devices, a complete sensing and monitoring module is formed, capable of collecting and transmitting traffic data in real time. A sensing and monitoring module typically includes multiple different types of sensing devices to enhance comprehensive traffic monitoring capabilities.

[0036] By analyzing historical data and sorting traffic, we can identify key traffic areas and rationally deploy sensing equipment to comprehensively monitor traffic flow, vehicle speed, and traffic events, thus avoiding unnecessary waste of resources, concentrating monitoring and traffic management resources on the most critical areas, and improving the efficiency of overall traffic management.

[0037] S200: Acquire traffic perception data through multi-source sensors carried by multiple robots, upload the traffic perception data to an edge computing node for local processing, and obtain standard perception data.

[0038] Furthermore, the present application S200 includes:

[0039] The edge computing node performs spatiotemporal alignment processing on the received traffic perception data to obtain first perception data; performs anomaly detection on the first perception data, performs data compensation based on the anomaly detection result to obtain second perception data; and performs standardization processing on the second perception data to obtain the standard perception data.

[0040] Specifically, each robot is equipped with different types of sensors, such as cameras (to capture images or video streams of the robot's surroundings and identify road obstacles and traffic signs), GPS (to obtain the robot's location coordinates), IMU (to obtain the robot's posture information), and ultrasonic sensors (to detect obstacles at close range). These sensors collect traffic perception data in real time around the robot. Through the multi-source sensors onboard multiple robots, traffic perception data is collected, including the robot's position, speed, posture, and status (such as battery level and operating status), as well as traffic information on the surrounding road, such as obstacles, other vehicles, or pedestrians.

[0041] Traffic sensing data is transmitted to edge computing nodes via wireless networks for processing, ensuring that all data is synchronized under the same timestamp and spatial framework. Edge computing nodes are computing devices deployed close to the data source (such as near the robot's sensors) and are used to process, store, and analyze data. Data collected by different sensors is uniformly aligned according to time and space to ensure consistency between the data. Different sensors may have different collection times and locations. Through spatiotemporal alignment, it can be ensured that the data can be analyzed and processed under the same timestamp and spatial coordinates.

[0042] After spatiotemporal alignment, the first perception data is obtained. This data has been matched according to a unified time and space, ensuring that all data has consistent timestamps and spatial locations, providing a reliable foundation for subsequent anomaly detection and data compensation. The first perception data is analyzed using an anomaly detection algorithm to identify abnormal data or errors. The mean and standard deviation of the first perception data are calculated using the Z-score method to identify outliers that are outside the normal range. For example, a vehicle speed exceeding a certain threshold (such as 200 km / h) may be considered an outlier. Suppose the sensor records a vehicle speed of 1000 km / h at a certain moment, which is an obviously unreasonable value.

[0043] For identified abnormal data, i.e., anomaly detection results, data compensation is typically performed through interpolation or smoothing methods to correct outliers or missing data identified during the anomaly detection phase. Missing or abnormal data points are filled in using the values ​​of nearby data points. For example, if vehicle speed data is lost at a certain moment due to a sensor failure, the edge computing node can linearly interpolate the speed values ​​of the preceding and following data points to fill in the missing data. Assuming the vehicle speeds were 30 km / h and 32 km / h two seconds before and after, respectively, the edge computing node can interpolate the speed at the moment of loss to 31 km / h. After compensation, the missing or abnormal data is resolved, forming secondary perception data that is more reliable and complete. Through compensation, the data no longer contains errors or missing data, improving its quality.

[0044] Standardize secondary sensor data, uniformly formatting and normalizing the data to a consistent scale and units. This includes unit unification and normalization (scaling the data to a consistent range). For example, some sensors may measure data in meters per second, but converting this to kilometers is not necessary. Unifying values ​​collected by different sensors to the range [0, 1] eliminates the impact of dimensionality on analysis. Through spatiotemporal alignment, anomaly detection, and data compensation, high data quality is ensured, the impact of errors and outliers is reduced, and data quality and accuracy are significantly improved.

[0045] Furthermore, the present application further comprises the following steps:

[0046] The traffic perception data is time-aligned according to the collection timestamp and sampling frequency to obtain traffic time series data; the multi-source sensor is calibrated according to the edge computing node to obtain the spatial position relationship of the multi-source sensors; based on the spatial position relationship and in combination with the collection data characteristics of the multi-source sensors, a spatial coordinate system is constructed; the traffic time series data is projected onto the spatial coordinate system to obtain first perception data.

[0047] Specifically, each sensor collects data at a preset sampling frequency, and the collected data has a corresponding timestamp. Because different sensors have different update frequencies, resulting in different sampling frequencies and sampling timestamps, the data needs to be aligned to the same time scale so that the data can be compared and fused at the same moment. Interpolation algorithms (such as linear interpolation and spline interpolation) are used to align data with different timestamps to the same point in time. For all sensors, their clocks are synchronized, which can be achieved through methods such as GPS time synchronization and network clock synchronization. For example, suppose that the lidar updates data once per second, while the camera updates data every 30 milliseconds. The edge computing node uses interpolation technology to align all data to the same point in time, such as updating every millisecond. After time alignment, the data from all sensors has a consistent timestamp, generating traffic time series data that represents the changes in the traffic environment at different time points.

[0048] The goal of sensor calibration is to accurately measure the relative position and attitude relationships between different sensors. Using known reference points, the spatial relationship between multiple sensors—that is, the spatial configuration and relative position of multiple sensors, including information such as the distance and angle between each sensor—is determined. Based on the spatial relationship between the multiple sensors, the relative distance and orientation of each sensor in a global coordinate system are determined. Each sensor has unique data acquisition characteristics, such as sampling frequency, measurement range, and accuracy. A sensor is selected as the reference coordinate system. Based on the spatial relationship between the other sensors and the reference coordinate system, the data from the other sensors is transformed into the reference coordinate system using translation and rotation matrices. Each sensor has its own data acquisition characteristics, which are integrated to form a spatial coordinate system that represents the position and orientation of the sensor or object in space.

[0049] Once all sensor data has been converted to a unified coordinate system, traffic time series data can be projected into this coordinate system. After projecting all sensor data into the spatial coordinate system, the resulting primary sensor data has been accurately aligned and mapped in time and space, and the data has a unified spatiotemporal reference. By constructing a unified spatial coordinate system, data from different sensors can be accurately fused together, and the advantages of different sensors can be complemented. LiDAR environmental perception, IMU motion estimation, and GPS positioning information can be integrated in the same coordinate system, enabling the traffic management system to have a more comprehensive understanding of the traffic environment. Through the calibration process and the construction of the spatial coordinate system, the measurement data of each sensor is ensured to be mapped in a precise spatial coordinate system, thereby improving the spatial accuracy of the data.

[0050] S300: Acquire multiple real-time status information of multiple robots, combine the standard perception data, adjust the initial task allocation plan, and obtain a real-time task allocation plan.

[0051] Furthermore, the present application S300 includes:

[0052] Based on the traffic demand and the historical traffic data set, multiple traffic task types are determined; based on the importance and timeliness of the multiple traffic task types, task priorities are determined; the target area is divided according to the task priorities to determine multiple sub-areas; based on task load balancing and multiple real-time status information of the multiple robots, multiple task allocation schemes for the multiple sub-areas are matched in the initial task allocation scheme; and the multiple task allocation schemes are integrated to obtain the real-time task allocation scheme.

[0053] Specifically, based on the target area's traffic demand and combined with historical traffic datasets, the type of traffic task—the traffic task to be assigned to the robot—is determined, such as traffic flow monitoring, traffic accident handling, road condition analysis, and traffic light optimization. Traffic demand refers to the demand for traffic management and resources within the target area, including traffic flow, road conditions, and incident frequency. Historical traffic datasets refer to data accumulated over a period of time, including traffic flow, incident records, traffic accident information, and road conditions. Assume that by analyzing traffic data from the past three months and combining it with traffic demand, the following traffic task types are identified: flow monitoring, accident handling, and traffic light optimization. The flow monitoring task primarily monitors traffic flow trends, especially during peak hours in the morning and evening. The accident handling task is responsible for identifying high-risk areas based on historical accident data and rapidly dispatching robots to handle incidents. The traffic light optimization task is responsible for optimizing traffic light control strategies during periods of traffic congestion.

[0054] Perform importance and timeliness analysis on multiple traffic task types, sort tasks based on factors such as urgency, importance, and timeliness, and ensure that the most urgent and important tasks are given priority. Tasks with higher priority need to be executed as soon as possible, while tasks with lower priority can be delayed or executed later. The importance of a task refers to the degree of impact of the task on the overall traffic management goals. A highly important task means that its completion or failure will directly affect traffic flow, traffic safety, or system stability. The timeliness of a task refers to the time required for task execution and the time window for task completion. Tasks with strong timeliness require rapid completion, such as handling emergencies; while tasks with lower timeliness can be completed within a certain time frame, such as routine traffic monitoring.

[0055] Each person is assigned a different weight based on their overall impact and urgency. A weighted scoring method is used to evaluate the importance and timeliness of each person's role, resulting in a comprehensive priority for each person, which together constitutes the task priority. The target area is divided into multiple sub-areas based on task priority. For example, areas with heavy traffic and frequent accidents are classified as high-priority sub-areas. Based on the distribution of different task types, areas with a high concentration of tasks are classified as high-priority areas. The target area is divided into multiple sub-areas based on task priority, and each sub-area will perform tasks of different priorities. The division of sub-areas helps to concentrate resources and improve task processing efficiency.

[0056] Task load balancing ensures that the workload is properly distributed among multiple robots when they participate in a task, avoiding uneven task execution, where some robots are overloaded while others are underloaded. Task load balancing is performed based on the robots' real-time status information (such as battery level, load capacity, and current location). In other words, each sub-area is assigned a different number and type of robots based on task priority and difficulty. Each robot's execution path is determined based on its real-time status information, forming a task allocation plan for each sub-area. The task allocation plan includes the number of robots required for the area, the type of robots, and the execution path for each robot.

[0057] Multiple task allocation plans corresponding to all sub-areas are integrated to generate a real-time task allocation plan, including the tasks to be performed by each area and each robot, their priority, and the execution order. This plan also determines the tasks to be performed by each robot, the order in which they are executed, and the robot's path of action. This real-time task allocation plan intelligently allocates tasks based on real-time traffic demand and task priority, ensuring efficient and orderly traffic management. Task scheduling is performed based on each robot's real-time status information, avoiding uneven task burdens on robots and improving overall operational efficiency. A clear task priority allocation mechanism and robot load balancing ensure that high-priority tasks are completed in the shortest possible time, improving the responsiveness of traffic management.

[0058] Furthermore, the present application further comprises the following steps:

[0059] Step a: randomly select a first sub-area from the multiple sub-areas; step b: extract multiple robot positions from the multiple real-time status information, determine the number and type of robots in the first sub-area in combination with the task priority, and obtain multiple first-area robots; step c: obtain the self-state data of the first robot among the multiple first-area robots, send it to at least one collaborative robot in the same area, and receive the collaborative state data of at least one of the collaborative robots; step d: combine the self-state data and the collaborative state data, and under the first area constraint condition, perform path planning with the goal of optimal path and optimal task execution, and generate a first task allocation plan, wherein the first area constraint condition includes robot load balancing and collaborative obstacle avoidance; traverse the multiple sub-areas, repeat steps a to d, and generate the multiple task allocation plans.

[0060] Specifically, the following steps are performed for each of the multiple sub-areas, resulting in multiple task allocation schemes corresponding to the sub-areas. For example, the first sub-area is randomly selected from the multiple sub-areas. Based on the real-time status information of multiple robots, the current position, speed, load, remaining battery life, and other data of all robots are obtained. The current position of all robots is extracted from this data. Combined with the previously determined task priorities, the number and type of robots required for the first sub-area are determined, thereby determining the multiple first-area robots, i.e., all the robots required for the first area.

[0061] A robot is randomly selected from multiple robots in a zone. The robot's status data, including its current position, speed, and task status, is then transmitted to other collaborative robots in the same zone. These collaborative robots then feed back their own status data (such as position, status, and speed) to the first robot, ensuring collaborative task completion and coordination, avoiding conflicts and improving efficiency. By sharing status data, robots can better coordinate task execution, improving the efficiency and safety of collaborative work.

[0062] For example, the first robot is moving from point A to point B and has completed part of its mission. It then sends information such as its position and speed to a collaborative robot (such as the second robot) via wireless communication. After receiving this information, the second robot also feeds back its status data (such as its position and status) to the first robot.

[0063] Load balancing and collaborative obstacle avoidance are used as constraints. Load balancing ensures that each robot performs appropriate tasks based on its capabilities and status. If some robots have low battery or are overloaded with tasks, task allocation is automatically adjusted to prevent these robots from being overloaded. Collaborative obstacle avoidance refers to avoiding collisions or interference between multiple robots when they work together, as well as avoiding obstacles on the road. Optimal path and optimal task execution are used as objective functions, and path planning is performed using the Rapid Randomized Tree Algorithm (RRT). RRT is an algorithm for path planning that is particularly suitable for path search in high-dimensional spaces and dynamic environments. By randomly selecting sample points and expanding the tree structure, feasible paths are gradually explored, and an approximate optimal path from the starting point to the end point can be quickly found. It has fast search capabilities and is particularly suitable for dynamic path planning in complex environments.

[0064] Randomly select a point within the target area and explore the path by expanding the tree structure. Expand a new node from the end of the current tree until the node is close to the target position. If the point is feasible (that is, there is no obstacle blocking it), the tree structure will continue to expand. The algorithm generates a path from the starting point to the end point by continuously expanding the tree structure. When expanding each node, the RRT algorithm will detect whether the current path collides with an obstacle. If a collision occurs, the algorithm will abandon the path and continue to try to expand in other directions. Ensure that the extended path does not collide with known obstacles (such as buildings, roadblocks, etc.). In a multi-robot environment, the RRT algorithm also needs to consider the presence of dynamic obstacles (such as other robots). By updating the position of each robot in real time, the path is prevented from colliding with other robots.

[0065] At the same time, the robot's assigned tasks are dynamically adjusted based on each robot's current position, power level, load capacity, and other information. As the robots' work progresses and their status changes, task allocation is continuously monitored and adjusted to ensure a consistently balanced load. In addition to avoiding collisions with static and dynamic obstacles (such as other robots), collaboration and obstacle avoidance between robots are also essential. By applying virtual forces between robots, the robots are guided toward the target location while avoiding other robots, ensuring that they avoid collisions as they move.

[0066] Based on the robot's acquired map of the target area, it identifies obstacles and passes this information to the RRT algorithm. The RRT algorithm accounts for static obstacles when generating a path. As road conditions change (such as the appearance of new obstacles), the RRT algorithm updates the path in real time, ensuring the robot consistently avoids these obstacles.

[0067] Through the above steps, each robot obtains multiple feasible paths. These paths are evaluated using the objective functions of path optimization and task execution optimization to obtain a score for each path. The path with the highest score is selected as the path for each robot to execute. The first task allocation plan is formed based on the number of robots in a zone, the paths executed by multiple robots in a zone, and the types of robots in a zone.

[0068] The above steps are repeated for other sub-areas, resulting in multiple task allocation plans corresponding to each sub-area. Each sub-area's task allocation plan is optimized based on factors such as the task requirements, robot status, and path planning for that area. By dynamically adjusting task allocation for each sub-area based on task priorities and real-time status information, the robots are able to quickly respond and efficiently handle different types of tasks. By exchanging status data, robots can effectively coordinate their respective tasks and paths, avoiding collisions and interference with each other and achieving efficient collaborative work. In complex and dynamic traffic environments, task allocation plans are adjusted in real time based on changes in task requirements and robot status to ensure efficient task completion.

[0069] S400: Matching multiple task paths of the multiple robots in the real-time task allocation solution, and collaboratively managing the multiple robots.

[0070] Specifically, the real-time task allocation plan includes not only the specific tasks to be performed by the robots, but also their priorities, the order in which they are to be executed, and the paths each robot must follow. The real-time task allocation plan includes the number of robots required for each area, their types, and the paths they must follow. These are then allocated to multiple robots, ensuring that each robot can reach its corresponding sub-area and perform its tasks. By matching each robot's task path with the task requirements within the area, multiple task paths are generated for multiple robots, enabling collaborative management of multiple robots.

[0071] The task path of each matched robot is sent to the corresponding robot through the control center, and the robot executes the task according to the task path. While performing a task, the robot sends its status information (such as current location, battery level, task progress, etc.) to the control center in real time, and the task is adjusted based on this real-time information. If a robot malfunctions, is overloaded, or has insufficient battery, the task is reassigned to another robot to ensure that the task is not affected. By matching the task paths of multiple robots in a real-time task allocation plan and through collaborative management, multiple robots can collaborate to complete the task, avoiding conflicts and collisions, and ensuring smooth task execution.

[0072] Furthermore, the present application further comprises the following steps:

[0073] The task execution status of the multiple robots is monitored in real time, and the task execution status is uploaded to the edge computing node for task evaluation to obtain task feedback information; and the multiple task paths are updated according to the task feedback information.

[0074] Specifically, the multi-source sensors onboard multiple robots monitor their task execution in real time, including information such as robot position, speed, path deviation, and battery level. This information is then uploaded to the edge computing node via wireless communication. As the robots perform their tasks, they continuously collect task-related data, tracking their progress in real time and verifying that each robot's task is proceeding according to plan. This collected real-time data is then uploaded to the edge computing node via wireless communication.

[0075] After receiving this data, the edge computing node evaluates the execution status of each robot task, including a comprehensive analysis of task progress, path feasibility, obstacle avoidance, and robot status, generating task feedback. This evaluation considers not only task progress but also multiple dimensions, including path feasibility, obstacle avoidance, and robot status, to comprehensively assess task performance and any issues.

[0076] After the edge computing node completes the task evaluation, it generates task feedback information, including issues encountered during task execution, changes in the robot's execution path, and task completion progress. The system dynamically adjusts the task path based on this feedback. If the robot encounters new obstacles, deviates from the path, or experiences inefficiencies during task execution, the task path is adjusted based on this feedback. The robot adjusts to the new task path and provides real-time feedback on the execution status. The system then re-evaluates and adjusts the path. This way, task execution is continuously optimized until the task is completed.

[0077] Through real-time monitoring and task evaluation, we can accurately understand the task execution status of each robot and quickly optimize the task path based on feedback information to improve task execution efficiency. We can also make adaptive adjustments based on the ever-changing environment and task requirements to cope with complex and dynamic traffic management tasks, and enhance the flexibility and adaptability of intelligent traffic management.

[0078] In summary, the multi-robot collaboration method for intelligent transportation provided by this application has the following beneficial effects:

[0079] By acquiring the traffic demand in the target area and extracting real-time traffic data from the sensor monitoring module, an initial task allocation plan is generated based on the traffic demand and the real-time traffic data. Traffic perception data is acquired through multi-source sensors carried by multiple robots, and the traffic perception data is uploaded to the edge computing node for local processing to obtain standard perception data. Multiple real-time states of the multiple robots are acquired and, combined with the standard perception data, the initial task allocation plan is adjusted to obtain a real-time task allocation plan. Multiple task paths for the multiple robots are determined based on the real-time task allocation plan, and the multiple robots are collaboratively managed. In other words, by acquiring traffic demand and real-time traffic data, an initial task allocation plan is generated accordingly, and the status information of the multiple robots and changes in the traffic environment are continuously monitored, and the task allocation plan is adjusted in real time, so that the robots can flexibly respond to changing traffic demand according to actual conditions, avoiding the problem of inflexible task allocation and improving overall traffic management efficiency.

[0080] Example 2: Based on the same inventive concept as the multi-robot collaborative method for intelligent transportation in the aforementioned Example 1, this application also provides a multi-robot collaborative system for intelligent transportation, see the attached Figure 2 , the multi-robot collaborative system for intelligent transportation includes:

[0081] The initial plan determination module 11 is used to obtain the traffic demand of the target area and extract real-time traffic data from the sensor monitoring module, and generate an initial task allocation plan based on the traffic demand and the real-time traffic data; the localization processing module 12 is used to obtain traffic perception data through multi-source sensors carried by multiple robots, upload the traffic perception data to the edge computing node for localization processing, and obtain standard perception data; the allocation plan determination module 13 is used to obtain multiple real-time states of multiple robots, and adjust the initial task allocation plan in combination with the standard perception data to obtain a real-time task allocation plan; the collaborative management module 14 is used to determine multiple task paths of the multiple robots based on the real-time task allocation plan, and collaboratively manage the multiple robots.

[0082] Furthermore, the initial solution determination module 11 in the multi-robot collaborative system for intelligent transportation is further configured to:

[0083] Based on the historical traffic data set of the target area, multiple traffic flows of multiple sub-areas are obtained; the multiple traffic flows are sorted in descending order to determine the key traffic area; the historical traffic data set is traversed to obtain historical traffic event records of the key traffic area, and multiple sensing points are identified; and multiple traffic sensing sensors are deployed in the key traffic area according to the multiple sensing points to obtain the sensing monitoring module.

[0084] Furthermore, the localization processing module 12 in the multi-robot collaborative system for intelligent transportation is further configured to:

[0085] The edge computing node performs spatiotemporal alignment processing on the received traffic perception data to obtain first perception data; performs anomaly detection on the first perception data, performs data compensation based on the anomaly detection result to obtain second perception data; and performs standardization processing on the second perception data to obtain the standard perception data.

[0086] Furthermore, the localization processing module 12 in the multi-robot collaborative system for intelligent transportation is further configured to:

[0087] The traffic perception data is time-aligned according to the collection timestamp and sampling frequency to obtain traffic time series data; the multi-source sensor is calibrated according to the edge computing node to obtain the spatial position relationship of the multi-source sensors; based on the spatial position relationship and in combination with the collection data characteristics of the multi-source sensors, a spatial coordinate system is constructed; the traffic time series data is projected onto the spatial coordinate system to obtain first perception data.

[0088] Furthermore, the allocation scheme determination module 13 in the multi-robot collaborative system for intelligent transportation is further configured to:

[0089] Based on the traffic demand and the historical traffic data set, multiple traffic task types are determined; based on the importance and timeliness of the multiple traffic task types, task priorities are determined; the target area is divided according to the task priorities to determine multiple sub-areas; based on task load balancing and multiple real-time status information of the multiple robots, multiple task allocation schemes for the multiple sub-areas are matched in the initial task allocation scheme; and the multiple task allocation schemes are integrated to obtain the real-time task allocation scheme.

[0090] Furthermore, the allocation scheme determination module 13 in the multi-robot collaborative system for intelligent transportation is further configured to:

[0091] Step a: randomly select a first sub-area from the multiple sub-areas; step b: extract multiple robot positions from the multiple real-time status information, determine the number and type of robots in the first sub-area in combination with the task priority, and obtain multiple first-area robots; step c: obtain the self-state data of the first robot among the multiple first-area robots, send it to at least one collaborative robot in the same area, and receive the collaborative state data of at least one of the collaborative robots; step d: combine the self-state data and the collaborative state data, and under the first area constraint condition, perform path planning with the goal of optimal path and optimal task execution, and generate a first task allocation plan, wherein the first area constraint condition includes robot load balancing and collaborative obstacle avoidance; traverse the multiple sub-areas, repeat steps a to d, and generate the multiple task allocation plans.

[0092] Furthermore, the multi-robot collaborative system for intelligent transportation further includes a monitoring and feedback module, which is further configured to:

[0093] The task execution status of the multiple robots is monitored in real time, and the task execution status is uploaded to the edge computing node for task evaluation to obtain task feedback information; and the multiple task paths are updated according to the task feedback information.

[0094] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. Figure 1 The multi-robot collaborative method and specific examples of the intelligent transportation system in Example 1 are also applicable to the multi-robot collaborative system in this embodiment. The detailed description of the multi-robot collaborative method for intelligent transportation will clearly indicate the multi-robot collaborative system in this embodiment. For the sake of brevity, a detailed description will not be given here. The system disclosed in this embodiment corresponds to the method disclosed in this embodiment, so the description is relatively simple. For relevant details, refer to the method description.

[0095] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present application. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application is not limited to the embodiments shown herein, but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

[0096] Obviously, for those skilled in the art, several improvements and modifications can be made to the present application without departing from the principles of the present application, and these improvements and modifications also fall within the scope of protection of the present application.

Claims

1. A multi-robot collaboration method for intelligent transportation, characterized in that: include: Obtaining traffic demand in a target area and extracting real-time traffic data from a sensor monitoring module, and generating an initial task allocation plan based on the traffic demand and the real-time traffic data; Traffic perception data is acquired through multi-source sensors carried by multiple robots, and the traffic perception data is uploaded to the edge computing node for local processing to obtain standard perception data; Acquire multiple real-time states of multiple robots, and adjust the initial task allocation plan in combination with the standard perception data to obtain a real-time task allocation plan; Based on the real-time task allocation scheme, multiple task paths of the multiple robots are determined, and the multiple robots are collaboratively managed.

2. The multi-robot collaborative method for intelligent transportation according to claim 1, characterized in that: The sensor monitoring module includes: Based on a historical traffic dataset of the target area, obtaining a plurality of traffic flows in a plurality of sub-areas; sorting the plurality of traffic flows in descending order to determine a key traffic area; Traversing the historical traffic data set, obtaining historical traffic event records of the key traffic area, and identifying multiple sensing points; A plurality of traffic sensing sensors are arranged in the key traffic area according to the plurality of sensing points to obtain the sensing monitoring module.

3. The multi-robot collaborative method for intelligent transportation according to claim 1, characterized in that: The traffic perception data is uploaded to the edge computing node for local processing to obtain standard perception data, including: The edge computing node performs spatiotemporal alignment processing on the received traffic perception data to obtain first perception data; Performing anomaly detection on the first sensed data, and performing data compensation according to the anomaly detection result to obtain second sensed data; The second perception data is standardized to obtain the standard perception data.

4. The multi-robot collaborative method for intelligent transportation according to claim 3, characterized in that: Performing spatiotemporal alignment processing on the received traffic perception data to obtain first perception data includes: Performing time alignment processing on the traffic sensing data according to the acquisition timestamp and sampling frequency to obtain traffic time series data; Calibrate the multi-source sensor according to the edge computing node to obtain the spatial position relationship of the multi-source sensor; Based on the spatial position relationship and in combination with the data acquisition characteristics of the multi-source sensors, a spatial coordinate system is constructed; The traffic time series data is projected onto the spatial coordinate system to obtain first perception data.

5. The multi-robot collaborative method for intelligent transportation according to claim 2, characterized in that: Acquiring multiple real-time status information of multiple robots, combining the standard perception data, and adjusting the initial task allocation plan to obtain a real-time task allocation plan, including: determining a plurality of traffic task types based on the traffic demand and the historical traffic dataset; Determining task priorities based on the importance and timeliness of the multiple traffic task types; Dividing the target area according to the task priority to determine a plurality of sub-areas; Matching the multiple task allocation plans of the multiple sub-areas in the initial task allocation plan based on task load balancing and multiple real-time status information of the multiple robots; The multiple task allocation schemes are integrated to obtain the real-time task allocation scheme.

6. The multi-robot collaborative method for intelligent transportation according to claim 5, characterized in that: Based on task load balancing and multiple real-time status information of the multiple robots, matching multiple task allocation schemes of the multiple sub-areas in the initial task allocation scheme includes: Step a: randomly selecting a first sub-region from the multiple sub-regions; Step b: extracting multiple robot positions from the multiple real-time status information, and determining the number and type of robots in the first sub-area based on the task priority to obtain multiple robots in one area; Step c: acquiring the self-state data of the first robot among the plurality of robots in the first zone, sending the self-state data to at least one collaborative robot in the same zone, and receiving the collaborative state data of at least one collaborative robot; Step d: combining the self-state data and the collaborative state data, and performing path planning with the goal of optimal path and optimal task execution under a first regional constraint, to generate a first task allocation plan, wherein the first regional constraint includes robot load balancing and collaborative obstacle avoidance; Traverse the multiple sub-areas and repeat steps a to d to generate the multiple task allocation solutions.

7. The multi-robot collaboration method for intelligent transportation according to claim 1, characterized in that: Also includes: Monitor the task execution status of the multiple robots in real time, upload the task execution status to the edge computing node for task evaluation, and obtain task feedback information; The multiple task paths are updated according to the task feedback information.

8. A multi-robot collaborative system for intelligent transportation, characterized in that: Steps for implementing the multi-robot collaborative method for intelligent transportation according to any one of claims 1 to 7, wherein the multi-robot collaborative system for intelligent transportation comprises: An initial plan determination module is used to obtain traffic demand in a target area and extract real-time traffic data from a sensor monitoring module, and generate an initial task allocation plan based on the traffic demand and the real-time traffic data; A localization processing module is used to obtain traffic perception data through multi-source sensors carried by multiple robots, upload the traffic perception data to the edge computing node for localization processing, and obtain standard perception data; An allocation scheme determination module is used to obtain multiple real-time states of multiple robots, and adjust the initial task allocation scheme in combination with the standard perception data to obtain a real-time task allocation scheme; A collaborative management module is used to determine multiple task paths of the multiple robots based on the real-time task allocation plan and to collaboratively manage the multiple robots.

Images