Indoor and outdoor integrated unmanned driving navigation method

Abstract

The present invention proposes an integrated indoor and outdoor unmanned driving navigation method. First: multi-line laser radar point cloud 3D mapping and depth camera point cloud 3D mapping are performed. Then: the 3D maps need to be fused, and the actual position calculated by the multi-line laser radar, the actual position calculated by the depth camera, and the actual position calculated by the odometer are fused to achieve the fusion of the depth camera point cloud 3D map and the multi-line laser radar point cloud 3D map, and the actual position of the unmanned vehicle is calculated according to the proportion in various situations; finally: when the unmanned vehicle robot obtains a detailed map of the surrounding area, it can start navigation; when the unmanned vehicle navigation starts to be used, it is necessary to determine whether there are RTK and GPS signals for high-precision positioning, and calculate the appropriate navigation path based on different judgment results. The present invention effectively solves the problem that 5G networked vehicles cannot locate and navigate in areas with dense buildings, areas with weak satellite navigation signals, and indoor and outdoor intersections.

Description

Technical Field

[0001] The present invention relates to the field of unmanned driving technology, and in particular to an indoor and outdoor integrated unmanned driving navigation method. Background Art

[0002] With the development of mobile network technology, the low latency and high reliability of 5G networks have made remote control of unmanned devices such as drones and driverless cars possible. By utilizing the widely covered mobile network, remote control of unmanned aerial vehicles will have higher reliability, and route design will be more real-time, flexible and simple.

[0003] In order to better meet the development technical requirements of 5G connected vehicles, the indoor and outdoor integrated unmanned driving navigation technology solves the problems of connected vehicles being unable to locate and navigate in densely built areas, areas with weak satellite navigation signals, and indoor and outdoor intersection areas. Starting from the improvement and satisfaction of the unmanned vehicle needs in various fields, a unmanned vehicle robot project based on weak signal scenarios is carried out to create a full-system intelligent positioning driving navigation technology from public scenes to weak signal scenes. The present invention proposes an indoor and outdoor integrated unmanned driving navigation method to improve the above-mentioned problems and further promote the development of unmanned driving navigation technology for 5G connected vehicles. Summary of the Invention

[0004] To solve the above technical problems, the present invention proposes an integrated indoor and outdoor unmanned driving navigation method, which effectively solves the problem that 5G networked vehicles cannot locate and navigate in densely built areas, areas with weak satellite navigation signals, and indoor and outdoor intersection areas, and creates a full-system intelligent positioning driving navigation technology that can accurately navigate in public scenes and weak signal scenes.

[0005] To achieve the above object, the technical solution adopted by the present invention is:

[0006] The method for indoor and outdoor integrated unmanned driving navigation comprises the following steps:

[0007] S1: Multi-line LiDAR point cloud 3D mapping:

[0008] The unmanned vehicle robot is equipped with SLAM, GIS or a map interface provided by a map operator. The SLAM is connected to the multi-line laser radar installed on the unmanned vehicle robot itself to scan the road conditions in unknown environments and realize 3D mapping of the multi-line laser radar point cloud through the SLAM, GIS or the map interface provided by the map operator.

[0009] S2: Depth camera point cloud 3D mapping:

[0010] The autonomous vehicle robot is also equipped with a VSLAM, GIS, or a map interface provided by a map operator. The VSLAM is connected to the depth camera installed on the autonomous vehicle robot itself to scan the road conditions in unknown environments and realize 3D mapping of the depth camera point cloud through the VSLAM, GIS, or the map interface provided by the map operator.

[0011] S3: 3D graph fusion:

[0012] The actual position calculated by the multi-line lidar, the actual position calculated by the depth camera, and the actual position inferred by the odometer are integrated to achieve the fusion of the depth camera point cloud 3D map and the multi-line lidar point cloud 3D map, and the actual position of the autonomous vehicle is calculated based on the proportion of various situations;

[0013] S4: Unmanned vehicle navigation begins:

[0014] When the autonomous vehicle robot obtains a detailed map of the surrounding area and receives a start signal, the autonomous vehicle robot begins to navigate according to the detailed map;

[0015] S4: Determine whether there is RTK or GPS signal for high-precision positioning:

[0016] In the absence of GPS and RTK, that is, in the absence of high-precision fusion positioning signals:

[0017] (1) The unmanned vehicle uses a multi-line laser radar to scan the surrounding area and compare it with the map to determine the vehicle's current location;

[0018] (2) The unmanned vehicle identifies the vehicle’s current location by comparing the depth camera’s point cloud data with the map data;

[0019] (3) The encoder obtains the current wheel speed information data in real time and transmits it to the edge computing terminal device;

[0020] (4) The gyroscope outputs the current angle and speed information data to the edge computing terminal device;

[0021] (5) Integrate the encoder speed data and gyroscope angle data to calculate the vehicle's driving state and deduce the current position of the unmanned vehicle;

[0022] (6) Fusing the actual position calculated by the multi-line laser radar, the actual position calculated by the depth camera, and the actual position calculated by the odometer, the actual position of the unmanned vehicle is calculated based on the proportion of each situation;

[0023] In the case of GPS and RTK, that is, when there is a high-precision fusion positioning signal and a 3D map is established:

[0024] (1) The unmanned vehicle uses a multi-line laser radar to scan the surrounding area and compare it with the map to determine the vehicle's current location;

[0025] (2) The unmanned vehicle identifies the vehicle’s current location by comparing the depth camera’s point cloud data with the map data;

[0026] (3) The encoder obtains the current wheel speed information data in real time and transmits it to the edge computing terminal device;

[0027] (4) The gyroscope outputs the current angle and speed information data to the edge computing terminal device;

[0028] (5) Integrate the encoder speed data and gyroscope angle data to calculate the vehicle's driving state and deduce the current position of the unmanned vehicle;

[0029] (6) Fusing the actual position calculated by the multi-line laser radar, the actual position calculated by the depth camera, and the actual position calculated by the odometer, the actual position of the unmanned vehicle is calculated based on the proportion of each situation;

[0030] (7) Through GPS and RTK high-precision positioning, the map is mapped to the map built by VSLAM through the GIS map system, and the position derived from the above perception calculation is calibrated in real time. The position information at this time is also the most accurate.

[0031] In the case of GPS and RTK, that is, high-precision fusion positioning signals but no 3D map is established:

[0032] (1) In areas where no maps are established, use GPS, RTK, or high-precision fusion positioning to determine your location;

[0033] (2) Identify surrounding obstacle information through multi-line laser radar and depth camera point cloud data;

[0034] (3) Odometer data is integrated with multi-line laser radar and depth camera point cloud data to calculate the relative position of the unmanned vehicle on the map in real time;

[0035] (4) Real-time positioning is performed based on GPS, RTK, or high-precision fusion positioning information, supplemented by odometer, multi-line lidar, and depth camera data.

[0036] In the above structure: The present invention proposes an indoor and outdoor integrated unmanned driving navigation method to solve the navigation and positioning technology of unmanned vehicles that need to move indoors and outdoors at the same time. By combining VSLAM technology and RTK high-precision positioning algorithm, a high-precision navigation and positioning algorithm of visual perception, satellite positioning, base station positioning and multi-sensor fusion is realized.

[0037] In indoor areas, basements, densely populated areas, and other areas where GPS and RTK cannot be used for positioning, the 3D map collected in advance by the SLAM algorithm and VSLAM algorithm by the multi-line lidar and depth camera is used to accurately determine the position of the unmanned vehicle through the fusion of real-time data from multiple sensors such as the multi-line lidar, depth camera, inertial navigation, and encoder odometer information.

[0038] In the case of GPS and RTK high-precision positioning, combined with lidar, depth camera, inertial navigation, encoder odometer information, the positioning position is verified in real time through the GPS and RTK high-precision positioning system to ensure the position perception and real-time positioning of the unmanned vehicle through GPS and RTK high-precision positioning when VSLM 3D map acquisition technology cannot be performed in advance in open areas.

[0039] As a preferred technical solution of the present invention: in step S1: the SLAM includes sensor data, a visual odometry, a backend and mapping, the multi-line lidar is connected to the sensor data for transmitting map data in an unknown environment, the sensor data is connected to the visual odometry, the visual odometry is connected to the backend, and the backend is connected to mapping to realize multi-line lidar 3D point cloud mapping.

[0040] As a preferred technical solution of the present invention: it also includes loop detection, the sensor data is connected to the loop detection, and the loop detection is connected to the back end, which is used to identify whether the map data transmitted by the sensor data is consistent with the map data in the previous database. If the map data is consistent, the data in the database is directly used for optimization and multi-line laser radar 3D point cloud mapping.

[0041] In the above structure: SLAM includes sensor data, visual odometry, backend and mapping. Multi-line lidar is connected to sensor data to transmit map data in unknown environments. Sensor data sends map data to the visual odometry to estimate the relative position of the unmanned robot at different times, including the application of algorithms such as feature matching and direct registration. The backend is mainly used to optimize the cumulative error caused by the visual odometry, including the application of algorithms such as filters and graph optimization. Finally, based on the optimized map data, multi-line lidar 3D point cloud mapping is performed.

[0042] The loop detection connection backend is used to identify whether the map data transmitted by the sensor data is consistent with the map data in the previous database. If the map data is consistent, the data in the database is directly used for optimization and multi-line lidar 3D point cloud mapping.

[0043] As a preferred technical solution of the present invention: in step S2: the VSLAM includes sensor data, a visual odometry, a backend and mapping, the depth camera is connected to the sensor data for transmitting map data in an unknown environment, the sensor data is connected to the visual odometry, the visual odometry is connected to the backend, and the backend is connected to the mapping to realize 3D point cloud mapping by the depth camera.

[0044] As a preferred technical solution of the present invention: it also includes loop detection, the sensor data is connected to the loop detection, and the loop detection is connected to the back end, which is used to identify whether the map data transmitted by the sensor data is consistent with the map data in the previous database. If the map data is consistent, the data in the database is directly used for optimization and depth camera 3D point cloud mapping.

[0045] In the above structure: VSLAM includes sensor data, visual odometry, backend and mapping. The depth camera is connected to the sensor data to transmit map data in unknown environments. The sensor data sends the map data to the visual odometry for estimating the relative position of the unmanned robot at different times, including the application of algorithms such as feature matching and direct registration. The backend is mainly used to optimize the cumulative error caused by the visual odometry, including the application of algorithms such as filters and graph optimization. Finally, based on the optimized map data, the depth camera 3D point cloud mapping is performed.

[0046] The loop detection connection backend is used to identify whether the map data transmitted by the sensor data is consistent with the map data in the previous database. If the map data is consistent, the data in the database is directly used for optimization and depth camera 3D point cloud mapping.

[0047] As a preferred technical solution of the present invention: the unmanned vehicle robot is provided with a central control system and a chassis drive, the central control system is provided with a serial port, a network port and a parallel port, and the central control system is connected to the chassis drive through the serial port, the network port and the parallel port.

[0048] In the above structure: the central control system is used to process data in the unmanned vehicle robot. The central control system is connected to the chassis drive through the serial port, network port and parallel port. When the unmanned vehicle robot obtains a detailed map of the surrounding area, the central control system sends a control signal to the chassis drive after receiving the start signal, and the unmanned vehicle robot starts to move according to the planned path.

[0049] As an optimal technical solution of the present invention: the chassis drive is provided with a single-chip microcomputer and a chassis motor drive, the single-chip microcomputer is connected to the chassis motor drive, the single-chip microcomputer is provided with a serial port, a network port and a parallel port, the single-chip microcomputer is connected to the hollow system through the serial port, the network port and the parallel port, the central control system sends a control signal to the single-chip microcomputer, and the single-chip microcomputer sends a control command to the chassis motor drive.

[0050] In the above structure: the chassis drive is provided with a single-chip microcomputer and a chassis motor drive. After receiving the start signal, the central control system sends a control signal to the single-chip microcomputer in the chassis drive. After receiving the control signal, the single-chip microcomputer sends a control command to the chassis motor drive. The chassis motor drive starts to run, and the unmanned robot starts to move forward according to the planned path.

[0051] As a preferred technical solution of the present invention: the central control system is also provided with a customized interface, and the central control system is connected to the corresponding modules on the unmanned vehicle robot through the customized interface.

[0052] In the above structure: a customized interface is also provided on the central control system for connecting to the corresponding module on the unmanned vehicle robot to realize the switching of different functions.

[0053] As the preferred technical solution of the present invention: the central control system is connected to a multi-line laser radar and a depth camera, and the multi-line laser radar and the depth camera realize mapping, positioning and path planning through SLAM, VSLAM, 5G, and UWB

[0054] As a preferred technical solution of the present invention: the central control system is connected to the cloud platform through a wireless network module, and the wireless network module is 4G, 5G or WIFI.

[0055] Compared with the prior art, the present invention has the following beneficial effects:

[0056] The present invention combines VSLAM technology with RTK high-precision positioning algorithms to realize visual perception, satellite positioning, base station positioning, and multi-sensor fusion high-precision navigation and positioning algorithms, enabling unmanned vehicle robots to accurately locate and navigate in densely built areas, areas with weak satellite navigation signals, and indoor and outdoor intersection areas, improving and meeting the needs of unmanned vehicles in various fields, and creating a full-system intelligent positioning driving and navigation technology that integrates everything from public scenes to weak signal scenes.

[0057] In indoor areas, basements, densely populated areas, and other areas where GPS and RTK cannot be used for positioning, the 3D map collected in advance by the VSLAM algorithm using multi-line lidar and depth cameras is used to accurately determine the position of the unmanned vehicle through the fusion of real-time data from multiple sensors such as multi-line lidar, depth cameras, inertial navigation, and encoder odometer information.

[0058] In the case of GPS and RTK high-precision positioning, combined with multi-line lidar, depth camera, inertial navigation, encoder odometer information, the positioning position is verified in real time through the GPS and RTK high-precision positioning system to ensure the position perception and real-time positioning of the unmanned vehicle through GPS and RTK high-precision positioning in open areas where VSLM 3D map acquisition technology cannot be performed in advance. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 It is the overall flow chart of the present invention;

[0060] Figure 2 It is a schematic diagram of the chassis interface in the present invention;

[0061] Figure 3 This is a schematic diagram of the central control system interface in the present invention;

[0062] Figure 4 This is a schematic diagram showing the principle of connecting the multi-line laser radar and SLAM in the present invention;

[0063] Figure 5 It is a schematic diagram of the principle structure of the connection between the depth camera and VSLAM in the present invention. DETAILED DESCRIPTION

[0064] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments:

[0065] like Figure 1-5 As shown: The present invention proposes an indoor and outdoor integrated unmanned driving navigation method, comprising the following steps:

[0066] S1: Multi-line LiDAR point cloud 3D mapping:

[0067] The unmanned vehicle robot is equipped with SLAM, GIS or a map interface provided by a map operator. The SLAM is connected to the multi-line laser radar installed on the unmanned vehicle robot itself to scan the road conditions in unknown environments and realize 3D mapping of the multi-line laser radar point cloud through the SLAM, GIS or the map interface provided by the map operator.

[0068] S2: Depth camera point cloud 3D mapping:

[0069] The autonomous vehicle robot is also equipped with a VSLAM, GIS, or a map interface provided by a map operator. The VSLAM is connected to the depth camera installed on the autonomous vehicle robot itself to scan the road conditions in unknown environments and realize 3D mapping of the depth camera point cloud through the VSLAM, GIS, or the map interface provided by the map operator.

[0070] S3: 3D graph fusion:

[0071] The actual position calculated by the multi-line lidar, the actual position calculated by the depth camera, and the actual position inferred by the odometer are integrated to achieve the fusion of the depth camera point cloud 3D map and the multi-line lidar point cloud 3D map, and the actual position of the autonomous vehicle is calculated based on the proportion of various situations;

[0072] S4: Unmanned vehicle navigation begins:

[0073] When the autonomous vehicle robot obtains a detailed map of the surrounding area and receives a start signal, the autonomous vehicle robot begins to navigate according to the detailed map;

[0074] S4: Determine whether there is RTK or GPS signal for high-precision positioning:

[0075] In the absence of GPS and RTK, that is, in the absence of high-precision fusion positioning signals:

[0076] (1) The unmanned vehicle uses a multi-line laser radar to scan the surrounding area and compare it with the map to determine the vehicle's current location;

[0077] (2) The unmanned vehicle identifies the vehicle’s current location by comparing the depth camera’s point cloud data with the map data;

[0078] (3) The encoder obtains the current wheel speed information data in real time and transmits it to the edge computing terminal device;

[0079] (4) The gyroscope outputs the current angle and speed information data to the edge computing terminal device;

[0080] (5) Integrate the encoder speed data and gyroscope angle data to calculate the vehicle's driving state and deduce the current position of the unmanned vehicle;

[0081] (6) Fusing the actual position calculated by the multi-line laser radar, the actual position calculated by the depth camera, and the actual position calculated by the odometer, the actual position of the unmanned vehicle is calculated based on the proportion of each situation;

[0082] In the case of GPS and RTK, that is, when there is a high-precision fusion positioning signal and a 3D map is established:

[0083] (1) The unmanned vehicle uses a multi-line laser radar to scan the surrounding area and compare it with the map to determine the vehicle's current location;

[0084] (2) The unmanned vehicle identifies the vehicle’s current location by comparing the depth camera’s point cloud data with the map data;

[0085] (3) The encoder obtains the current wheel speed information data in real time and transmits it to the edge computing terminal device;

[0086] (4) The gyroscope outputs the current angle and speed information data to the edge computing terminal device;

[0087] (5) Integrate the encoder speed data and gyroscope angle data to calculate the vehicle's driving state and deduce the current position of the unmanned vehicle;

[0088] (6) Fusing the actual position calculated by the multi-line laser radar, the actual position calculated by the depth camera, and the actual position calculated by the odometer, the actual position of the unmanned vehicle is calculated based on the proportion of each situation;

[0089] (7) Through GPS and RTK high-precision positioning, the map is mapped to the map built by VSLAM through the GIS map system, and the position derived from the above perception calculation is calibrated in real time. The position information at this time is also the most accurate.

[0090] In the case of GPS and RTK, that is, high-precision fusion positioning signals but no 3D map is established:

[0091] (1) In areas where no maps are established, use GPS, RTK, or high-precision fusion positioning to determine your location;

[0092] (2) Identify surrounding obstacle information through multi-line laser radar and depth camera point cloud data;

[0093] (3) Odometer data is integrated with multi-line laser radar and depth camera point cloud data to calculate the relative position of the unmanned vehicle on the map in real time;

[0094] (4) Real-time positioning is performed based on GPS, RTK, or high-precision fusion positioning information, supplemented by odometer, multi-line lidar, and depth camera data.

[0095] The present invention proposes an indoor and outdoor integrated unmanned driving navigation method to solve the navigation and positioning technology of unmanned vehicles that need to operate indoors and outdoors at the same time. By combining VSLAM technology and RTK high-precision positioning algorithm, a high-precision navigation and positioning algorithm combining visual perception, satellite positioning, base station positioning, and multi-sensor fusion is realized.

[0096] First, perform 3D mapping of multi-line lidar point clouds and depth camera point clouds. SLAM connects to the multi-line lidar installed on the unmanned vehicle to scan road conditions in unknown environments. This is achieved through the map interface provided by SLAM, GIS, or map operators. VSLAM connects to the depth camera installed on the unmanned vehicle to scan road conditions in unknown environments. This is achieved through the map interface provided by VSLAM, GIS, or map operators. This is achieved through the map interface provided by VSLAM, GIS, or map operators.

[0097] Then, the 3D images need to be fused. The actual position calculated by the multi-line LiDAR, the actual position calculated by the depth camera, and the actual position inferred by the odometer are integrated to achieve the fusion of the depth camera point cloud 3D image and the multi-line LiDAR point cloud 3D image. The actual position of the autonomous vehicle is calculated based on the proportion of each situation.

[0098] Finally: the unmanned vehicle navigation can begin. When the unmanned vehicle robot obtains a detailed map of the surrounding area and receives a start signal, the unmanned vehicle robot begins to navigate according to the detailed map;

[0099] When unmanned vehicle navigation is first used, it is necessary to determine whether there are RTK and GPS signals for high-precision positioning.

[0100] In indoor areas, basements, densely populated areas, and other areas where GPS and RTK cannot be used for positioning, the 3D map collected in advance by the SLAM algorithm and VSLAM algorithm by the multi-line lidar and depth camera is used to accurately determine the position of the unmanned vehicle through the fusion of real-time data from multiple sensors such as the multi-line lidar, depth camera, inertial navigation, and encoder odometer information.

[0101] In the case of GPS and RTK high-precision positioning, combined with lidar, depth camera, inertial navigation, encoder odometer information, the positioning position is verified in real time through the GPS and RTK high-precision positioning system to ensure the position perception and real-time positioning of the unmanned vehicle through GPS and RTK high-precision positioning when VSLM 3D map acquisition technology cannot be performed in advance in open areas.

[0102] This paper primarily utilizes multi-line laser radar mapping and depth camera VSLAM mapping. SLAM is primarily used to solve the positioning, navigation, and map-building problems faced by autonomous vehicles operating in unknown environments. SLAM typically involves the following components: feature extraction, data association, state estimation, state update, and feature update. Multiple methods exist for each of these components.

[0103] VSLAM is a mapping and navigation technology that uses a depth camera to capture images of the surrounding environment, filter and calculate these images, determine its own position, identify its path, and make navigation decisions. Visual navigation operates passively, requires simple equipment, is low-cost, and has a wide range of applications. Its key features are autonomy and real-time performance. It does not rely on any external equipment and simply calculates information stored in the system and the environment to derive navigation information.

[0104] This method generates a map for practical application scenarios by fusing and complementing multi-line LiDAR point cloud maps with depth camera point cloud maps. Obstacles are identified in scanned low-lying areas and height-restricted areas, allowing the optimal route to be planned for the autonomous vehicle.

[0105] In this embodiment: In step S1: the SLAM includes sensor data, a visual odometry, a backend and mapping, the multi-line lidar is connected to the sensor data for transmitting map data in an unknown environment, the sensor data is connected to the visual odometry, the visual odometry is connected to the backend, and the backend is connected to mapping to realize multi-line lidar 3D point cloud mapping.

[0106] It also includes loop detection. The sensor data is connected to the loop detection, and the loop detection is connected to the back end to identify whether the map data transmitted by the sensor data is consistent with the map data in the previous database. If the map data is consistent, the data in the database is directly used for optimization and multi-line laser radar 3D point cloud mapping.

[0107] SLAM includes sensor data, visual odometry, backend, and mapping. Multi-line LiDAR is connected to sensor data to transmit map data in unknown environments. Sensor data sends map data to the visual odometry to estimate the relative position of the unmanned robot at different times. This includes the application of algorithms such as feature matching and direct registration. The backend is mainly used to optimize the cumulative error caused by the visual odometry, including the application of algorithms such as filters and graph optimization. Finally, based on the optimized map data, multi-line LiDAR 3D point cloud mapping is performed.

[0108] The loop detection connection backend is used to identify whether the map data transmitted by the sensor data is consistent with the map data in the previous database. If the map data is consistent, the data in the database is directly used for optimization and multi-line lidar 3D point cloud mapping.

[0109] In this embodiment: In step S2: the VSLAM includes sensor data, a visual odometry, a backend and mapping, the depth camera is connected to the sensor data for transmitting map data in an unknown environment, the sensor data is connected to the visual odometry, the visual odometry is connected to the backend, and the backend is connected to mapping to realize 3D point cloud mapping by the depth camera.

[0110] It also includes loop detection. The sensor data is connected to the loop detection, and the loop detection is connected to the back end to identify whether the map data transmitted by the sensor data is consistent with the map data in the previous database. If the map data is consistent, the data in the database is directly used for optimization and depth camera 3D point cloud mapping.

[0111] VSLAM includes sensor data, visual odometry, backend, and mapping. The depth camera connects to the sensor data to transmit map data in unknown environments. The sensor data sends the map data to the visual odometry to estimate the relative position of the unmanned robot at different times, including the application of algorithms such as feature matching and direct registration. The backend is mainly used to optimize the cumulative error caused by the visual odometry, including the application of algorithms such as filters and graph optimization. Finally, the depth camera 3D point cloud mapping is performed based on the optimized map data.

[0112] The loop detection connection backend is used to identify whether the map data transmitted by the sensor data is consistent with the map data in the previous database. If the map data is consistent, the data in the database is directly used for optimization and depth camera 3D point cloud mapping.

[0113] In this embodiment: the unmanned vehicle robot is provided with a central control system and a chassis drive, the central control system is provided with a serial port, a network port and a parallel port, and the central control system is connected to the chassis drive through the serial port, the network port and the parallel port.

[0114] The central control system is used to process data in the unmanned vehicle robot. The central control system is connected to the chassis drive through the serial port, network port and parallel port. When the unmanned vehicle robot obtains a detailed map of the surrounding area, the central control system sends a control signal to the chassis drive after receiving the start signal, and the unmanned vehicle robot begins to move according to the planned path.

[0115] The chassis drive is provided with a single-chip microcomputer and a chassis motor drive, the single-chip microcomputer is connected to the chassis motor drive, the single-chip microcomputer is provided with a serial port, a network port and a parallel port, the single-chip microcomputer is connected to the hollow system through the serial port, the network port and the parallel port, the central control system sends a control signal to the single-chip microcomputer, and the single-chip microcomputer sends a control command to the chassis motor drive.

[0116] The chassis drive is equipped with a single-chip microcomputer and a chassis motor drive. After receiving the start signal, the central control system sends a control signal to the single-chip microcomputer in the chassis drive. After receiving the control signal, the single-chip microcomputer sends a control command to the chassis motor drive. The chassis motor drive starts running, and the unmanned robot starts to move according to the planned path.

[0117] In this embodiment: the central control system is also provided with a customized interface, and the central control system is connected to the corresponding modules on the unmanned vehicle robot through the customized interface.

[0118] The central control system is also equipped with a customized interface for connecting to the corresponding modules on the unmanned vehicle robot to realize the switching of different functions.

[0119] In this embodiment: the central control system is connected to a multi-line laser radar and a depth camera, which realizes mapping, positioning and path planning through SLAM, VSLAM, 5G and UWB.

[0120] In this embodiment: the central control system is connected to the cloud platform via a wireless network module, and the wireless network module is 4G, 5G or WIFI.

[0121] The present invention combines VSLAM technology with RTK high-precision positioning algorithms to realize visual perception, satellite positioning, base station positioning, and multi-sensor fusion high-precision navigation and positioning algorithms, enabling unmanned vehicle robots to accurately locate and navigate in densely built areas, areas with weak satellite navigation signals, and indoor and outdoor intersection areas, improving and meeting the needs of unmanned vehicles in various fields, and creating a full-system intelligent positioning driving and navigation technology that integrates everything from public scenes to weak signal scenes.

[0122] In indoor areas, basements, densely populated areas, and other areas where GPS and RTK cannot be used for positioning, the 3D map collected in advance by the VSLAM algorithm using multi-line lidar and depth cameras is used to accurately determine the position of the unmanned vehicle through the fusion of real-time data from multiple sensors such as multi-line lidar, depth cameras, inertial navigation, and encoder odometer information.

[0123] In the case of GPS and RTK high-precision positioning, combined with multi-line lidar, depth camera, inertial navigation, encoder odometer information, the positioning position is verified in real time through the GPS and RTK high-precision positioning system to ensure the position perception and real-time positioning of the unmanned vehicle through GPS and RTK high-precision positioning in open areas where VSLM 3D map acquisition technology cannot be performed in advance.

[0124] The above description is merely a preferred embodiment of the present invention and does not constitute any other form of limitation to the present invention. Any modification or equivalent variation based on the technical essence of the present invention shall still fall within the scope of protection claimed by the present invention.

Claims

1. An integrated indoor and outdoor unmanned driving navigation method, characterized by: The steps include: S1: Multi-line LiDAR point cloud 3D mapping: The unmanned vehicle robot is equipped with SLAM, GIS or a map interface provided by a map operator. The SLAM is connected to the multi-line laser radar installed on the unmanned vehicle robot itself to scan the road conditions in unknown environments and realize 3D mapping of the multi-line laser radar point cloud through the SLAM, GIS or the map interface provided by the map operator. S2: Depth camera point cloud 3D mapping: The autonomous vehicle robot is also equipped with a VSLAM, GIS, or a map interface provided by a map operator. The VSLAM is connected to the depth camera installed on the autonomous vehicle robot itself to scan the road conditions in unknown environments and realize 3D mapping of the depth camera point cloud through the VSLAM, GIS, or the map interface provided by the map operator. S3: 3D graph fusion: The actual position calculated by the multi-line lidar, the actual position calculated by the depth camera, and the actual position inferred by the odometer are integrated to achieve the fusion of the depth camera point cloud 3D map and the multi-line lidar point cloud 3D map, and the actual position of the autonomous vehicle is calculated based on the proportion of various situations; S4: Unmanned vehicle navigation begins: When the autonomous vehicle robot obtains a detailed map of the surrounding area and receives a start signal, the autonomous vehicle robot begins to navigate according to the detailed map; S4: Determine whether there is RTK or GPS signal for high-precision positioning: In the absence of GPS and RTK, that is, in the absence of high-precision fusion positioning signals: (1) The unmanned vehicle uses a multi-line laser radar to scan the surrounding area and compare it with the map to determine the vehicle's current location; (2) The unmanned vehicle identifies the vehicle’s current location by comparing the depth camera’s point cloud data with the map data; (3) The encoder obtains the current wheel speed information data in real time and transmits it to the edge computing terminal device; (4) The gyroscope outputs the current angle and speed information data to the edge computing terminal device; (5) Integrate the encoder speed data and gyroscope angle data to calculate the vehicle's driving state and deduce the current position of the unmanned vehicle; (6) Fusing the actual position calculated by the multi-line laser radar, the actual position calculated by the depth camera, and the actual position calculated by the odometer, the actual position of the unmanned vehicle is calculated based on the proportion of each situation; In the case of GPS and RTK, that is, when there is a high-precision fusion positioning signal and a 3D map is established: (1) The unmanned vehicle uses a multi-line laser radar to scan the surrounding area and compare it with the map to determine the vehicle's current location; (2) The unmanned vehicle identifies the vehicle’s current location by comparing the depth camera’s point cloud data with the map data; (3) The encoder obtains the current wheel speed information data in real time and transmits it to the edge computing terminal device; (4) The gyroscope outputs the current angle and speed information data to the edge computing terminal device; (5) Integrate the encoder speed data and gyroscope angle data to calculate the vehicle's driving state and deduce the current position of the unmanned vehicle; (6) Fusing the actual position calculated by the multi-line laser radar, the actual position calculated by the depth camera, and the actual position calculated by the odometer, the actual position of the unmanned vehicle is calculated based on the proportion of each situation; (7) Through GPS and RTK high-precision positioning, the map is mapped to the map built by VSLAM through the GIS map system, and the position derived from the above perception calculation is calibrated in real time. The position information at this time is also the most accurate. In the case of GPS and RTK, that is, high-precision fusion positioning signals but no 3D map is established: (1) In areas where no maps are established, use GPS, RTK, or high-precision fusion positioning to determine your location; (2) Identify surrounding obstacle information through multi-line laser radar and depth camera point cloud data; (3) Odometer data is integrated with multi-line laser radar and depth camera point cloud data to calculate the relative position of the unmanned vehicle on the map in real time; (4) Real-time positioning is performed based on GPS, RTK, or high-precision fusion positioning information, supplemented by odometer, multi-line lidar, and depth camera data.

2. The indoor and outdoor integrated unmanned driving navigation method according to claim 1, characterized in that: In step S1: the SLAM includes sensor data, visual odometry, backend and mapping. The multi-line lidar is connected to the sensor data to transmit map data in an unknown environment. The sensor data is connected to the visual odometry, the visual odometry is connected to the backend, and the backend is connected to the mapping to realize multi-line lidar 3D point cloud mapping.

3. The indoor and outdoor integrated unmanned driving navigation method according to claim 2, characterized in that: It also includes loop detection. The sensor data is connected to the loop detection, and the loop detection is connected to the back end to identify whether the map data transmitted by the sensor data is consistent with the map data in the previous database. If the map data is consistent, the data in the database is directly used for optimization and multi-line laser radar 3D point cloud mapping.

4. The indoor and outdoor integrated unmanned driving navigation method according to claim 1, characterized in that: In step S2: the VSLAM includes sensor data, visual odometry, backend and mapping. The depth camera is connected to the sensor data to transmit map data in an unknown environment. The sensor data is connected to the visual odometry, the visual odometry is connected to the backend, and the backend is connected to the mapping to realize 3D point cloud mapping by the depth camera.

5. The indoor and outdoor integrated unmanned driving navigation method according to claim 4, characterized in that: It also includes loop detection. The sensor data is connected to the loop detection, and the loop detection is connected to the back end to identify whether the map data transmitted by the sensor data is consistent with the map data in the previous database. If the map data is consistent, the data in the database is directly used for optimization and depth camera 3D point cloud mapping.

6. The indoor and outdoor integrated unmanned driving navigation method according to claim 1, characterized in that: The unmanned vehicle robot is provided with a central control system and a chassis drive. The central control system is provided with a serial port, a network port and a parallel port. The central control system is connected to the chassis drive via the serial port, the network port and the parallel port.

7. The indoor and outdoor integrated unmanned driving navigation method according to claim 6, characterized in that: The chassis drive is provided with a single-chip microcomputer and a chassis motor drive, the single-chip microcomputer is connected to the chassis motor drive, the single-chip microcomputer is provided with a serial port, a network port and a parallel port, the single-chip microcomputer is connected to the hollow system through the serial port, the network port and the parallel port, the central control system sends a control signal to the single-chip microcomputer, and the single-chip microcomputer sends a control command to the chassis motor drive.

8. The indoor and outdoor integrated unmanned driving navigation method according to claim 6, characterized in that: The central control system is also provided with a customized interface, and the central control system is connected to the corresponding modules on the unmanned vehicle robot through the customized interface.

9. The indoor and outdoor integrated unmanned driving navigation method according to claim 6, characterized in that: The central control system is connected to a multi-line laser radar and a depth camera, which realize mapping, positioning and path planning through SLAM, VSLAM, 5G and UWB.

10. The indoor and outdoor integrated unmanned driving navigation method according to claim 6, characterized in that: The central control system is connected to the cloud platform via a wireless network module, and the wireless network module is 4G, 5G or WIFI.

Images