Port air-ground cooperative intelligent autonomous inspection robot system and method

Abstract

The invention discloses a full-automatic intelligent port inspection robot system and method. The system mainly comprises an unmanned vehicle and an unmanned aerial vehicle which are autonomously coordinated. The unmanned vehicle platform comprises a vehicle body, an omni-directional driving wheel set arranged at the bottom of the vehicle body, and a multi-source sensing system, a 5G / Beidou dual-mode communication module, an emergency equipment cabin, an edge computing controller and an intelligent charging system which are arranged on the vehicle body, and real-time data interaction is achieved between the communication module and the unmanned aerial vehicle as well as between the communication module and the port central dispatching system. According to the invention, it is ensured that the robot system can complete core inspection tasks such as container three-dimensional scanning, container body state monitoring and storage yard path planning in an all-weather manner, automatic docking of the unmanned aerial vehicle and the unmanned vehicle with emergency equipment is realized, and a'sensing-decision-execution 'closed-loop system is formed; the functions of high-precision visual identification, multi-mode environment perception, dynamic obstacle avoidance planning, convenient replacement of unmanned aerial vehicle equipment, intelligent charging and the like are integrated, the port inspection efficiency can be remarkably improved, and the new-generation intelligent port equipment of the emergency response process can be optimized.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of robots, and particularly relates to a port open space cooperative inspection robot. BACKGROUND

[0002] In recent years, with the continuous growth of global trade volume, the pressure faced by port operations is also increasing. The problems of low efficiency of yard maintenance, frequent equipment failures and the like are increasingly prominent, which are largely due to the limitations of the traditional manual operation mode. The current port inspection mainly relies on manual work, which not only has many blind areas of inspection, inaccurate data recording and the like, but also is low in efficiency in the link of equipment maintenance. A modern large port usually occupies an area of several square kilometers, contains tens of thousands of container stacking positions, dozens of shore cranes and yard cranes, and the traditional operation mode has been difficult to meet the demand of efficient operation.

[0003] Under the wave of intelligent port construction, intelligent operation systems have become the development trend of the industry. Some advanced ports have begun to try to use robot systems, which not only can realize 7x24 hours uninterrupted operation, but also can improve the inspection accuracy and efficiency. However, the existing solutions are mostly single-function devices, which are difficult to meet the dual needs of inspection and emergency maintenance.

[0004] The robot systems currently applied in ports mainly have the following technical limitations:

[0005] 1. Insufficient multi-function integration:

[0006] The existing systems often separate the functions of inspection and emergency maintenance, resulting in low utilization rate of the devices. The ground robots mostly only have simple inspection functions and cannot complete emergency maintenance and the like.

[0007] 2. Lack of cooperative operation ability:

[0008] Single ground robots are difficult to cover the detection needs of high-level containers in the container yard, and drones cannot perform physical interaction. The two lack effective cooperative mechanisms and cannot form a complementary operation system.

[0009] 3. Poor adaptability to dynamic environment:

[0010] The port operation environment has high dynamicity, and the existing systems are difficult to respond to changes such as container shifting and device moving in real time, and the logistics path planning often deviates.

[0011] 4. Insufficient application of data fusion:

[0012] The inspection data and the logistics management system are isolated from each other, which cannot provide intelligent decision support for port operation and is difficult to realize real digital operation.

[0013] After searching, application publication number CN111300372B, a kind of air-ground coordinated intelligent inspection robot, including robot platform, unmanned aerial vehicle, robot platform includes car body, wheel and drive component, mechanical arm, environmental perception component, communication device, robot controller and power supply component, communication device realizes communication connection to unmanned aerial vehicle and base station.Patrol method includes:air-ground coordinated multi-robot positioning and mapping:including perception positioning calculation, map creation, multi-information fusion positioning, air-ground coordinated tracking and control:including unmanned aerial vehicle flight control design, robot platform trajectory tracking control, unmanned aerial vehicle self-help landing control.

[0014] The air-ground coordinated intelligent inspection robot of patent CN111300372B is mainly for the chemical industry, although it realizes the basic cooperation of unmanned aerial vehicle and unmanned vehicle, but there are deficiencies in actual application.1.The function design is relatively single, only focuses on gas leakage detection and equipment state inspection, lacks emergency maintenance expansion capability, and is difficult to meet the multi-task demand in complex scene.2.The coordination mechanism is limited, the unmanned aerial vehicle cannot replace the equipment independently, the high-altitude and ground operation are separated, and the efficiency is low.3.The wireless charging rate is slow and the power consumption is large, which greatly affects the operation efficiency.

[0015] The port air-ground coordinated inspection robot of the application solves the defects of CN111300372B through multiple technologies.1.It adopts modular design, integrates emergency equipment cabin and intelligent charging system, supports unmanned aerial vehicle to quickly replace mechanical arm, fireball and other equipment, realizes the integration of inspection and maintenance functions.2.For dynamic environment, the path planning algorithm of large model (such as DeepSeek) of the application improves the navigation accuracy and obstacle avoidance ability in complex conditions, combined with YOLO abnormal detection model, realizes the whole process automation from data acquisition to intelligent decision.3.For charging problem, the application optimizes charging success rate and efficiency through contact and structure design. SUMMARY

[0016] The application aims to solve the problems of the above prior art. A port air-ground coordinated intelligent autonomous inspection robot system and method are proposed. The technical scheme of the application is as follows:

[0017] A port open space cooperative intelligent autonomous inspection robot system, comprising: a ground unmanned vehicle platform, which is equipped with a 5G / Beidou dual-mode communication module for realizing real-time data interaction with a unmanned aerial vehicle; the ground unmanned vehicle platform is also provided with an edge computing controller for processing multi-source perception data, realizing environment three-dimensional modeling and obstacle identification; an intelligent charging system, which comprises a circular groove base and a charging contact array, for realizing automatic and accurate landing and charging of the unmanned aerial vehicle; an emergency equipment cabin, which is modularly designed and is equipped with a standard circular docking disc and a lifting platform, for realizing quick equipment replacement of the unmanned aerial vehicle; a unmanned aerial vehicle, which is integrated with a multi-modal environment perception system, including a three-dimensional laser radar, a depth camera, a GPS device, an inertial measurement unit (IMU), for obtaining environment information.

[0018] Further, it also includes an environment perception component, which includes a laser radar and a depth camera 3. The laser radar emits laser pulses and receives reflected signals to construct a precise three-dimensional point cloud model of the surrounding environment in real time, providing high-reliability spatial information for the robot; the depth camera obtains depth information of the surrounding environment in real time through infrared projection and sensor reception, and constructs a precise three-dimensional scene model.

[0019] Further, the unmanned vehicle platform includes a platform body for landing and charging of the unmanned aerial vehicle, which provides a carrying and landing platform for the unmanned aerial vehicle. The platform body is a circular groove, and the circular groove base is arranged on the top of the vehicle body. The circular groove adopts a concentric circular stepped structure design, including a conical guide ring wall. The center of the circular groove base is integrated with a charging contact array, the conical guide ring wall is inclined at an angle of 40°-50° with the horizontal plane and the inner surface is provided with a wear-resistant layer, and the charging contact adopts an elastic copper alloy contact design.

[0020] Further, the unmanned aerial vehicle is configured with a distributed power system, including a brushless motor, a matched electronic speed regulator and a propeller arranged at four corners of the body. The brushless motor is used to provide the core driving force for the flight of the unmanned aerial vehicle, the electronic speed regulator is used to accurately adjust the speed of the brushless motor, and the propeller is used to convert the mechanical energy of the motor rotation into lift. The power system adopts a modular design, and each component has waterproof and dustproof performance. The brushless motor and the electronic speed regulator shell are made of high-strength aluminum alloy material,

[0021] Further, the unmanned vehicle adopts an Ackerman steering structure power system, including a front axle steering mechanism, a rear axle driving motor and a distributed electric control unit, the power system and the vehicle control system realize data interaction through a CAN bus, providing power control and data feedback; the front axle steering mechanism is driven by a servo motor, integrating a steering angle sensor and a torque feedback device to realize precise steering control; the rear axle driving motor is equipped with an electronic differential function, and the left and right wheel torques can be independently adjusted to improve the stability of the vehicle on curved roads and uneven roads.

[0022] Further, the multi-source perception system realizes environment three-dimensional modeling and obstacle identification function by communicating with the edge computing controller, and the data of the three-dimensional laser radar and the depth camera are fused and processed by the edge computing controller for path planning and obstacle avoidance decision.

[0023] Further, the air-ground cooperative system is composed of an unmanned vehicle and an unmanned aerial vehicle, and the overall structure is divided into three layers; the upper layer: the top of the unmanned vehicle is a landing pad for the unmanned aerial vehicle, equipped with openable cabin doors; when the unmanned aerial vehicle lands, the cabin doors automatically open to provide a precise landing platform; when the unmanned aerial vehicle is stored, the cabin doors are closed to protect the unmanned aerial vehicle from the external environment; the second layer: an emergency equipment warehouse for storing emergency equipment and docking with the unmanned aerial vehicle; the lower layer: the power chassis of the unmanned vehicle, integrating the Ackerman steering structure power system and the environment perception system, supporting ground navigation and cargo transportation.

[0024] Further, the emergency equipment warehouse contains symmetrically arranged fixed columns and lifting platforms for carrying two replaceable emergency equipment: a general mechanical arm for aerial equipment maintenance of the unmanned aerial vehicle and a general fire extinguishing ball for aerial fire extinguishing of the unmanned aerial vehicle; the emergency equipment is integrated with a standard circular docking disc, which is fixed by the elastic clamping mechanism of the fixed column in the non-working state and lifted to the docking position by the lifting platform when working; the standard interface of the disc realizes automatic docking of the equipment with the unmanned aerial vehicle, and the equipment can also be returned to the emergency equipment warehouse for reinstallation.

[0025] A method for inspection based on any of the robot systems, comprising the following steps:

[0026] Multi-source perception data fusion and pose estimation, using extended Kalman filter to fuse the environment information obtained by the ground unmanned vehicle platform and the unmanned aerial vehicle, realizing high-precision positioning;

[0027] Abnormality detection, based on the multi-sensor data of the ground unmanned vehicle platform and the unmanned aerial vehicle and the pre-trained YOLO model, realizing abnormal feature extraction and identification;

[0028] Task scheduling and path planning, using a large model to perform dynamic task allocation of the ground unmanned vehicle platform and the unmanned aerial vehicle to realize optimal inspection path planning;

[0029] Intelligent charging of drones, based on visual recognition composite positioning identification and the drone's IMU data, enables precise landing and automatic charging of the drone.

[0030] Furthermore, the air-ground collaborative positioning and mapping and high-precision positioning steps specifically include: a perception positioning step, a map creation step and a high-precision positioning step;

[0031] The specific steps of UAV intelligent control and mission planning include: UAV flight control system design, anomaly detection, collaborative path algorithm based on large models, and UAV autonomous landing and charging.

[0032] The advantages and beneficial effects of the present invention are as follows:

[0033] The structural features and innovative features of the present invention are as follows:

[0034] 1. Air-Ground Collaborative Inspection Robot System: This system combines unmanned vehicles and drones to form an intelligent collaborative operation system. This hybrid robot design is not easily conceived in the field of port inspection, as it requires overcoming the challenges of integrating the different operating environments and methods of air and ground operations. Consequently, it significantly improves inspection coverage and efficiency while reducing the risks of manual operation.

[0035] 2. Intelligent Charging System Design: The circular grooved base on the top of the unmanned vehicle automatically docks with the power receiving component on the bottom of the drone for charging, which is a significant innovation in the field of robotics. The challenge lies in ensuring precise and stable electrical contact between the drone and the vehicle in a dynamic environment. This design enables automatic charging of the drone, extending operating time and reducing manual intervention.

[0036] 3. Emergency Equipment Rapid Replacement System: The emergency equipment compartment located on the unmanned vehicle enables rapid docking with drones to replace equipment such as robotic arms and fire spheres. This rapid replacement capability is difficult to implement in traditional inspection and emergency response systems. This design significantly improves the robot's adaptability and response speed, enabling timely response to various port emergencies.

[0037] Innovations in methods and steps

[0038] 1. Multimodal environmental perception data fusion: This technology fuses data from sensors such as lidar, depth cameras, and IMUs through an edge computing controller to achieve high-precision positioning and 3D modeling. While this fusion technology, while challenging in dealing with differences and noise in the data from different sensors, significantly improves the robot's perception accuracy and decision-making capabilities.

[0039] 2. Air-ground collaborative global map creation: Utilize multi-source sensor data of drones and unmanned vehicles to construct a global map of the port. The key of this innovation is to effectively integrate high-altitude and ground perspective information, which provides comprehensive and accurate environmental information for the robot system, supporting efficient path planning and dynamic obstacle avoidance.

[0040] 3. Collaborative path algorithm based on large models: Introduce large models for task scheduling and path planning. This method is not easy to think of in multi-agent systems, as it needs to handle complex decision spaces and real-time dynamic environments. By using large models, the system can achieve intelligent task allocation and path optimization, improving the efficiency of port inspection and the accuracy of emergency response.

[0041] The present application provides a port air-ground collaborative inspection robot system, which integrates inspection and maintenance distribution. Through intelligent cooperation of ground inspection robots and drones, it realizes functions such as container yard inspection, cargo state monitoring, dynamic path planning and emergency maintenance. The system uses multi-modal environment perception, Internet of Things, laser SLAM positioning and edge computing technology, and has the ability of autonomous navigation, dynamic obstacle avoidance, intelligent loading and unloading and automatic charging, and can adapt to complex dynamic environments in the port all day long. Compared with traditional operation mode, the system significantly improves the inspection accuracy and maintenance efficiency, reduces the risk of manual operation, optimizes the warehouse management and emergency response process, and provides an efficient and reliable innovative solution for smart port construction. BRIEF DESCRIPTION OF DRAWINGS

[0042] Figure 1 is the overall diagram of the port air-ground collaborative inspection robot provided by the present application;

[0043] Figure 2 is the overall flowchart of port air-ground collaborative inspection;

[0044] Figure 3 is the local perception and positioning flowchart of air-ground;

[0045] Figure 4 is the air-ground collaborative global map creation flowchart;

[0046] Figure 5 is the air-ground collaborative high-precision positioning flowchart;

[0047] Figure 6 is the schematic diagram of drone anomaly detection;

[0048] Figure 7 is the air-ground collaborative anomaly response flowchart;

[0049] Figure 8 is the air-ground collaborative inspection based on large model collaborative path planning schematic diagram.

[0050] Figure 9 is a schematic diagram of an air-ground cooperative intelligent charging system. DETAILED DESCRIPTION

[0051] The technical solutions in the embodiments of the present application will be described clearly and in detail below with reference to the drawings in the embodiments of the present application. The described embodiments are only some of the embodiments of the present application.

[0052] The technical solution of the present application to solve the above technical problems is:

[0053] In the description of the present application, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer" and the like indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, and 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, and therefore cannot be understood as a limitation on the present application. In addition, the terms "first", "second", "third" are only for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0054] In the description of the present application, it should be noted that unless otherwise explicitly specified and limited, the terms "mounting", "connection", "connection" should be understood broadly, for example, it can be fixedly connected, or it can be detachably connected, or integrally connected; it can be mechanically connected, or it can be electrically connected; it can be directly connected, or it can be indirectly connected through an intermediate medium, or it can be connected inside two elements. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0055] As Figure 1 shown in one air-ground cooperative intelligent inspection robot, comprising an unmanned vehicle platform, an unmanned aerial vehicle. The unmanned vehicle platform, the unmanned aerial vehicle will be described in detail below.

[0056] The unmanned vehicle platform comprises wheels and driving components arranged at the bottom of the vehicle body 1, and a depth camera, a three-dimensional laser radar, an edge computing controller, a storage compartment and a power chassis arranged on the vehicle body 1. Among them:

[0057] The wheel and driving component is a power system using Ackerman steering structure, including a front axle steering mechanism, a rear axle driving motor and a distributed electric control unit.

[0058] The environmental perception component includes a laser radar 4 and a depth camera 3. Among them:

[0059] Lidar 4 constructs a precise three-dimensional point cloud model of the surrounding environment in real time by emitting laser pulses and receiving reflected signals, providing high-reliability spatial information for robots. On unmanned vehicle platforms, unmanned aerial vehicles, the laser radar is used to realize autonomous navigation, obstacle detection and avoidance, and environmental mapping functions. Its high-frequency scanning and long-range measurement characteristics ensure the stability and safety of robots in complex dynamic environments. Laser radar can capture object distance and shape with centimeter-level accuracy, adapt to various lighting conditions, and have higher robustness than traditional sensors, significantly improving the perception ability and operation efficiency of robots in unknown or unstructured scenes.

[0060] Depth camera 3 acquires depth information of the surrounding environment in real time by infrared projection and sensor reception, and constructs a precise three-dimensional scene model. It can measure object distance and contour with millimeter-level accuracy, unaffected by lighting conditions, suitable for various complex environments indoors and outdoors. In unmanned aerial vehicle and robot applications, depth cameras are used for close-range obstacle detection, precise positioning, and three-dimensional reconstruction, with high frame rate and low latency characteristics ensuring the rapid response capability of the system.

[0061] Communicator 5 is a wireless communicator, mainly responsible for information transmission with unmanned aerial vehicles and base stations to ensure smooth information transmission. In the communicator 5, there are integrated inertial navigation equipment and GPS equipment.

[0062] Device 9 is an edge computing controller, responsible for integrating data from lidar, vision sensors and other devices, processing environmental information in real time and generating navigation instructions.

[0063] In addition, the unmanned vehicle platform includes a platform body 6 for the take-off and landing of unmanned aerial vehicles and charging. The platform body 6 provides a take-off and landing platform for unmanned aerial vehicles, and the circular groove of the platform body 6 is more suitable for random errors in the landing position of unmanned aerial vehicles. At the same time, the platform body 6 also provides electrical functions for unmanned aerial vehicles, which must be charged in time after completing a patrol task. Therefore, after the unmanned aerial vehicle returns to the take-off and landing platform, the platform body 6 charges it using a metal contact charging method. The high-power contact power supply technology provided by the platform body 6 can completely eliminate manual operation during charging of the unmanned aerial vehicle.

[0064] The unmanned vehicle platform further comprises an emergency equipment warehouse 13. The emergency equipment warehouse 13 comprises fixed columns 12 and lifting platforms 11 arranged symmetrically, and is used for carrying two replaceable emergency equipments: a universal mechanical arm 15 for unmanned aerial vehicle air equipment maintenance, and a universal fire ball 16 for unmanned aerial vehicle air fire extinguishing. The emergency equipments are integrated with standard circular docking discs 14, are fixed by elastic clamping mechanisms of the fixed columns 12 in a non-working state, and are lifted to a docking position by the lifting platforms 11 in a working state. The integrated design enables the unmanned aerial vehicle to quickly replace emergency equipments with different functions, realizes automatic docking of the equipments and the unmanned aerial vehicle through the standard interfaces of the discs 14, and can also return to the emergency equipment warehouse for reinstallation.

[0065] As shown in Figure 1 : the unmanned aerial vehicle is composed of the following core components: a circular support foot 10, a rotor system, a power supply unit and a sensing unit, which are integrated on a fuselage 2. The main control system comprises a flight management module and a wireless communication module. The flight management module is composed of a waterproof onboard box, a flight control core, a power distribution unit and a parameter debugging port assembled in the box, and the three are interconnected through electrical lines. The flight control core is integrated with a high-precision atmospheric pressure sensor, a three-dimensional space positioning device and a flight attitude compensation system. The wireless communication module adopts an independent sealing design, and is provided with an onboard data / image transceiver device and a ground station device matched with the device. The onboard device establishes a data link with the flight control core.

[0066] The power supply unit adopts an explosion-proof safety design, and mainly comprises a battery pack installed in an energy storage cabin, a power distribution hub, an electronic speed regulation device, a motor mounting bracket and a driving motor carried by the bracket. The battery pack is in communication with the power distribution unit and supplies power to the electronic speed regulator through the power hub. The speed regulator receives control signals from the debugging port at an input end, and directly drives the motors to operate at an output end. All electrical connections meet the explosion-proof safety standards.

[0067] As shown in Figure 2 , the following method for air-ground cooperative intelligent inspection is specifically described:

[0068] One: air-ground cooperative positioning and mapping and high-precision positioning:

[0069] Mainly includes perception positioning, map creation, high-precision positioning, specifically:

[0070] As shown in Figure 3 : the perception positioning step comprises:

[0071] As shown in the figure, perception and positioning are implemented in four parts: environmental data acquisition, sensor data preprocessing, feature extraction and matching, and pose estimation. Three-dimensional lidar data collects information about the robot's surroundings; lidar data is tightly coupled with IMU data to achieve high-frequency pose estimation; a depth camera captures visual data; and a GPS device provides global positioning information. The raw data collected by each sensor is synchronized in time and space, followed by feature extraction and matching. Features are extracted from point cloud data and image data and matched with information in existing maps or databases. The position and pose of the drone and unmanned vehicle in the global coordinate system are calculated based on sensor data, and algorithms are used to fuse multi-sensor data to improve positioning accuracy.

[0072] like Figure 4 As shown: The map creation steps include:

[0073] As shown in the figure, map creation is implemented in three steps: local mapping, data preprocessing, and global map assembly. Drones and autonomous vehicles each collaborate on multiple sensors to complete local mapping. The drone uses 3D LiDAR and a depth camera to collect high-altitude point clouds and depth data. After filtering and segmentation, it constructs a local map primarily from a bird's-eye view, while also recording its position and orientation. The autonomous vehicle, on the other hand, uses LiDAR and a high-definition camera to capture ground point clouds and visual features, generating a local ground map that includes detailed structure. Both vehicles require preparatory steps, such as sensor calibration and initial GPS positioning, to ensure that each local map has a traceable spatial reference.

[0074] Global map generation is then achieved through spatiotemporal alignment and multi-source data fusion. Data synchronization begins by aligning sensor data streams using timestamps and establishing a unified coordinate system using GPS or shared features. Maps are then stitched together using a two-step process of ICP coarse registration and image optimization fine registration, ultimately fusing wide-area coverage from an aerial perspective with detailed features from the ground.

[0075] like Figure 5 As shown: The high-precision positioning steps include:

[0076] As shown in the figure, high-precision positioning involves local positioning, global positioning, and extended Kalman filter fusion. UAVs and autonomous vehicles achieve local positioning through multi-sensor collaboration, using IMU, LiDAR, and depth camera data to generate high-frequency pose estimates. Global positioning relies on a global port map to provide an absolute position reference. Finally, an extended Kalman filter (EKF) is used to fuse local and global information, resulting in centimeter-level high-precision pose estimates.

[0077] 2. UAV Intelligent Control and Mission Planning

[0078] Mainly includes unmanned aerial vehicle flight control system design, anomaly detection, large model based collaborative path algorithm, unmanned aerial vehicle autonomous landing charging. Specifically:

[0079] Unmanned aerial vehicle flight control system design:

[0080] A six-degree-of-freedom system dynamics mathematical model is constructed for the unmanned aerial vehicle, and the aircraft actuation system is composed of motor-driven units and rotor aerodynamics modules. The air dynamic characteristic parameters of the aircraft are inversed through the lift and torque characteristic curves measured by experiments. Based on the input airflow velocity information and the rotational speed and angular velocity data of the four rotors, the current lift and torque effects of each execution unit can be solved in real time, thereby completing system simulation verification.

[0081] Based on the kinematics and dynamics of the aircraft, the system model can be divided into attitude dynamics and position dynamics two subsystems, and the corresponding motion control strategy is also divided into position servo control and attitude stabilization control two parts. A specific space coordinate and the combination of aircraft attitude are defined as a waypoint, and the flight path planning of the unmanned aerial vehicle can be regarded as an ordered set of multiple waypoints, and the aircraft needs to reach each preset waypoint in the given order.

[0082] As shown in Figure 6 , anomaly detection:

[0083] Port anomaly detection is implemented in three parts, including multi-source data acquisition and preprocessing, anomaly feature extraction, anomaly identification classification and response. During the port inspection task of the unmanned aerial vehicle, the IMU, laser radar and depth camera carried by the unmanned aerial vehicle cooperatively collect environmental and equipment state data, and the data are preprocessed to adapt to the input of the YOLO model. The trained YOLO model extracts abnormal features from the preprocessed data, generates boundary boxes, categories (fire abnormality, equipment abnormality, etc.) and confidence.

[0084] As shown in Figure 7 , air-ground collaborative anomaly response:

[0085] After the unmanned aerial vehicle identifies the abnormal classification during the port inspection task, the abnormal response mechanism is triggered, and the equipment replacement and task adjustment process are executed according to the anomaly detection result. The unmanned aerial vehicle carries the default equipment for the regular inspection task. When the anomaly detection identifies a scene inconsistent with the default equipment operation scene, the unmanned aerial vehicle automatically returns to the unmanned vehicle and arrives at the emergency equipment cabin. In the equipment cabin, the unmanned aerial vehicle replaces the required equipment through the docking platform and updates the task parameters. After the replacement is completed, the unmanned aerial vehicle replans the path based on the updated configuration and flies to the abnormal area to perform targeted operation.

[0086] As shown in Figure 8 , large model based collaborative path algorithm:

[0087] In the constructed map model, the inspection task area range and key inspection targets can be planned. As the total area of the robot patrol, a suitable cooperative inspection path is planned using a large model, in which the unmanned vehicle can release the unmanned aerial vehicle at any position, perform partial ground inspection after releasing the unmanned aerial vehicle, and then receive the unmanned aerial vehicle at another position. At the same time, the unmanned vehicle can also inspect the inspection dead angle of the unmanned aerial vehicle in the air to obtain specific information.

[0088] Specifically, in the path planning stage, the present research considers using a large model for task scheduling: using a relatively mature general language large model (such as Deepseek, Tongyi Qianwen, etc.) as the basis, a task loss function is set as the evaluation standard, and a teacher model is obtained based on this evaluation standard. Iterative direction to get in cooperative scheduling and path planning; on this basis, the expert model can be used to train a lightweight student model as the final lightweight expert model.

[0089] The map model, agent information, task point position information, and other restriction information are all input to the expert model, which can plan a scheduling scheme and a path according to the actual situation. The final path planning is obtained, which covers the release and recovery of the unmanned aerial vehicle by the unmanned vehicle.

[0090] At the same time, when the unmanned aerial vehicle or the unmanned vehicle detects an abnormality, it will notify the other party of the abnormal point information, and then the unmanned vehicle will provide repair tools (if the correct equipment is not carried on the vehicle, it will return to the warehouse to find it). Finally, the unmanned aerial vehicle uses the tool to repair the abnormal point and reports the abnormal information.

[0091] As shown in Figure 9 The unmanned aerial vehicle intelligent charging system:

[0092] The ground mobile robot platform (UGV) as the cooperative work base station of the unmanned aerial vehicle (UAV) has the function of full-automatic take-off and landing guidance of the unmanned aerial vehicle (UAV). After completing the inspection task or during the task when the power of the unmanned aerial vehicle is insufficient, the UAV identifies the composite positioning mark set on the platform body based on machine vision, which is a high-contrast concentric circle coding pattern, and embeds absolute position information based on QR code. Through the image processing unit, the center coordinates and attitude angle of the mark are calculated, and a six-degree-of-freedom relative pose estimation model is constructed combined with IMU data; the flight control system adjusts the thrust vector of the brushless motor according to the pose data, realizes the three-axis cooperative compensation of roll / pitch / yaw of the landing trajectory, and finally realizes the docking of the power receiving interface of the unmanned aerial vehicle support frame and the platform charging contact point.

[0093] In one aspect of this application, a drone uses a tapered guide ring wall to assist in precise landing when landing on the intelligent charging system on top of an unmanned vehicle. The tapered guide ring wall is tilted at a specific angle to the horizontal and has a wear-resistant layer on its inner surface, ensuring accurate and safe landing in all weather conditions. Technically, the tapered guide ring wall design leverages the drone's natural downward tendency during landing. By tilting at a specific angle, it can accurately guide the drone to the charging point even under wind. In principle, the tapered structure reduces the contact area between the drone and the top of the unmanned vehicle, reducing deviation caused by wind resistance. At the same time, the wear-resistant layer improves durability during long-term use. Effectively, the technology in this embodiment ensures safe landing of drones in complex environments, reduces the risk of equipment damage caused by inaccurate landing, and improves the overall stability and operational efficiency of the system. In other embodiments, landing accuracy and reliability can be further improved by adding visual markers to assist landing or adopting a magnetic landing method.

[0094] In one aspect of the present application, the specific angle is further 45°. Technically, the selection of the angle of 45° is the optimal value based on a large amount of experimental data, which can effectively balance the guidance effect and wind resistance of the drone during landing. In principle, the angle design follows the landing dynamics principle of the drone, ensuring the stability and safety of the drone during the landing process. In terms of effect, the technology in this embodiment significantly improves the landing success rate of the drone by optimizing the landing angle, reduces mission interruptions caused by landing failures, and enhances the continuous operation capability of the system. In other embodiments, the landing angle can also be adjusted according to different drone models and weights to accommodate more types of drones and improve the compatibility of the system.

[0095] According to one aspect of the present application, the charging contacts in the intelligent charging system are designed with elastic copper alloy contacts, which have the ability to automatically adapt to the power receiving components of the drone to achieve a stable and reliable electrical connection. Technically, the elastic copper alloy contact design overcomes the problem of slight displacement that may exist when the drone lands, ensuring that a stable electrical connection can be established even in the case of slight shaking. In principle, the application of elastic materials combined with the high conductivity of copper alloys form charging contacts that are both strong and have good conductive properties, thereby improving charging efficiency and safety. In terms of effect, the technology in this embodiment realizes rapid and automatic charging of drones, shortens the non-operating time of drones, and extends their continuous operating cycle. In other embodiments, wireless charging technology can be used to further reduce physical contact, reduce maintenance costs, and improve the convenience and safety of charging.

[0096] Further, the edge computing controller of the ground unmanned vehicle platform not only performs three-dimensional modeling of the environment and obstacle recognition, but also has task allocation and path optimization capabilities. Based on real-time environmental information and task requirements, the inspection strategy is adjusted to ensure efficient operation of the robot system in complex dynamic environments. Technically, the edge computing controller integrates advanced algorithms and processors to quickly process large amounts of data on site without relying on remote servers. In principle, through the fusion processing of multi-source perception data, the edge computing controller can generate an environmental map in real time, identify potential obstacles, and plan the optimal path. In terms of effects, the technology in this embodiment significantly improves the autonomous navigation capability of the unmanned vehicle and the inspection efficiency of the unmanned aerial vehicle, and reduces the decision lag caused by network delay or data transmission problems. In other embodiments, the edge computing controller can also be equipped with machine learning functions to enable self-learning and optimization, further improving the intelligence and adaptability of the system.

[0097] Further, the 5G / Beidou dual-mode communication module realizes ultra-high speed and low latency wireless communication in the port area for data interaction between the ground unmanned vehicle platform and the unmanned aerial vehicle. Technically, the 5G communication technology provides high-speed data transmission rate and low-latency communication characteristics, while the Beidou system provides high-precision positioning services. In principle, the high speed and low latency characteristics of 5G technology ensure the real-time and accuracy of data transmission between the unmanned vehicle and the unmanned aerial vehicle, and the high-precision positioning of the Beidou system provides reliable position information for the autonomous navigation of the unmanned vehicle. In terms of effects, the technology in this embodiment greatly improves the efficiency of the unmanned vehicle and the unmanned aerial vehicle working together, especially in emergency situations that require immediate response, it can quickly transmit critical information and speed up the decision-making process. In other embodiments, additional communication redundancies such as Wi-Fi or satellite communication can be added to enhance the communication stability and reliability of the system, especially in port environments with severe signal blockage.

[0098] An aspect of the present application further provides that the emergency equipment cabin of the unmanned vehicle platform can be equipped with various types of emergency equipment, including but not limited to rescue ropes, medical kits, lighting lamps, etc., which are axially aligned and coupled through standard circular docking discs and unmanned aerial vehicle bottom docking mechanisms. Technically, the emergency equipment cabin adopts a modular design, which can quickly replace equipment according to different task requirements. In principle, the axial alignment and coupling mechanism of the standard circular docking disc and the unmanned aerial vehicle bottom docking mechanism ensures the accuracy and stability of the unmanned aerial vehicle when replacing equipment. In terms of effect, the technology in this embodiment enables the unmanned aerial vehicle to flexibly change equipment according to task requirements, such as replacing lighting equipment during night patrol or replacing maintenance tools when equipment fails, greatly improving the emergency response capability and operational flexibility of the system. In other embodiments, the efficiency of replacing equipment can be further improved and the manual intervention can be reduced by adding an automatic replacement function of the equipment, such as using a mechanical arm to automatically grab and install the equipment.

[0099] An aspect of the present application further provides that the power system of the unmanned aerial vehicle adopts a modular waterproof and dustproof design to maintain stable operation in harsh environments. The bottom of the unmanned aerial vehicle is provided with a power receiving assembly, which automatically forms an electrical connection with the charging contact array when the unmanned aerial vehicle lands on the intelligent charging system on the top of the unmanned vehicle, realizing fast charging of the unmanned aerial vehicle. Technically, the modular design makes it more convenient to maintain and upgrade the power system, and the waterproof and dustproof design improves the adaptability and durability of the unmanned aerial vehicle in harsh environments. In principle, the modular design of the power system follows the standardization principle of the hardware of the unmanned aerial vehicle, facilitating quick replacement and maintenance, while the waterproof and dustproof design is based on the actual needs of the unmanned aerial vehicle in the port environment, ensuring the normal operation of the power system in rainy, dusty and other conditions. In terms of effect, the technology in this embodiment significantly improves the operational continuity and safety of the unmanned aerial vehicle, reducing the task interruption caused by equipment failure. In other embodiments, the endurance and power efficiency of the unmanned aerial vehicle can be further improved by using higher performance motors and more efficient battery management systems.

[0100] Further, the ground unmanned vehicle platform power system adopts Ackerman steering structure and has electronic differential function, realizing flexible steering and smooth driving in complex terrain. In terms of technology, the combination of Ackerman steering structure and electronic differential function enables the unmanned vehicle to maintain synchronization and coordination of all wheels during turning, improving the maneuverability and stability of the vehicle. In principle, the Ackerman steering structure follows the principles of vehicle dynamics to ensure the stability of the vehicle during turning, while the electronic differential function adjusts the speed of different wheels to achieve stable driving on uneven or slippery surfaces. In terms of effect, the technology in this embodiment significantly improves the adaptability and work efficiency of the unmanned vehicle in complex port environments, reducing operational errors caused by terrain limitations. In other embodiments, the off-road performance of the vehicle can be further improved by increasing the size of the tires or the strength of the suspension system, further improving the traffic capacity of the unmanned vehicle in harsh terrain conditions.

[0101] Further, the emergency equipment cabin of the ground unmanned vehicle platform is linked with the vehicle-mounted communication system to upload key data to the central dispatch platform, facilitating remote monitoring and maintenance management, and optimizing the emergency response process. In terms of technology, the linkage of the emergency equipment cabin and the vehicle-mounted communication system realizes real-time monitoring of equipment status and remote transmission of data. In principle, the vehicle-mounted communication system uploads real-time equipment status, remaining power and other key information in the emergency equipment cabin to the central dispatch platform, enabling dispatch personnel to monitor the availability and health of the equipment. In terms of effect, the technology in this embodiment significantly improves the remote monitoring capability and emergency response efficiency of the system, reducing operational delays caused by equipment failure or improper resource allocation. In other embodiments, the self-diagnosis function of the equipment can be further improved by adding sensors and algorithms, further improving the self-checking capability and preventive maintenance level of the equipment, and reducing the equipment failure rate.

[0102] Further, the unmanned aerial vehicle intelligent charging is based on visual recognition composite positioning markers and IMU data of the unmanned aerial vehicle to realize precise landing and automatic charging of the unmanned aerial vehicle. In terms of technology, the visual recognition composite positioning markers combined with IMU data form a dual positioning mechanism for the landing of the unmanned aerial vehicle. In principle, visual recognition composite positioning markers provide visual guidance for the landing of the unmanned aerial vehicle, while IMU data is used to correct attitude deviation during landing to ensure precise positioning and stable landing of the unmanned aerial vehicle. In terms of effect, the technology in this embodiment significantly improves the success rate of automatic charging of the unmanned aerial vehicle, reduces charging failures caused by inaccurate landing, and prolongs the continuous operation time of the unmanned aerial vehicle. In other embodiments, environmental adaptability protection mechanisms such as wind and rain protection facilities can be added to further improve the landing safety and charging reliability of the unmanned aerial vehicle in adverse weather conditions.

[0103] Further, in an aspect of the present application, the visual recognition composite positioning mark includes a logo pattern on the top of the ground unmanned vehicle platform, which matches the visual positioning system of the unmanned aerial vehicle, guiding the unmanned aerial vehicle to accurately align. In terms of technology, the design of the composite positioning mark takes into account the visual recognition ability of the unmanned aerial vehicle and the layout of the unmanned vehicle platform, ensuring that the unmanned aerial vehicle can quickly recognize and position when landing. In principle, the composite positioning mark is easy to capture by the visual system of the unmanned aerial vehicle through special design of color contrast and geometric shape, thereby guiding the unmanned aerial vehicle to adjust the landing posture and achieve accurate alignment. In terms of effect, the technology in this embodiment significantly improves the landing accuracy of the unmanned aerial vehicle, reduces the damage of the equipment caused by landing deviation, and improves the overall stability and operation efficiency of the system. In other embodiments, the recognition ability of the unmanned aerial vehicle in the night or low light conditions can be further improved by increasing the brightness of the mark or using night vision materials, to ensure accurate landing in all-weather conditions.

[0104] Further, in an aspect of the present application, the intelligent charging system is also provided with an environmental adaptability protection mechanism for protecting the charging contact point in extreme weather conditions to avoid short circuit or poor contact, ensuring the safety and reliability of the unmanned aerial vehicle charging. In terms of technology, the environmental adaptability protection mechanism adopts multiple protection measures such as waterproof, dustproof, and anti-freezing, ensuring the normal work of the charging contact point in various harsh weather conditions. In principle, the waterproof design prevents rainwater from entering, the dustproof design avoids dust accumulation, and the anti-freezing design maintains the temperature of the charging contact point through heating or insulation technology, which together ensures the electrical performance and mechanical strength of the charging contact point. In terms of effect, the technology in this embodiment significantly improves the stability and safety of the unmanned aerial vehicle charging, reduces the charging failure caused by environmental factors, and prolongs the continuous operation period of the unmanned aerial vehicle. In other embodiments, the self-cleaning function of the charging contact point can be added, such as using vibration or airflow sweeping to remove surface dust, to further improve the long-term performance and reliability of the charging contact point.

[0105] Further, the ground unmanned vehicle platform is provided with a self-balancing control system for maintaining the vehicle body stable on uneven terrain and ensuring continuous data collection and transmission during the inspection process. In terms of technology, the self-balancing control system uses advanced sensors and control algorithms to monitor the attitude changes of the unmanned vehicle in real time and restore balance by adjusting the speed and direction of the wheels. In principle, the self-balancing control system obtains the vehicle body attitude information based on sensors such as gyroscopes and accelerometers, adjusts the motion state of the wheels through the PID control algorithm, and realizes the dynamic balance of the vehicle body. In terms of effect, the technology in this embodiment significantly improves the stability of the unmanned vehicle in complex terrain and the reliability of data transmission, reducing data loss or miscommunication caused by vehicle body shaking. In other embodiments, the suspension system of the vehicle can be increased, such as using air springs or hydraulic shock absorbers, to further improve the comfort and stability of the unmanned vehicle on bumpy roads, ensuring the continuity and accuracy of data collection.

[0106] Further, the data interaction protocol between the unmanned aerial vehicle and the ground unmanned vehicle platform supports dynamic routing selection to cope with frequent signal blocking and interference in the port environment, ensuring the smoothness and integrity of data transmission. In terms of technology, the dynamic routing selection protocol uses adaptive hopping and re-routing mechanism to automatically find the best transmission path when encountering signal blocking or interference. In principle, the dynamic routing selection protocol is based on real-time analysis of network topology structure, and dynamically adjusts the data transmission path by evaluating the signal quality of different transmission nodes to avoid the influence of signal blocking and interference. In terms of effect, the technology in this embodiment significantly improves the data transmission quality and stability between the unmanned vehicle and the unmanned aerial vehicle, reduces communication interruption caused by signal problems, and ensures the continuity and efficiency of the inspection task. In other embodiments, additional relay nodes or multi-band communication technology can be added to further improve the communication coverage range and anti-interference ability of the system, adapting to larger-scale port inspection needs.

[0107] Further, the aspect of the present application, the anomaly detection is based on the multi-sensor data of the ground unmanned vehicle platform and the unmanned aerial vehicle and the pre-trained YOLO model, realizes the extraction and identification of abnormal features, and improves the inspection efficiency and safety. In terms of technology, the anomaly detection system adopts multi-sensor fusion technology and deep learning algorithm, which can extract abnormal features from complex environmental data and accurately identify them. In principle, the multi-sensor data fusion technology comprehensively processes the data of laser radar, depth camera, GPS and other sensors to generate a three-dimensional model of the environment, and the pre-trained YOLO model quickly identifies and classifies abnormal features based on these model data. In terms of effect, the technology in the embodiment significantly improves the accuracy and response speed of anomaly detection, reduces the safety hazards caused by not discovering abnormal conditions in time, and improves the overall efficiency and safety of port inspection. In other embodiments, the self-learning function of anomaly detection can also be added, such as using online learning algorithm to continuously optimize the YOLO model, further improving the intelligence and adaptability of anomaly detection, and coping with the dynamic changes of the port environment.

[0108] Further, the aspect of the present application, the real-time interaction of the data of the ground unmanned vehicle platform and the unmanned aerial vehicle also includes monitoring the health status of the unmanned aerial vehicle, implementing preventive maintenance, and prolonging the service life of the unmanned aerial vehicle. In terms of technology, the unmanned aerial vehicle health status monitoring system adopts sensor data acquisition and data analysis technology, which can monitor the key parameters of the unmanned aerial vehicle in real time, such as battery voltage, motor temperature, flight attitude, etc. In principle, the unmanned aerial vehicle health status monitoring system is based on real-time analysis of sensor data, sets threshold and early warning rules, and discovers potential faults of the unmanned aerial vehicle in time, such as battery overheating and motor wear, so as to take preventive maintenance measures in advance. In terms of effect, the technology in the embodiment significantly improves the maintenance efficiency and service life of the unmanned aerial vehicle, reduces the interruption of work caused by equipment failure, and reduces the maintenance cost. In other embodiments, the self-repair function of the unmanned aerial vehicle can also be added, such as using redundant system design and adaptive flight control algorithm, to further improve the reliability and operation continuity of the unmanned aerial vehicle, even in the case of partial component failure, the unmanned aerial vehicle can continue to perform tasks.

[0109] The working process or use process of the present application is as follows:

[0110] In the working process of the port open space cooperative intelligent autonomous inspection robot system, the unmanned vehicle and the unmanned aerial vehicle cooperate to realize the all-around inspection of the port environment. First, the 5G / Beidou dual-mode communication module equipped on the unmanned vehicle ensures real-time communication with the unmanned aerial vehicle and the central dispatching system. The edge computing controller on the unmanned vehicle processes multi-source sensing data, performs three-dimensional modeling of the environment and obstacle identification, and plans the optimal inspection path. When the unmanned aerial vehicle executes the inspection task and the power is insufficient, the intelligent charging system on the top of the unmanned vehicle guides the unmanned aerial vehicle to land accurately through visual recognition composite positioning marks to realize automatic charging. When encountering abnormal situations, the emergency equipment cabin of the unmanned vehicle provides modularized emergency equipment, and the unmanned aerial vehicle quickly replaces the equipment through the standard circular docking disc, such as replacing the fire-fighting equipment for fire extinguishing. The self-balancing control system of the unmanned vehicle ensures stable driving in complex terrain, and the dynamic routing protocol between the unmanned aerial vehicle and the unmanned vehicle ensures the smoothness of data transmission. The whole system can timely discover and handle abnormal situations through abnormal detection technology, and monitor the health status of the unmanned aerial vehicle to implement preventive maintenance, prolong the service life of the unmanned aerial vehicle, and improve the efficiency and safety of port inspection.

[0111] The embodiment can achieve the following beneficial effects:

[0112] 1. Cooperation of unmanned vehicle and unmanned aerial vehicle

[0113] The application designs a robot platform carrying an unmanned aerial vehicle, forming an integrated mobile inspection system. The unmanned vehicle not only solves the transportation and landing site restriction problem of the unmanned aerial vehicle, but also is equipped with a high-capacity battery to provide endurance support for the unmanned aerial vehicle, effectively prolonging the operation time. The cooperative working mode fully utilizes the advantages of wide-area coverage and flexible maneuvering of aerial inspection of the unmanned aerial vehicle, and fully combines the stable bearing and continuous supply characteristics of the ground platform. The system is equipped with an intelligent task allocation algorithm, which can automatically plan the optimal inspection path and cooperative strategy according to the actual operation demand. This cooperative working mode not only takes advantage of the aerial inspection of the unmanned aerial vehicle, but also utilizes the bearing and supply capacity of the ground platform, improving the comprehensive inspection efficiency and adaptability in complex environments.

[0114] 2. Synchronous mapping and positioning of dynamic environment:

[0115] In view of the complex and changeable characteristics of the port production environment, the application introduces the air-ground cooperative multi-robot SLAM technology, realizes the full-coverage environment modeling of the large-scale port area through the cooperative work of the unmanned aerial vehicle and the ground robot. The system integrates multi-source sensing data such as laser radar and visual sensor, significantly improves the positioning accuracy in the typical chemical scene with dense pipelines and complex equipment, and has reliable positioning with centimeter-level accuracy in various mixed environments, providing strong technical support for port safety production.

[0116] 3. Large model-based path planning technology

[0117] For the case of complex port environment and large number of agents, it is difficult to obtain good results using conventional deep learning algorithms. The present application introduces a large model to solve the multi-agent dynamic task allocation and collaborative task planning problem in the task scheduling stage, and finally obtains the collaborative path to detect and repair the abnormal operation. This provides a strong technical guarantee for solving such complex problems.

[0118] 4. Efficient UAV equipment replacement and operation

[0119] In response to the demand for rapid response to emergency tasks, the present application designs a modular unmanned aerial vehicle equipment quick replacement system. The system uses standardized docking interfaces and automatic lifting mechanisms to support the quick switching of various functional modules such as maintenance and fire fighting, achieving flexible configuration of unmanned aerial vehicles during task execution.

[0120] The system, device, module or unit illustrated in the above embodiments can be specifically implemented by a computer chip or entity, or by a product with certain functions.

[0121] It should be further noted that the terms "comprising", "including", or any other variant thereof are intended to cover non-exclusive inclusion, so that processes, methods, articles or devices including a series of elements not only include those elements, but also include other elements not explicitly listed, or further include elements inherent in such processes, methods, articles or devices. Without more limitations, the element defined by the statement "including a" does not exclude the presence of additional identical elements in the process, method, article or device including the element.

[0122] The above examples are understood to be used only to illustrate the present application and not to limit the protection scope of the present application. After reading the content of the present application, the skilled person can make various changes or modifications to the present application, and these equivalent changes and modifications also fall within the scope defined by the claims of the present application.

Claims

1. A port air-ground collaborative intelligent autonomous inspection robot system, characterized by: include: A ground unmanned vehicle platform is equipped with a 5G / Beidou dual-mode communication module for realizing real-time data interaction with drones. The ground unmanned vehicle platform is also provided with an edge computing controller for processing multi-source perception data to realize three-dimensional environmental modeling and obstacle recognition. An intelligent charging system comprises a circular groove base and a charging contact array for realizing automatic and precise landing and charging of drones. An emergency equipment cabin is modularly designed and equipped with a standard circular docking disc and a lifting platform for realizing rapid equipment replacement of the drone. A drone is integrated with a multimodal environmental perception system, including a three-dimensional lidar, a depth camera, a GPS device, and an inertial measurement unit (IMU) for acquiring environmental information.

2. The port air-ground collaborative intelligent autonomous inspection robot system according to claim 1 is characterized in that: It also includes an environmental perception component, which includes a lidar and a depth camera 3. The lidar emits laser pulses and receives reflected signals to construct an accurate 3D point cloud model of the surrounding environment in real time, providing the robot with highly reliable spatial information. The depth camera uses infrared projection and sensor reception to obtain real-time depth information of the surrounding environment and construct an accurate 3D scene model.

3. The port air-ground collaborative intelligent autonomous inspection robot system according to claim 1 is characterized in that: The unmanned vehicle platform includes a platform body for drone take-off, landing and charging. The platform body provides a carrying and take-off and landing platform for the drone. The platform body is a circular groove, and the circular groove base is set on the top of the vehicle body. The circular groove adopts a concentric circle stepped structure design, including a tapered guide ring wall; the center of the circular groove base is integrated with a charging contact array, the tapered guide ring wall is inclined at a 40°-50° angle to the horizontal plane and the inner surface is provided with a wear-resistant layer, and the charging contacts are designed with elastic copper alloy contacts.

4. The port air-ground collaborative intelligent autonomous inspection robot system according to claim 3 is characterized in that: The drone is equipped with a distributed power system, including brushless motors, matching electronic speed regulators and propellers installed at the four corners of the body. The brushless motors are used to provide the core driving force for the drone's flight, the electronic speed regulators are used to accurately adjust the speed of the brushless motors, and the propellers are used to convert the mechanical energy of the motor's rotation into lift. The power system adopts a modular design, and each component is waterproof and dustproof; the brushless motor and electronic speed regulator casings are made of high-strength aluminum alloy.

5. The port air-ground collaborative intelligent autonomous inspection robot system according to claim 3 is characterized in that: The unmanned vehicle adopts a power system with an Ackerman steering structure, including a front axle steering mechanism, a rear axle drive motor and a distributed electronic control unit. The power system and the vehicle control system realize data exchange through the CAN bus, providing power control and data feedback; the front axle steering mechanism is driven by a servo motor and integrates a steering angle sensor and a torque feedback device to achieve precise steering control; the rear axle drive motor is equipped with an electronic differential function, and the torque of the left and right wheels can be adjusted independently, thereby improving the stability of the vehicle on curves and uneven roads.

6. The port air-ground collaborative intelligent autonomous inspection robot system according to claim 1 is characterized in that: The multi-source perception system realizes the functions of three-dimensional modeling of the environment and obstacle recognition by communicating with the edge computing controller. The data of the three-dimensional lidar and depth camera are fused and processed by the edge computing controller for path planning and obstacle avoidance decisions.

7. The port air-ground collaborative intelligent autonomous inspection robot system according to claim 1 is characterized in that: The air-ground collaborative system consists of unmanned vehicles and drones, with a three-layer structure. On the top layer, the top of the unmanned vehicle houses a drone landing pad with a retractable hatch. When the drone lands, the hatch automatically opens, providing a precise landing platform. When the drone is stored, the hatch is sealed to protect the drone from the outside environment; the upper second floor: emergency equipment warehouse, used to store emergency equipment and dock with the drone; Lower layer: The unmanned vehicle's power chassis, which integrates the Ackerman steering structure power system and environmental perception system to support ground navigation and cargo transportation.

8. The port air-ground collaborative intelligent autonomous inspection robot system according to claim 1 is characterized in that: The emergency equipment warehouse includes symmetrically arranged fixed columns and a lifting platform, which are used to carry two types of replaceable emergency equipment: a universal robotic arm for drone aerial equipment maintenance, and a universal fire ball for drone aerial firefighting; the emergency equipment are all integrated with a standard circular docking disc, which is fixed by the elastic clamping mechanism of the fixed column in the non-working state and lifted to the docking position by the lifting platform when working; the equipment can be automatically docked with the drone through the standard interface of the disc, and can also be returned to the emergency equipment warehouse for reinstallation.

9. An inspection method based on the robot system according to any one of claims 1 to 8, characterized in that: The following steps are involved: Multi-source perception data fusion and pose estimation, using extended Kalman filtering to fuse the environmental information obtained by the ground unmanned vehicle platform and the UAV to achieve high-precision positioning; Anomaly detection, based on the multi-sensor data of the ground unmanned vehicle platform and the drone and the pre-trained YOLO model, to achieve anomaly feature extraction and identification; Task scheduling and path planning: using a large model to dynamically allocate tasks between the ground unmanned vehicle platform and the drone to achieve optimal inspection path planning; Intelligent charging of drones, based on visual recognition composite positioning identification and the drone's IMU data, enables precise landing and automatic charging of the drone.

10. The inspection method according to claim 9, characterized in that: The air-ground collaborative positioning and mapping and high-precision positioning steps specifically include: a perception positioning step, a map creation step and a high-precision positioning step; The specific steps of UAV intelligent control and mission planning include: UAV flight control system design, anomaly detection, collaborative path algorithm based on large models, and UAV autonomous landing and charging.

Images