Autonomous navigation patrol robot and remote control system based on three-dimensional laser radar and GNSS

Abstract

The invention relates to an autonomous navigation patrol robot based on a three-dimensional laser radar and a GNSS and a remote control system, and belongs to the technical field of intelligent patrol and robots. The objective of the invention is to solve the problems of inaccurate positioning, weak obstacle avoidance capability and lack of remote cooperative control of an existing patrol robot in a complex environment. The robot integrates a three-dimensional laser radar, a global navigation satellite system, an inertial measurement unit and a camera, and realizes high-precision map construction, autonomous path planning and dynamic obstacle avoidance in combination with an automatic driving navigation framework. A matched remote control system supports multi-robot task scheduling, state monitoring and video viewing. The system has the advantages of being high in environmental adaptability, high in patrol efficiency, convenient to deploy and maintain, flexible in remote operation and the like, and is suitable for intelligent patrol application in scenes such as parks, campuses and commercial streets.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of intelligent patrol and robot technology, and relates to an autonomous navigation patrol robot based on a three-dimensional laser radar and GNSS and a remote control system. BACKGROUND

[0002] With the continuous advancement of smart city construction, open or semi-open public spaces such as parks, commercial districts, and campuses are increasing. These scenarios have characteristics such as wide spatial range, high personnel density, and complex structure, and higher requirements for the real-time, coverage, and intelligence of security patrol.

[0003] Currently, traditional security patrol mainly relies on a combination of manual patrol and fixed-point monitoring. Manual patrol has problems such as limited patrol frequency, high personnel cost, and fatigue leading to neglect of abnormal situations. At the same time, fixed video monitoring often has monitoring blind spots due to limited camera viewing angles, and cannot respond to emergencies in a timely and autonomous manner.

[0004] In recent years, patrol robots have begun to be applied in some scenarios as an auxiliary or alternative to manual patrol. However, existing patrol robot solutions still have many technical limitations. Some robots use only two-dimensional laser radars to obtain environmental information, making it difficult to perceive and understand complex three-dimensional structures in outdoor environments such as trees, building pillars, or suspended obstacles, thus having obvious deficiencies in path planning and obstacle avoidance capabilities. Some solutions also rely on the Global Positioning System (GPS) for positioning, but in areas with dense buildings or severe obstructions, GPS signals are weak or lost, leading to unstable or even failed positioning.

[0005] In addition, most existing patrol robots lack deep integration with remote management platforms, often only supporting basic single-machine operations, and cannot realize comprehensive functions such as multi-machine coordination and scheduling and remote video monitoring, making it difficult to meet the needs of large-scale intelligent patrol systems.

[0006] Therefore, there is an urgent need for an autonomous navigation patrol robot system suitable for outdoor park environments, with high-precision positioning capabilities, strong adaptability to complex terrains, and the ability to be managed uniformly through a remote platform, to improve patrol efficiency, reduce labor costs, and enhance the intelligence level of security systems. SUMMARY

[0007] Therefore, the present application provides an autonomous navigation patrol robot based on a three-dimensional laser radar and GNSS and a remote control system.

[0008] To achieve the above-mentioned purposes, the present application provides the following technical solutions:

[0009] An autonomous navigation patrol robot based on three-dimensional laser radar and GNSS, comprising:

[0010] A differential drive chassis for robot movement;

[0011] An upper mechanical structure mounted on the differential drive chassis for carrying a computing unit and a plurality of sensors;

[0012] A perception system comprising a three-dimensional laser radar, a GNSS receiving module, an inertial measurement unit (IMU) and at least one network camera, the perception system being used to obtain environmental perception information and pose information;

[0013] A computing unit connected to the perception system for performing positioning, path planning and obstacle avoidance control based on the Autoware autonomous driving framework;

[0014] A communication system comprising a gigabit switch and a 4G / 5G cellular network communication module for data transmission between the robot and the remote operation platform;

[0015] A power supply system using an independent dual battery structure, one of which supplies power to the chassis system, and the other supplies power to the perception and computing devices after being stepped down.

[0016] Further, the autonomous navigation performed by the computing unit comprises:

[0017] Starting the positioning step to obtain the initial pose by fusing GNSS and IMU;

[0018] The real-time positioning step matches the real-time laser radar point cloud with the pre-built map to realize continuous positioning;

[0019] The path planning step generates a global path according to the task target point and the map topological information;

[0020] The local path re-planning and obstacle avoidance step dynamically adjusts the local path based on real-time perception information;

[0021] The control instruction generation step converts the planning result into chassis motor control signals.

[0022] Further, the matching of the laser radar point cloud and the map uses the normal distribution transformation matching algorithm NDT or the VGICP algorithm.

[0023] An environment map construction method based on the robot, comprising:

[0024] Constructing a laser inertial odometry (LIO) based on laser radar and IMU to obtain a low-drift trajectory by optimizing the objective function; the objective function J is represented as:

[0025]

[0026] wherein, represents a laser radar matching factor; represents an IMU pre-integration factor; represents an edge factor; n is the number of states within the optimization window;

[0027] A GNSS factor and a loop closure factor are introduced for global optimization, wherein the objective function of the GNSS factor is represented as:

[0028] J GNSS = (T - T GNSS ) T Σ -1 (T - T GNSS )

[0029] wherein, T is the optimized pose, T GNSS is the measured position provided by GNSS, and Σ is the GNSS measurement covariance matrix;

[0030] A globally consistent point cloud map and a vector map are output, which are respectively used for perception matching and path planning.

[0031] A remote monitoring and operation system applied to the robot of claim 1, comprising:

[0032] a control end for remotely sending task instructions and receiving monitoring information;

[0033] a cloud server for information interaction with the robot, including task issuing and state returning;

[0034] a communication link for realizing data connection between the robot and the cloud server through a cellular network;

[0035] a multi-machine scheduling module for supporting simultaneous management of task states and path planning of multiple robots.

[0036] Further, the data uploaded by the robot includes real-time position information, running state, battery capacity and camera video stream.

[0037] Further, the instructions include but are not limited to starting a task, pausing a task, remote control and emergency stop.

[0038] An autonomous navigation and remote control method suitable for the robot system, comprising the following steps:

[0039] S1: constructing an environment map, including LIO trajectory estimation and GNSS / loop closure global optimization;

[0040] S2: performing robot initial positioning and loading a map;

[0041] S3: issuing a patrol task by a remote end or a local end;

[0042] S4: path planning based on map information;

[0043] S5: performing autonomous navigation, perceiving obstacles and dynamically avoiding obstacles;

[0044] S6: uploading state information in real time and receiving remote instructions;

[0045] S7: dynamically adjusting a task or controlling a robot action according to an input of a control end.

[0046] Further, the map construction adopts a map format output supported by Autoware, including a PCD format point cloud map and a Lanelet2 format vector map.

[0047] Further, the remote control is realized through a 4G / 5G cellular network, and an end-to-end delay of the video backhaul is not more than 230 ms.

[0048] The present application has the following advantages:

[0049] (1) The present application fuses information of various sensors such as a three-dimensional laser radar, a global navigation satellite system and an inertial measurement unit, constructs a high-precision map and realizes robust positioning. Compared with a traditional scheme relying on a single sensor, the present application can still maintain stable positioning and path control in a park road, building shielding or feature sparse scene, effectively improving navigation precision and reliability.

[0050] (2) By configuring a three-dimensional laser radar and a camera, the robot can comprehensively perceive static and dynamic obstacles in a three-dimensional space and dynamically adjust a driving track in path planning. The design makes the robot adapt to diversified park road structures, effectively avoids collision risks and improves operation safety.

[0051] (3) The present application designs a remote operation platform, realizes real-time communication between the robot and a cloud server. An operator can remotely issue a task, monitor a robot state and view video information at a control end. The platform also supports unified scheduling and management of multiple robots, is beneficial to large-scale deployment and cluster control, and significantly improves system scalability and management efficiency.

[0052] (4) The robot adopts a modular structure design, a chassis, a power supply, a sensor and a computing unit are highly integrated, the system has good stability and fast on-site deployment. A dual power supply design effectively isolates a high-power load and a computing device, guarantees operation reliability and reduces maintenance complexity.

[0053] (5) The system is built based on an open source software framework, supports function extension and software and hardware upgrade, and can add intelligent identification modules, upgrade network communication mode or access new type of sensing devices according to specific park scenes. The feature provides flexible secondary development capability for users and prolongs the product life cycle.

[0054] Other advantages, objects, and features of the application will be apparent to those skilled in the art from the following specification wherein are set forth, in part, some of the best modes contemplated for carrying out the application. The application is capable of other and different embodiments, and its several details are capable of modifications in various obvious respects, all without departing from the spirit and scope of the application. Accordingly, the drawings are to be regarded as illustrative in nature, and the specific details are to be interpreted not as limiting, but merely as a representative of the general method of operation. BRIEF DESCRIPTION OF DRAWINGS

[0055] In order to make the objects, technical solutions and advantages of the present application clearer, the preferred embodiments of the present application will be described in detail below with reference to the drawings, in which:

[0056] Fig. 1 is a schematic diagram of electrical connection of the robot;

[0057] Fig. 2 is a schematic diagram of network connection of the robot;

[0058] Fig. 3 is a GLIM architecture;

[0059] Fig. 4 is an Autoware architecture diagram. DETAILED DESCRIPTION

[0060] The embodiments of the present application will be described below through specific concrete examples, and those skilled in the art can easily understand other advantages and effects of the present application from the disclosure of the present specification. The present application can also be implemented or applied through other different specific embodiments, and various modifications or changes can be made to the details in the present specification based on different views and applications without departing from the spirit of the present application. It should be noted that the diagrams provided in the following embodiments only illustrate the basic concept of the present application in a schematic manner, and the following embodiments and features in the embodiments can be combined with each other without conflict.

[0061] The drawings are only used for illustrative explanation, and the representation is only a schematic diagram, not a physical diagram, and cannot be understood as a limitation on the present application; in order to better illustrate the embodiments of the present application, some components in the drawings may be omitted, enlarged or reduced, and do not represent the actual size of the product; it can be understood by those skilled in the art that some known structures and their descriptions in the drawings may be omitted.

[0062] The same or similar reference numerals in the drawings of the embodiments of the present application correspond to the same or similar components; in the description of the present application, it is understood that if the orientations or positional relationships indicated by the terms "upper", "lower", "left", "right", "front", "back" and the like are based on the orientations or positional relationships shown in the drawings, they are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the devices or elements referred to must have a particular orientation, be constructed and operated in a particular orientation, therefore the terms describing the positional relationship in the drawings are only used for exemplary illustration, and cannot be understood as a limitation on the present application, for those skilled in the art, the specific meanings of the above terms can be understood according to the specific circumstances.

[0063] Referring to Figs. 1-4 , the present application provides an autonomous navigation and remote operation method of a park patrol robot, which specifically comprises the following steps:

[0064] Step 1: Environment map construction

[0065] Execution subject: computing unit (such as Nvidia OrinNX) carried by robot.

[0066] The core of this step is to use a SLAM algorithm based on factor graph optimization, to construct a globally consistent high-precision map by fusing multi-sensor data. This process mainly includes two stages: laser radar-inertial odometry (LIO) and global optimization.

[0067] Laser radar-inertial odometry (LIO)

[0068] This stage tightly couples three-dimensional laser radar and IMU data to estimate the motion of the robot. By minimizing an objective function that combines laser radar matching cost and IMU pre-integration cost, low-drift odometry information is obtained. The objective function g LIO (X a ) can be expressed as:

[0069] g LIO (X a )=∑∑e PC (P i ,P j ,T i ,T j )+∑e IMU (x i-1 ,x i )+e MG (X a )

[0070] Wherein:

[0071] X ais the set of robot states (pose, velocity, IMU bias) within the window.

[0072] e PC is the laser-radar matching cost factor, representing the point cloud registration error between consecutive laser-radar frames

[0073] P i and P j . This scheme uses the VGICP algorithm to compute the distribution-to-distribution (D2D) distance between point clouds to measure this error.

[0074] e IMU is the IMU pre-integration factor, representing the error between the relative motion of x i-1 and x i predicted from IMU measurements and the actual estimated relative motion. This factor provides a crucial motion constraint when laser-radar features are degenerated (e.g., in open fields or long corridors).

[0075] e MG is the marginalization factor, used to keep old state information in the optimization.

[0076] Global optimization and mapping

[0077] To eliminate the accumulated error caused by long running of the LIO phase, the global optimization phase introduces the absolute positioning information of GNSS and loop closure detection to optimize the entire trajectory and map. This phase is achieved by minimizing a global cost function. The key lies in the introduction of the GNSS factor, which is used to constrain the absolute position of the robot. The GNSS factor e GNSS is defined as the Mahalanobis distance between the robot's estimated pose and the GNSS measurement value:

[0078]

[0079] where:

[0080] T i is the global pose of the robot to be optimized.

[0081] z i is the measured position provided by the GNSS module.

[0082] H(zi) is the conversion function that converts the GNSS measurement from the geographic coordinate system to the map coordinate system.

[0083] S i is the covariance matrix of the GNSS measurement, representing the uncertainty of this measurement.

[0084] Finally, a globally consistent high-precision 3D point cloud map is generated by solving a global optimization problem containing LIO factors, GNSS factors and loop closure factors, and is converted into formats supported by the Autoware framework (such as PCD format point cloud map and Lanelet2 format vector map).

[0085] Step 2: Autonomous navigation patrol

[0086] Execution subject: computing unit carried by robot.

[0087] Initialization and positioning: after the robot is started, the pre-constructed map is loaded. Through the fusion of the initial position of GNSS and the attitude information of IMU, the initial positioning of the robot is realized. During the running process, through matching the real-time collected laser radar point cloud with the high-precision map (such as NDT matching) and fusing the IMU data, continuous and high-precision real-time positioning is realized.

[0088] Task planning: the operator sets the patrol task for the robot through the remote control platform or locally, such as specifying a series of patrol path points.

[0089] Global path planning: the planning module of the robot plans a globally optimal path from the current position to the target point according to the task target and the road topological information in the vector map.

[0090] Local path planning and obstacle avoidance: when driving along the global path, the perception module processes the three-dimensional laser radar data in real time to detect static and dynamic obstacles. The local planner generates a smooth, collision-free and vehicle dynamics constraint-compliant local driving trajectory according to the obstacle information.

[0091] Motion control: the control module converts the target speed and steering angle output by the local planner into specific control instructions (such as speed instructions) for the four wheel hub motors of the differential chassis to drive the robot to drive along the planned trajectory.

[0092] Step 3: remote operation and monitoring

[0093] Execution subject: robot, cloud server, remote control end.

[0094] Data upload: the robot sends its state information (such as position, speed, power) and sensor data (such as camera video stream) to the cloud server in real time through the 4G / 5G communication module.

[0095] Instruction issuing: the remote operator can view the robot's video and status in real time on the control end software. The operator can issue control instructions, such as starting / stopping the task, emergency stop, manual remote control, etc. The instructions are forwarded to the robot via the server.

[0096] Multi-machine management: The server-side software is responsible for managing the robot queue, supporting multiple users operating different robots simultaneously, and realizing centralized scheduling and management of the entire robot patrol system.

[0097] The present application provides a park patrol robot, the structure and the connection relationship of each part are as follows:

[0098] 1. Overall structure:

[0099] Chassis: adopt integrated welded aluminum alloy frame, install four independent drive hub motors below, realize differential steering.

[0100] Upper structure: install support on the chassis through aluminum profile, form a multi-layer structure, used to accommodate and fix various equipment.

[0101] Housing: adopt one-time forming stainless steel shell, divided into three layers, install computing unit, sensor adapter, communication equipment, etc. inside, install laser radar and GNSS antenna on the top of the shell.

[0102] 2. Core components and connection relationship:

[0103] 2.1 Perception system:

[0104] 32-line three-dimensional laser radar (such as Speedten Helios): installed on the top of the robot, connected to the gigabit switch through Ethernet cable. Used for 360-degree environment perception and map construction.

[0105] GNSS / IMU module: GNSS antenna is installed on the top, receiver and IMU module are placed inside the shell, connected to the USB expansion dock through USB interface. Used to provide global position and attitude information.

[0106] Network camera (2): installed on the front and back of the robot, connected to the gigabit switch through Ethernet cable. Used to provide real-time video stream.

[0107] 2.2 Computing and control system:

[0108] Onboard computing unit (such as Nvidia Orin NX): as the "brain" of the robot, installed inside the shell. Connected to the gigabit switch through Ethernet cable, connected to the USB expansion dock through USB cable. Responsible for running operating system (ROS2), Autoware navigation framework, SLAM algorithm and remote communication program.

[0109] Bottom control board: integrated with the chassis, communicates with four motor drivers through CAN bus, connected with the USB expansion dock through RS232-USB adapter. Responsible for receiving motion instructions from the computing unit and executing.

[0110] 2.3 Communication system:

[0111] Gigabit switch: Connects laser radar, camera, computing unit, and constitutes the local area network inside the robot.

[0112] 4G / 5G router: Connected to the gigabit switch, provides the robot with access to the Internet, enabling communication with remote servers.

[0113] 2.4 Power system:

[0114] Dual power supply design: One battery (such as 29V) powers the chassis power system (motor, control board); another battery (such as 25.4V) is reduced to 12V by DC-DC, and powers the upper layer computing and sensing devices (computing unit, sensor, switch, etc.). Dual power supply design can avoid the large current of motor starting interfering with the computing unit, improving system stability.

[0115] For the purpose of the invention (park autonomous patrol), there are some alternative solutions in the technical route, but the scheme chosen by the invention has more advantages in comprehensive performance:

[0116] 1. Navigation framework alternative:

[0117] Alternative: Use the Navigation2 (Nav2) framework provided by ROS2.

[0118] Comparison: Nav2 is mainly designed for indoor or semi-structured environments, and its map model is relatively simple, with limited ability to express complex scenarios such as outdoor roads and traffic rules. Autoware.Universe is a framework designed specifically for autonomous driving, with more comprehensive functions, supporting high-precision vector maps (Lanelet2), and better handling of lane keeping and path planning tasks in park roads, making it more suitable for the application scenario of the invention.

[0119] 2. Perception and positioning scheme alternative:

[0120] Alternative 1: Pure visual SLAM scheme. That is, only using cameras for positioning and mapping.

[0121] Comparison: The pure visual scheme is low in cost, but it is very sensitive to changes in light and weather, and has poor robustness in night or light changes, making it difficult to meet the needs of all-weather patrol. The laser radar-based scheme used in the invention is not affected by light and can work stably all day.

[0122] Alternative 2: 2D laser radar SLAM scheme.

[0123] Comparison: 2D laser radar can only obtain information of a horizontal plane, and cannot perceive obstacles in three-dimensional space (such as overhanging branches and road edges), which is prone to errors in complex outdoor terrain. The application uses a three-dimensional laser radar to obtain rich three-dimensional environmental information, and has stronger obstacle avoidance capability and positioning robustness.

[0124] 3. Communication scheme alternatives:

[0125] Alternative: use the Wi-Fi network in the park for communication.

[0126] Comparison: The park Wi-Fi network may have incomplete coverage, unstable signal, and bandwidth competition among multiple users, which can easily cause remote control interruption. The application uses a 4G / 5G cellular network, which has a wide coverage range and is more stable and reliable in connection, ensuring the real-time and smoothness of remote operation.

[0127] Embodiment one: autonomous map construction implementation process in a park environment

[0128] This embodiment is aimed at the first deployment of a robot in a target park environment, and constructs a high-precision map to support subsequent autonomous navigation tasks.

[0129] The workflow is as follows:

[0130] After starting the robot, the internal computing unit initializes the three-dimensional laser radar, the inertial measurement unit, and the global navigation satellite system module in sequence;

[0131] The robot moves slowly along the main road of the park, while collecting laser point cloud, inertial measurement data, and satellite positioning information;

[0132] Through tight coupling processing of laser and inertial data, the initial motion trajectory is calculated;

[0133] On this basis, satellite positioning information is introduced for global optimization to eliminate cumulative errors;

[0134] For some areas with looped paths, a loop detection algorithm is used to further optimize the consistency of the trajectory;

[0135] A unified format three-dimensional point cloud map and vector road map are generated and stored in the robot's internal database for subsequent navigation module calling.

[0136] Embodiment two: autonomous patrol implementation process based on high-precision map

[0137] This embodiment is based on the constructed park map, and realizes the continuous autonomous patrol task of the robot in the specified area.

[0138] The workflow is as follows:

[0139] The operator issues a patrol task through a remote control platform, including a starting point, a sequence of path points, and an ending point position.

[0140] The robot loads a corresponding map and completes initialization positioning according to current global positioning system information and inertial measurement information;

[0141] Based on the road topology structure in the map, an optimal path from the current position to the target path point is calculated;

[0142] During navigation, three-dimensional laser radar data are acquired in real time and matched with the high-precision map to realize continuous positioning;

[0143] If a static or dynamic obstacle is detected in front, the robot automatically activates a local path planning module to adjust the current driving trajectory to avoid the obstacle;

[0144] The control module converts the navigation path into control instructions and sends them to four wheel hub motors to drive the robot to travel smoothly;

[0145] When the robot completes all path point patrols, it automatically enters a standby state and waits for the next task instruction.

[0146] Example Three: Remote monitoring and multi-robot scheduling implementation process

[0147] This example describes the centralized management and scheduling of multiple patrol robots through a remote operation platform.

[0148] The workflow is as follows:

[0149] Each robot accesses a cloud server through a cellular network and uploads running state, position information, and video data in real time;

[0150] The operator of the remote control terminal can log in to the management system to view the robot status panel and the park map interface;

[0151] The operator can issue task instructions for any one or more robots, such as changing a patrol route, pausing a task, or emergency stopping;

[0152] If an abnormal event occurs, such as an obstacle blocking or navigation failure, the system will automatically alarm and prompt the operator to intervene;

[0153] The multi-robot scheduling module dynamically allocates tasks and paths according to task priorities and current robot positions to avoid repeated coverage;

[0154] All data records will be stored in a background database for subsequent operation and maintenance analysis and optimized scheduling strategy.

[0155] Example Four: Precise positioning and navigation implementation process under heterogeneous perception fusion

[0156] This embodiment shows how the robot can achieve accurate positioning in complex environments through multi-sensor data fusion.

[0157] The workflow is as follows:

[0158] After the robot enters the park, it first collects the current laser radar data and inertial measurement data;

[0159] Match the current frame of laser point cloud with the pre-built point cloud map to obtain a preliminary position estimate;

[0160] At the same time, integrate the inertial measurement unit data to calculate the attitude change in a short period of time;

[0161] Use the laser registration result and the inertial calculation result for joint optimization to output the final high-precision pose;

[0162] In areas with strong GPS signals, introduce global positioning results to further correct the current pose;

[0163] If entering an area with weak GPS signals, the system automatically reduces the weight of GPS information and prioritizes laser and inertial data;

[0164] Through the above multi-source perception fusion mechanism, the robot can still maintain stable positioning and navigation in scenes with building obstructions or severe changes in lighting.

[0165] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and are not limiting. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced by equivalents without departing from the spirit and scope of the present application, and all should be covered in the scope of the claims of the present application.

Claims

1. An autonomous navigation patrol robot based on 3D laser radar and GNSS, characterized by: include: differential drive chassis for robot locomotion; an upper mechanical structure, mounted on the differential drive chassis and used for carrying a computing unit and a plurality of sensors; A perception system, comprising a three-dimensional laser radar, a GNSS receiver module, an inertial measurement unit (IMU), and at least one network camera, for acquiring environmental perception information and position information; A computing unit, connected to the perception system, configured to perform positioning, path planning, and obstacle avoidance control based on the Autoware autonomous driving framework; Communication system, including a gigabit switch and 4G / 5G cellular network communication module, for data transmission between the robot and the remote operation platform; The power supply system adopts an independent dual-battery structure, in which one battery supplies power to the chassis system, and the other battery supplies power to the sensing and computing equipment after voltage reduction.

2. The autonomous navigation patrol robot based on three-dimensional laser radar and GNSS according to claim 1 is characterized in that: The autonomous navigation performed by the computing unit includes: Start the positioning step and obtain the initial pose by fusing GNSS and IMU; Real-time positioning step, matching the real-time lidar point cloud with the pre-built map to achieve continuous positioning; Path planning step, generating a global path based on the mission target point and map topology information; Local path replanning and obstacle avoidance steps, dynamically adjusting the local path based on real-time perception information; The control instruction generation step converts the planning results into chassis motor control signals.

3. The autonomous navigation patrol robot based on three-dimensional laser radar and GNSS according to claim 2, characterized in that: The laser radar point cloud is matched with the pre-built map using the normal distribution transformation matching algorithm NDT or the VGICP algorithm.

4. A method for constructing an environmental map based on the robot according to any one of claims 1 to 3, characterized in that: include: Build a laser inertial odometry (LIO) based on lidar and IMU, and obtain low-drift trajectories by optimizing the objective function; The objective function J is expressed as: in, represents the lidar matching factor; Represents the IMU pre-integration factor; represents the marginalization factor; n is the number of states in the optimization window; The GNSS factor and loop closure factor are introduced for global optimization, where the objective function of the GNSS factor is expressed as: J GNSS =(T-T GNSS ) T Σ -1 (T-T GNSS ) Among them, T is the optimized pose, T GNSS is the measurement position provided by GNSS, Σ is the GNSS measurement covariance matrix; Output globally consistent point cloud maps and vector maps for perception matching and path planning respectively.

5. A remote monitoring and operating system for the robot according to claim 1, characterized in that: include: The control terminal is used to send task instructions remotely and receive monitoring information; Cloud server, which interacts with the robot, including task dispatching and status feedback; A communication link, which realizes data connection between the robot and the cloud server through a cellular network; The multi-machine scheduling module supports simultaneous management of the task status and path planning of multiple robots.

6. The remote monitoring and operating system according to claim 5, characterized in that: The data uploaded by the robot includes real-time location information, operating status, battery power and camera video stream.

7. The remote monitoring and operating system according to claim 5, characterized in that: The instructions include starting a task, pausing a task, remote control and emergency stop.

8. An autonomous navigation and remote control method, applicable to the robot system according to claim 1, characterized in that: The following steps are involved: S1: Build environment map, including LIO trajectory estimation and GNSS / loop closure global optimization; S2: Perform initial positioning of the robot and load the map; S3: The remote or local end issues patrol tasks; S4: Path planning based on map information; S5: performs autonomous navigation, perceives obstacles, and dynamically avoids them; S6: Upload status information in real time and receive remote commands; S7: Dynamically adjust tasks or control robot movements based on control-end input.

9. The autonomous navigation and remote control method according to claim 8, characterized in that: The construction environment map is output in a map format supported by Autoware, including a PCD format point cloud map and a Lanelet2 format vector map.

10. The autonomous navigation and remote control method according to claim 8, characterized in that: The remote control is achieved through the 4G / 5G cellular network, and the end-to-end delay of the video backhaul does not exceed 230ms.

Images