Control method and device of intelligent agent, intelligent device and computer storage medium

Abstract

The present application relates to the technical field of artificial intelligence and industrial application, and specifically provides an agent control method and device, an intelligent device and a computer storage medium, aiming to solve the problems of low global positioning accuracy and low motion safety of an agent in a large-scale complex environment. To this end, the agent control method of the present application comprises: in response to a first agent being located in a target area, acquiring real-time point cloud data of the environment near the first agent; performing point cloud matching on the real-time point cloud data and a point cloud map of the target area, determining a real-time position of the first agent in the target area based on the point cloud matching result; updating a first cost map of the target area based on the real-time point cloud data to obtain a second cost map of the target area; and based on the real-time position, the second cost map and a target position of the first agent, performing path planning on the first agent to obtain a planned path of the first agent.

Description

Technical Field

[0001] This invention relates to the fields of artificial intelligence and industrial application technology, and specifically to a control method, device, intelligent equipment, and computer storage medium for an intelligent agent. Background Technology

[0002] Enabling intelligent agents (such as robots / automated guided vehicles, AGVs) to navigate autonomously and avoid obstacles in complex environments is an important technological direction for fields such as intelligent manufacturing, smart logistics, and warehouse management.

[0003] In traditional production and logistics processes, material handling and delivery typically rely on manual operation of forklifts or trolleys. While this method offers some flexibility, it is overly dependent on human experience, resulting in high labor intensity, low efficiency, and high operating costs, making it difficult to meet the requirements of large-scale flexible production for both high efficiency and low cost.

[0004] Current automated guided vehicles (AGVs) typically rely on magnetic strips, QR codes, reflectors, or preset paths for navigation. This approach suffers from rigid paths, poor environmental adaptability, and difficulty in handling dynamically changing scenarios, limiting its application in unstructured or semi-structured environments.

[0005] Some intelligent agents use Simultaneous Localization and Mapping (SLAM) technology for mapping and localization, which suffers from low localization accuracy and low motion safety.

[0006] Accordingly, the field needs a new solution to ensure the global localization and path execution accuracy of intelligent agents in large-scale complex environments to address the above problems. Summary of the Invention

[0007] To overcome the above-mentioned shortcomings, this invention is proposed to solve or at least partially solve the technical problems of low global positioning accuracy and low motion safety of intelligent agents in large-scale complex environments.

[0008] In a first aspect, the present invention provides a method for controlling an intelligent agent, comprising:

[0009] In response to the first intelligent agent being located within the target area, real-time point cloud data of the environment near the first intelligent agent is acquired;

[0010] Point cloud matching is performed between the real-time point cloud data and the point cloud map of the target area, and the real-time position of the first intelligent agent in the target area is determined based on the point cloud matching result.

[0011] The first cost map of the target area is updated based on the real-time point cloud data to obtain the second cost map of the target area.

[0012] Based on the real-time location, the second cost map, and the target location of the first agent, path planning is performed on the first agent to obtain the planned path of the first agent.

[0013] In some embodiments of the present invention, before performing point cloud matching on the real-time point cloud data and the point cloud map of the target area, and determining the real-time position of the first intelligent agent in the target area based on the point cloud matching result, the method further includes:

[0014] Based on the real-time environmental point cloud data collected by the second agent while it moves within the target area and the pose of the second agent, an environmental map of the target area is constructed using an instantaneous localization and mapping method, resulting in a point cloud map of the target area.

[0015] In some embodiments of the present invention, before updating the first cost map of the target area based on the real-time point cloud data to obtain the second cost map of the target area, the method further includes:

[0016] The point cloud map of the target area is subjected to planar projection processing, and the first cost map is obtained based on the planar projection processing result.

[0017] In some embodiments of the present invention, updating the first cost map of the target area based on the real-time point cloud data to obtain a second cost map of the target area includes:

[0018] The real-time point cloud data is converted into two-dimensional laser scanning data;

[0019] Based on the two-dimensional laser scanning data, the real-time location of obstacles near the first intelligent agent is determined;

[0020] Based on the real-time location and preset safe distance of obstacles near the first agent, the first cost map is updated to obtain the second cost map.

[0021] In some embodiments of the present invention, the step of performing path planning on the first intelligent agent based on the real-time location, the second cost map, and the target location of the first intelligent agent to obtain the planned path of the first intelligent agent includes:

[0022] Obtain the first planned path of the first intelligent agent;

[0023] Using a time-elastic band trajectory optimizer, based on the real-time location, the second cost map, and the target location of the first agent, the first planned path of the first agent is adjusted to obtain the second planned path of the first agent, wherein the planned path of the first agent includes the second planned path.

[0024] In some embodiments of the present invention, after performing path planning on the first agent based on the real-time location, the second cost map, and the target location of the first agent to obtain the planned path of the first agent, the method further includes:

[0025] The first intelligent agent is controlled to move based on the planned path, so as to execute a preset task after the first intelligent agent reaches the target location;

[0026] The task execution status and results of the first intelligent agent are updated and fed back.

[0027] In some embodiments of the present invention, updating and providing feedback on the task execution status of the first intelligent agent includes: in response to detecting a delay or interruption in the task of the first intelligent agent, recording and issuing an anomaly alarm in real time.

[0028] In a second aspect, the present invention provides a control device for an intelligent agent, comprising:

[0029] The point cloud data acquisition module is used to acquire real-time point cloud data of the environment near the first intelligent agent in response to the first intelligent agent being located in the target area;

[0030] The positioning module is used to perform point cloud matching between the real-time point cloud data and the point cloud map of the target area, and determine the real-time position of the first intelligent agent in the target area based on the point cloud matching result.

[0031] The cost map update module is used to update the first cost map of the target area based on the real-time point cloud data to obtain the second cost map of the target area.

[0032] The path planning module is used to perform path planning for the first agent based on the real-time location, the second cost map, and the target location of the first agent, so as to obtain the planned path of the first agent.

[0033] In some embodiments of the present invention, the apparatus further includes:

[0034] The point cloud map creation module is used to perform environmental mapping of the target area based on real-time environmental point cloud data collected when the second agent moves within the target area and the pose of the second agent, using an instant localization and mapping method, to obtain a point cloud map of the target area.

[0035] In some embodiments of the present invention, the apparatus further includes:

[0036] The cost map acquisition module is used to perform planar projection processing on the point cloud map of the target area, and obtain the first cost map based on the planar projection processing result.

[0037] In some embodiments of the present invention, the cost map update module is used to convert the real-time point cloud data into two-dimensional laser scan data; the cost map update module is also used to determine the real-time position of obstacles near the first intelligent agent based on the two-dimensional laser scan data; the cost map update module is also used to update the first cost map based on the real-time position of obstacles near the first intelligent agent and a preset safety distance to obtain the second cost map.

[0038] In some embodiments of the present invention, the path planning module is used to obtain a first planned path of the first agent; the path planning module is also used to use a time elastic band trajectory optimizer to adjust the first planned path of the first agent based on the real-time location, the second cost map and the target location of the first agent, to obtain a second planned path of the first agent, wherein the planned path of the first agent includes the second planned path.

[0039] In some embodiments of the present invention, the apparatus further includes:

[0040] A control module is used to control the first intelligent agent to move based on the planned path, so as to execute a preset task after the first intelligent agent reaches the target position;

[0041] The feedback update module is used to update and provide feedback on the task execution status and results of the first intelligent agent.

[0042] In some embodiments of the present invention, the feedback update module is used to record and issue an anomaly alarm in real time in response to the detection of task delay or task interruption of the first intelligent agent.

[0043] When the above technical solution is adopted, the agent obtains its real-time position by loading the real-time point cloud data collected by the agent and combining it with the point cloud map of the target area during the movement of the agent. The cost map is updated by the real-time point cloud data. Then, path planning or path adjustment is performed based on the real-time position of the agent and the updated cost map. Compared with traditional positioning and path planning methods, this can effectively reduce the accumulated error and ensure the global positioning and path execution accuracy of the robot in large-scale complex environments.

[0044] Furthermore, this invention can perceive and dynamically update information about dynamic obstacles near the agent in real time. When a dynamic obstacle disappears, it is promptly cleared from the cost map, preventing the agent from frequently stopping abruptly or getting stuck in local dead zones due to lingering obstacle information in dynamic environments. Combined with the time constraint optimization of the TEB local planner, the robot can still achieve smooth and safe trajectory adjustments in complex environments with continuous movement of people and obstacles, thereby significantly improving the system's adaptability to dynamic environments. The agent can maintain continuity and fluency during task execution, significantly improving overall operational efficiency and task completion reliability.

[0045] Furthermore, this invention adopts a fusion approach of global path planning and local trajectory optimization. The global path ensures the accessibility and optimality of the task objective, while local planning (TEB) optimizes the trajectory in real time based on the latest obstacle information, enabling the robot to maintain the global task direction without deviation and flexibly avoid dynamic obstacles, achieving a navigation effect of "global stability and local flexibility".

[0046] Thanks to the point cloud and SLAM fusion positioning and dynamic mapping capabilities, the intelligent agent can operate stably in environments with high population density and complex obstacles, such as workshops, warehouses, hospitals, and shopping malls. This overcomes the limitations of traditional navigation methods and greatly expands the application scenarios of intelligent agents in intelligent manufacturing, smart logistics, and public services. Attached Figure Description

[0047] The preferred embodiments of the present invention are described below with reference to the accompanying drawings, in which:

[0048] Figure 1 This is a flowchart illustrating the control method of the intelligent agent in some embodiments of the present invention;

[0049] Figure 2 This is a schematic diagram of trajectory optimization in one example of the present invention;

[0050] Figure 3 This is a structural block diagram of the control device for the intelligent agent in some embodiments of the present invention;

[0051] Figure 4 This is a structural block diagram of the intelligent agent in some embodiments of the present invention. Detailed Implementation

[0052] Some embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0053] In the description of this invention, "module" and "processor" can include hardware, software, or a combination of both. A module can include hardware circuitry, various suitable sensors, communication ports, memory, and may also include software components, such as program code, or a combination of software and hardware. A processor can be a central processing unit, microprocessor, image processor, digital signal processor, or any other suitable processor. The processor has data and / or signal processing capabilities. The processor can be implemented in software, in hardware, or a combination of both. Computer-readable storage media include any suitable medium capable of storing program code, such as magnetic disks, hard disks, optical disks, flash memory, read-only memory, random access memory, etc. The term "A and / or B" means all possible combinations of A and B, such as only A, only B, or A and B. The terms "at least one A or B" or "at least one of A and B" have a similar meaning to "A and / or B" and can include only A, only B, or A and B. The singular terms "a" or "this" can also include plural forms.

[0054] Here we will first explain some of the terms involved in this invention.

[0055] SLAM is a core technology that enables robots to locate themselves in unknown environments and build environmental maps in real time using sensors. Its core principle is to fuse measurement data from sensors (such as LiDAR, cameras, and millimeter-wave radar) to simultaneously estimate the robot's own trajectory and the coordinates of environmental features.

[0056] AGVs are intelligent logistics equipment based on automatic navigation technology, which achieve autonomous movement through technologies such as magnetic strips, lasers, RFID, and SLAM.

[0057] An Inertial Measurement Unit (IMU) is a device that measures an object's three-axis attitude angles (or angular rates) and acceleration. An IMU typically contains three single-axis accelerometers and three single-axis gyroscopes. The accelerometers detect the object's acceleration signals along three independent axes of the carrier's coordinate system, while the gyroscopes detect the carrier's angular velocity signals relative to the navigation coordinate system. By measuring the object's angular velocity and acceleration in three-dimensional space, the object's attitude can be calculated.

[0058] Robot Operating System 2 (ROS2) is a software library and toolset for building robot applications. It is an improvement and update of ROS1 and is suitable for production environments such as industrial automation, healthcare, and autonomous driving.

[0059] The Timed Elastic Band (TEB) trajectory optimizer performs subsequent corrections on the initial trajectory generated by the global path planner, thereby optimizing the robot's motion trajectory. It belongs to local path planning.

[0060] Figure 1 This is a flowchart illustrating the control method of the intelligent agent in some embodiments of the present invention. For example... Figure 1 As shown, the control method for the intelligent agent in this embodiment of the invention includes the following steps:

[0061] S1: In response to the first agent being located within the target area, acquire real-time point cloud data of the environment near the first agent.

[0062] In this embodiment, the first intelligent agent may include at least one of a robot, an automated guided vehicle (AGV), and a robotic dog, and has autonomous navigation and obstacle avoidance capabilities. The target area may be the working area of ​​the first intelligent agent, such as a portion of a smart factory.

[0063] There are several ways to determine whether the first intelligent agent is located within the target area. For example, a monitoring system installed within the target area can detect the first intelligent agent's presence within the system's image acquisition range to determine if the first intelligent agent is located within the target area. Alternatively, a marker scanning device installed within the target area can scan the marker on the first intelligent agent to determine if the first intelligent agent is located within the target area. A positioning device installed on the first intelligent agent can also be used to determine if the first intelligent agent is located within the target area.

[0064] After determining that the first intelligent agent is located within the target area, a radar scanning device installed on the first intelligent agent collects real-time point cloud data of the vicinity of the first intelligent agent through radar scanning. The radar scanning device may include at least one of lidar and millimeter-wave radar.

[0065] S2: Perform point cloud matching between real-time point cloud data and point cloud map of target area, and determine the real-time position of the first agent in target area based on the point cloud matching result.

[0066] In this embodiment, point cloud matching is performed between real-time point cloud data and local areas in the point cloud map of the target area. For example, the optimal pose transformation between real-time point cloud data and point cloud map can be calculated by point cloud registration methods such as Iterative Closest Point (ICP) or Normal Distribution Transform (NDT) to determine the position of the first agent in the target area.

[0067] During point cloud matching, the initial approximate position of the first agent within the target area can be manually specified in a visualization interface (such as RViz) to obtain an initial pose estimate. A local point cloud region in the point cloud map near this initial pose is then selected for matching calculations, thereby improving matching speed and stability. Subsequently, the pose is continuously optimized through iterative point cloud matching algorithms to minimize the distance error between the real-time point cloud data and the point cloud at the location on the point cloud map, thus obtaining the precise pose of the first agent in the map coordinate system.

[0068] By matching and correcting real-time point cloud data with a global point cloud map, the cumulative odometer error generated during the movement of the agent can be periodically corrected, thereby effectively reducing positioning drift during long-term operation and improving the positioning accuracy and stability of the agent in complex environments.

[0069] In one implementation, before S2, the following steps may be included: based on the real-time environmental point cloud data collected when the second agent moves within the target area and the pose of the second agent, an environmental map of the target area is constructed using an instantaneous localization and mapping method to obtain a point cloud map of the target area.

[0070] In this embodiment, during the initial mapping phase of the target area, the second agent is controlled to move within the target area. A LiDAR system mounted on the second agent acquires 3D point cloud data of the surrounding environment, and this data is combined with IMU data stored within the second agent to perform SLAM mapping, generating a high-precision 3D point cloud map (PCD). During SLAM mapping, the LiDAR point cloud data and IMU data are synchronized in time. Furthermore, during the movement of the second agent, attitude change information provided by the IMU is used to assist in correcting the LiDAR point cloud data, reducing point cloud distortion errors caused by the movement of the second agent. The second agent can be the same as or different from the first agent.

[0071] During SLAM mapping, the currently acquired LiDAR point cloud is matched with the established local map. The pose of the second agent in the map is calculated using the point cloud matching algorithm, and the pose is corrected by combining the pose information provided by the IMU, thereby obtaining a more stable localization result.

[0072] By using the above-mentioned multi-source data (point cloud data + IMU data) fusion method, the cumulative error generated during the long-term mapping process can be effectively reduced, the accuracy and stability of the 3D point cloud map can be improved, and a reliable environmental map can be provided for subsequent intelligent agent navigation and path planning.

[0073] S3: Update the first cost map of the target area based on real-time point cloud data to obtain the second cost map of the target area.

[0074] In this embodiment, the first cost map of the target area can be the initial cost map of the target area, that is, the first cost map is obtained by performing planar projection processing on the 3D point cloud map obtained after mapping the target area environment, based on the planar projection processing result. Alternatively, the first cost map of the target area can also be the cost map of the target area after the last update.

[0075] Obtaining the initial cost map may include the following steps:

[0076] The 3D point cloud map obtained by SLAM mapping is filtered by height range, retaining only the point cloud data within the set height range to remove ground noise and irrelevant structures at high altitudes.

[0077] The filtered point cloud is subjected to noise filtering, and the point cloud after noise filtering is corrected by coordinate rotation and translation according to the navigation coordinate system of the second intelligent agent so that the point cloud map is aligned with the navigation coordinate system.

[0078] The processed 3D point cloud is projected vertically onto a 2D plane and divided into grids according to a preset map resolution. When a grid contains more point cloud data than a preset threshold, the grid is marked as occupied; when a grid contains no point cloud data, it is marked as an empty or unknown area, thereby generating a 2D grid map for intelligent agent navigation, i.e., the first cost map.

[0079] Since real-time point cloud data can include the location and size information of dynamic obstacles in the environment near the first agent, as well as the location and size information of newly appearing static obstacles such as static debris, the real-time point cloud data is projected onto the plane where the first cost map is located for update processing to obtain the second cost map of the target area.

[0080] In one implementation, S3 may include the following steps:

[0081] S3-1: Convert real-time point cloud data into two-dimensional laser scan data.

[0082] S3-2: Determine the real-time location of obstacles near the first intelligent agent based on two-dimensional laser scanning data.

[0083] S3-3: Based on the real-time location of obstacles near the first agent and the preset safe distance, update the first cost map to obtain the second cost map.

[0084] Real-time point cloud data is converted into two-dimensional laser scan data (LaserScan) to facilitate obstacle detection and map updates in the navigation system. Based on the converted laser scan data, obstacle information is updated in the local cost map surrounding the first agent within the first cost map. When the scan data detects a new obstacle, the corresponding area is marked as occupied, and an obstacle region is generated in the cost map. Combined with a preset distance threshold representing the safe distance near the obstacle, a second cost map is generated, enabling the first agent to plan new obstacle avoidance paths in a timely manner.

[0085] In scenarios where dynamic obstacles exist in the target area, when a person or mobile device enters the perception range of the first intelligent agent, its corresponding area is quickly marked as an obstacle area. Once these dynamic obstacles leave the sensor's field of view, due to the system's non-persistent voxel update mechanism, if an area is not detected again in subsequent perception cycles, its occupancy status is automatically cleared, thus preventing dynamic obstacles from leaving residues or trailing on the map. Through this costly map update mechanism, while ensuring the real-time nature of the local map, false obstacle information generated by dynamic obstacles is effectively reduced, thereby improving the obstacle avoidance accuracy and operational safety of the intelligent agent in complex industrial environments.

[0086] In one implementation, the layered structure of the second cost map may include: a static layer, a non-persistent voxel layer, and an inflated layer.

[0087] The static layer stores the 2D raster map information generated by SLAM mapping. This map is converted from the PCD point cloud map and mainly contains fixed environmental structures such as walls and equipment. This layer serves as the basic environmental map, providing a global reference for path planning.

[0088] A non-persistent voxel layer is used to process dynamic obstacle information around the first agent. Existing non-persistent voxel layer plugins are used to implement dynamic obstacle updates, processing data acquired in real-time by sensors. When a new obstacle is detected near the first agent, the corresponding area is marked as an obstacle region. When the obstacle is not detected again in subsequent sensing, the obstacle information is automatically cleared, thus preventing obstacle information from remaining on the map after people or equipment have left. In this way, real-time updates of the dynamic environment can be achieved.

[0089] The expansion layer is used to generate a safe distance around obstacles. When a grid cell is marked as an obstacle, the system expands the cost region within a certain range around it, allowing the robot to automatically maintain a safe distance from obstacles during path planning, thereby improving navigation safety.

[0090] The static layer, non-persistent voxel layer, and dilated layer are superimposed and fused to form the final second cost map. This second cost map contains both static environment information and real-time dynamic obstacle information, and serves as input to path planning algorithms (such as the TEB trajectory planner). Through this hierarchical structure, the first agent can achieve real-time obstacle avoidance in dynamic environments while maintaining the stability of the static map.

[0091] S4: Based on the real-time location, the second cost map, and the target location of the first agent, perform path planning for the first agent to obtain the planned path of the first agent.

[0092] In this embodiment, TEB is used to perform path planning for the first agent. TEB represents the planned path of the first agent from its real-time location to its target location as a series of discrete trajectory points, and introduces time interval constraints between adjacent trajectory points, thereby forming a trajectory sequence containing spatial location and time information.

[0093] During trajectory optimization, TEB comprehensively considers the following constraints: robot kinematic constraints (velocity and acceleration limits), obstacle distance constraints (maintaining a safe distance from environmental obstacles), time optimization constraints (minimizing motion time as much as possible), and trajectory smoothness constraints (reducing unnecessary turning or oscillation).

[0094] By comprehensively optimizing the above constraints, TEB continuously adjusts the position and time interval of trajectory points, enabling the first agent to generate a smooth and executable motion trajectory while avoiding obstacles.

[0095] In one implementation, S4 may include the following steps:

[0096] S4-1: Obtain the first planned path for the first agent.

[0097] When the first agent moves from the initial planned position to the current real-time position along the initial planned trajectory and no obstacle affecting the first agent's movement is found (there is no obstacle on this movement trajectory), the first planned path can be the initial planned path.

[0098] When the first agent encounters an obstacle that requires changing its initial planned trajectory while moving from its initial planned position to its current real-time position, the first planned path becomes the planned path updated last time.

[0099] S4-2: Using TEB, based on the real-time location, the second cost map, and the target location of the first agent, adjust the first planned path of the first agent to obtain the second planned path of the first agent. The planned path of the first agent includes the second planned path.

[0100] Figure 2 This is a schematic diagram of trajectory optimization in one example of the present invention. For example... Figure 2 As shown, the TEB interacts with the cost map (e.g., the first cost map) in real time. When a new obstacle is detected, the TEB recalculates the local trajectory based on the updated cost map (e.g., the second cost map), enabling the first agent to dynamically adjust its direction of travel, thereby achieving real-time obstacle avoidance.

[0101] Because quadruped robots have high mobility and a large range of posture changes during movement, the TEB planning method can simultaneously consider spatial path and time constraints, making it more suitable for trajectory planning and dynamic obstacle avoidance of quadruped robots in complex environments.

[0102] In some embodiments, the following steps may be included after S4:

[0103] S5: Control the first intelligent agent to move according to the planned path, so as to execute a preset task after the first intelligent agent reaches the target location. The preset task may include detecting whether the target location is abnormal through image acquisition and image recognition, placing a specified item carried by the first intelligent agent at the target location, etc.

[0104] S6: Update and provide feedback on the task execution status and results of the first agent. The task execution status may include: task in progress, task delayed, task interrupted, etc.

[0105] In some implementations, S6 may include the following steps: in response to detecting a task delay or interruption of the first agent, record and issue an anomaly alarm in real time. For example, it can be determined whether a task delay has occurred based on the planned task duration and the current task duration, or it can be determined whether a task interruption has occurred based on the current task being executed by the first agent (the first agent has executed other tasks without completing its original task or the pending task is empty).

[0106] When the above technical solution is adopted, the agent obtains its real-time position by loading the real-time point cloud data collected by the agent and combining it with the point cloud map of the target area during the movement of the agent. The cost map is updated by the real-time point cloud data. Then, path planning or path adjustment is performed based on the real-time position of the agent and the updated cost map. Compared with traditional positioning and path planning methods, this can effectively reduce the accumulated error and ensure the global positioning and path execution accuracy of the robot in large-scale complex environments.

[0107] Furthermore, this invention can perceive and dynamically update information about dynamic obstacles near the agent in real time. When a dynamic obstacle disappears, it is promptly cleared from the cost map, preventing the agent from frequently stopping abruptly or getting stuck in local dead zones due to lingering obstacle information in dynamic environments. Combined with the time constraint optimization of the TEB local planner, the robot can still achieve smooth and safe trajectory adjustments in complex environments with continuous movement of people and obstacles, thereby significantly improving the system's adaptability to dynamic environments. The agent can maintain continuity and fluency during task execution, significantly improving overall operational efficiency and task completion reliability.

[0108] Furthermore, this invention adopts a fusion approach of global path planning and local trajectory optimization. The global path ensures the accessibility and optimality of the task objective, while local planning (TEB) optimizes the trajectory in real time based on the latest obstacle information, enabling the robot to maintain the global task direction without deviation and flexibly avoid dynamic obstacles, achieving a navigation effect of "global stability and local flexibility".

[0109] Thanks to the point cloud and SLAM fusion positioning and dynamic mapping capabilities, the intelligent agent can operate stably in environments with high population density and complex obstacles, such as workshops, warehouses, hospitals, and shopping malls. This overcomes the limitations of traditional navigation methods and greatly expands the application scenarios of intelligent agents in intelligent manufacturing, smart logistics, and public services.

[0110] It should be noted that although the steps in the above embodiments are described in a specific order, those skilled in the art will understand that in order to achieve the effects of the present invention, different steps do not necessarily have to be executed in such an order. They can be executed simultaneously (in parallel) or in other orders. These adjusted solutions are equivalent to the technical solutions described in the present invention and therefore will also fall within the protection scope of the present invention.

[0111] Those skilled in the art will understand that all or part of the processes in the method of the above-described embodiment of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable storage medium can include any entity or device capable of carrying computer program code, media, USB flash drive, portable hard drive, magnetic disk, optical disk, computer memory, read-only memory, random access memory, electrical carrier signals, telecommunication signals, and software distribution media, etc.

[0112] Figure 3 This is a structural block diagram of the control device for the intelligent agent in some embodiments of the present invention. For example... Figure 3 As shown, the control device for the intelligent agent includes:

[0113] The point cloud data acquisition module 100 is used to acquire real-time point cloud data of the environment near the first intelligent agent in response to the first intelligent agent being located in the target area;

[0114] The positioning module 200 is used to perform point cloud matching between real-time point cloud data and point cloud map of the target area, and determine the real-time position of the first intelligent agent in the target area based on the point cloud matching result.

[0115] The cost map update module 300 is used to update the first cost map of the target area based on real-time point cloud data to obtain the second cost map of the target area.

[0116] The path planning module 400 is used to perform path planning for the first intelligent agent based on the real-time location, the second cost map, and the target location of the first intelligent agent, so as to obtain the planned path of the first intelligent agent.

[0117] In some embodiments of the present invention, the control device for the intelligent agent further includes:

[0118] The point cloud map creation module is used to create an environmental map of the target area based on the real-time environmental point cloud data collected when the second agent moves within the target area and the pose of the second agent, using instant localization and map building methods.

[0119] In some embodiments of the present invention, the control device for the intelligent agent further includes:

[0120] The cost map acquisition module is used to perform planar projection processing on the point cloud map of the target area, and obtain the first cost map based on the planar projection processing result.

[0121] In some embodiments of the present invention, the cost map update module 300 is used to convert real-time point cloud data into two-dimensional laser scan data; the cost map update module 300 is also used to determine the real-time position of obstacles near the first intelligent agent based on the two-dimensional laser scan data; the cost map update module 300 is also used to update the first cost map based on the real-time position of obstacles near the first intelligent agent and a preset safety distance to obtain a second cost map.

[0122] In some embodiments of the present invention, the path planning module 400 is used to obtain a first planned path of the first intelligent agent; the path planning module 400 is also used to use a time elastic band trajectory optimizer to adjust the first planned path of the first intelligent agent based on the real-time location, the second cost map and the target location of the first intelligent agent to obtain a second planned path of the first intelligent agent, wherein the planned path of the first intelligent agent includes the second planned path.

[0123] In some embodiments of the present invention, the control device for the intelligent agent further includes:

[0124] The control module is used to control the first intelligent agent to move according to the planned path, so as to execute a preset task after the first intelligent agent reaches the target position;

[0125] The feedback update module is used to update and provide feedback on the task execution status and results of the first intelligent agent.

[0126] In some embodiments of the present invention, the feedback update module is used to record and issue an anomaly alarm in real time in response to the detection of a delay or interruption in the task of the first intelligent agent.

[0127] Another aspect of the present invention provides a computer-readable storage medium.

[0128] It should be noted that the specific implementation of the control device of the intelligent agent in this embodiment is similar to the specific implementation of the control method of the intelligent agent in this embodiment, and the technical effects of the control device of the intelligent agent in this embodiment are similar to the technical effects of the control method of the intelligent agent in this embodiment. For details, please refer to the description of the control method of the intelligent agent. In order to reduce redundancy, it will not be repeated.

[0129] In one embodiment of a computer-readable storage medium according to the present invention, the computer-readable storage medium may be configured to store a program for executing the control method of an intelligent agent in the above-described method embodiments. This program may be loaded and run by a processor to implement the control method of the intelligent agent. For ease of explanation, only the parts related to the embodiments of the present invention are shown; for specific technical details not disclosed, please refer to the method section of the embodiments of the present invention. The computer-readable storage medium may be a storage device comprising various electronic devices. Optionally, in the embodiments of the present invention, the computer-readable storage medium is a non-transitory computer-readable storage medium.

[0130] Another aspect of the present invention provides a smart device.

[0131] In one embodiment of a smart device according to the present invention, the smart device may include at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program, which, when executed by the at least one processor, implements the method described in any of the above embodiments. The smart device of the present invention may include driving equipment, intelligent vehicles, robots, and other devices. See Appendix Figure 4 , Figure 4 The example shows a memory 10 and a processor 20 connected via a bus communication connection.

[0132] In some embodiments of the present invention, the smart device may further include at least one sensor for sensing information. The sensor is communicatively connected to any type of processor mentioned in the present invention. Optionally, the smart device may further include an autonomous driving system for guiding the smart device to drive autonomously or assisting in driving. The processor communicates with the sensor and / or the autonomous driving system to perform the methods described in any of the above embodiments.

[0133] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after such changes or substitutions will all fall within the scope of protection of the present invention.

Claims

1. A control method for an intelligent agent, characterized in that, include: In response to the first intelligent agent being located within the target area, real-time point cloud data of the environment near the first intelligent agent is acquired; Point cloud matching is performed between the real-time point cloud data and the point cloud map of the target area, and the real-time position of the first intelligent agent in the target area is determined based on the point cloud matching result. The first cost map of the target area is updated based on the real-time point cloud data to obtain the second cost map of the target area. Based on the real-time location, the second cost map, and the target location of the first agent, path planning is performed on the first agent to obtain the planned path of the first agent.

2. The method according to claim 1, characterized in that, Before performing point cloud matching on the real-time point cloud data and the point cloud map of the target area, and determining the real-time position of the first intelligent agent in the target area based on the point cloud matching result, the method further includes: Based on the real-time environmental point cloud data collected by the second agent while it moves within the target area and the pose of the second agent, an environmental map of the target area is constructed using an instantaneous localization and mapping method, resulting in a point cloud map of the target area.

3. The method according to claim 1, characterized in that, Before updating the first cost map of the target area based on the real-time point cloud data to obtain the second cost map of the target area, the method further includes: The point cloud map of the target area is subjected to planar projection processing, and the first cost map is obtained based on the planar projection processing result.

4. The method according to claim 1, characterized in that, The step of updating the first cost map of the target area based on the real-time point cloud data to obtain the second cost map of the target area includes: The real-time point cloud data is converted into two-dimensional laser scanning data; Based on the two-dimensional laser scanning data, the real-time location of obstacles near the first intelligent agent is determined; Based on the real-time location and preset safe distance of obstacles near the first agent, the first cost map is updated to obtain the second cost map.

5. The method according to claim 1, characterized in that, The step of performing path planning on the first agent based on the real-time location, the second cost map, and the target location of the first agent to obtain the planned path of the first agent includes: Obtain the first planned path of the first intelligent agent; Using a time-elastic band trajectory optimizer, based on the real-time location, the second cost map, and the target location of the first agent, the first planned path of the first agent is adjusted to obtain the second planned path of the first agent, wherein the planned path of the first agent includes the second planned path.

6. The method according to claim 1 or 5, characterized in that, After performing path planning on the first agent based on the real-time location, the second cost map, and the target location of the first agent to obtain the planned path of the first agent, the method further includes: The first intelligent agent is controlled to move based on the planned path, so as to execute a preset task after the first intelligent agent reaches the target location; The task execution status and results of the first intelligent agent are updated and fed back.

7. The method according to claim 6, characterized in that, The step of updating and providing feedback on the task execution status of the first intelligent agent includes: In response to the detection of task delay or interruption of the first intelligent agent, the system records the information in real time and issues an alarm for abnormality.

8. A control device for an intelligent agent, characterized in that, include: The point cloud data acquisition module is used to acquire real-time point cloud data of the environment near the first intelligent agent in response to the first intelligent agent being located in the target area; The positioning module is used to perform point cloud matching between the real-time point cloud data and the point cloud map of the target area, and determine the real-time position of the first intelligent agent in the target area based on the point cloud matching result. The cost map update module is used to update the first cost map of the target area based on the real-time point cloud data to obtain the second cost map of the target area. The path planning module is used to perform path planning for the first agent based on the real-time location, the second cost map, and the target location of the first agent, so as to obtain the planned path of the first agent.

9. A smart device, characterized in that, include: At least one processor; And, a memory communicatively connected to the at least one processor; The memory stores a computer program that, when executed by the at least one processor, implements the method of any one of claims 1 to 7.

10. A computer-readable storage medium storing a plurality of program codes, characterized in that, The program code is adapted to be loaded and run by a processor to perform the method of any one of claims 1 to 7.

Images