Satellite positioning-based unmanned aerial vehicle autonomous inspection obstacle avoidance system

Abstract

This invention discloses an autonomous inspection and obstacle avoidance system for unmanned aerial vehicles (UAVs) based on satellite positioning. It relates to the field of UAV inspection and includes a vehicle-mounted mobile base station, satellite positioning, and multi-source perception fusion. The vehicle-mounted mobile base station serves as the central hub, constructing a localized communication network and providing support for all-terrain deployment, continuous power supply, and equipment takeoff and landing recovery. The satellite positioning module adapts to multi-mode signals and achieves centimeter-level positioning using RTK technology, switching to a backup fusion positioning method when satellite rejection occurs. The multi-source perception fusion module integrates three types of sensors, generating perception data through noise reduction and weighted fusion. The environmental modeling module constructs a 3D map and safety thresholds based on SLAM technology. The path planning module plans global and local paths through improved algorithms, enabling multi-device collaborative scheduling of UAVs and robot dogs. The safety management module implements access control, anomaly alarms, and emergency switching. The system does not rely on external communication, solving the management challenges in remote, signal-free scenarios, improving inspection efficiency and safety, and can be integrated with existing systems.

Description

Technical Field

[0001] This invention relates to the field of autonomous inspection and obstacle avoidance for unmanned aerial vehicles (UAVs), specifically to an autonomous inspection and obstacle avoidance system for UAVs based on satellite positioning. Background Technology

[0002] Construction sites, such as power line inspections, mountain engineering projects, and facility maintenance in remote areas, often face complex scenarios with no communication signal coverage. Traditional operation modes rely on manual inspections or single-equipment operation, resulting in low efficiency, poor safety, and high management difficulty. With the development of drone and robotic dog collaborative operation technology, intelligent inspection and construction management are becoming the trend. However, existing technologies still have significant pain points in scenarios without signal coverage and in vehicle-mounted mobile deployments:

[0003] Lack of on-site management in signal-free scenarios: Traditional construction management systems heavily rely on public network communication signals. In signal-free environments such as deep mountains and remote construction areas, remote control, task distribution, and data transmission of drones and robotic dogs are impossible, leading to equipment collaboration failures and inability to monitor operational status. This necessitates close-range manual operation, which is insufficient to meet the needs of large-scale, refined construction management. Existing multi-device collaboration systems rely on external communication networks for data synchronization. In signal-free scenarios, control commands, operational data, and equipment status cannot be transmitted, causing drones and robotic dogs to become disconnected, hindering the orderly progress of construction tasks. Furthermore, the lack of local data storage and interaction mechanisms means that work records cannot be retained in real time, making subsequent traceability and data analysis difficult. Summary of the Invention

[0004] The purpose of this invention is to provide a satellite positioning-based autonomous inspection and obstacle avoidance system for unmanned aerial vehicles (UAVs) to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: an autonomous inspection and obstacle avoidance system for unmanned aerial vehicles based on satellite positioning, comprising a vehicle-mounted mobile base station module, a satellite positioning module, a multi-source perception fusion module, an environmental modeling module, a path planning module, a multi-device collaborative scheduling module, and a safety management and monitoring feedback module.

[0006] The vehicle-mounted mobile base station module, serving as the system deployment and communication hub, integrates an all-terrain carrying platform, a dual-link communication unit, and a UPS power supply. It builds a local communication network in scenarios without communication signals, providing support for mobile deployment, one-click take-off and landing, automatic charging, and return-to-base recovery for drones and robot dogs. At the same time, it enables collaborative management and data transmission of multiple devices.

[0007] The satellite positioning module interfaces with the vehicle-mounted mobile base station module, adapts to multi-mode satellite positioning signals, and uses RTK carrier phase differential technology to obtain centimeter-level positioning and attitude information. In satellite denial scenarios, it automatically switches to a backup positioning method that combines laser, inertial odometry and vision, inertial odometry. The positioning data is synchronized to the environment modeling module and the path planning module.

[0008] The multi-source perception fusion module transmits data through the communication link of the vehicle-mounted mobile base station module, integrates orthogonal lidar, binocular vision sensor and millimeter-wave radar, collects environmental perception data, and generates a perception dataset after noise reduction and weighted fusion, and feeds it back to the environmental modeling module.

[0009] The environmental modeling module, based on the positioning data from the satellite positioning module, the perception dataset from the multi-source perception fusion module, and prior information on power transmission lines, towers, and mountain terrain, uses SLAM technology to construct a 3D dense point cloud map and an occupied grid map, dynamically sets a safe distance threshold to generate an environmental model, and outputs it to the path planning module.

[0010] The path planning module receives the environmental model from the environmental modeling module and the inspection task issued by the vehicle-mounted mobile base station module. It presets multi-scenario strategies and plans the global path and local trajectory through the improved fast random tree algorithm and the improved artificial potential field method. The planning results are transmitted to the multi-device collaborative scheduling module.

[0011] The multi-device collaborative scheduling module relies on the localized communication network of the vehicle-mounted mobile base station module to receive path instructions from the path planning module, establish an action collaboration mechanism between the drone and the robot dog, transmit control instructions through dual links to achieve real-time synchronization of task distribution and action trajectory, and feed back the device status to the safety management and monitoring feedback module.

[0012] The security management and monitoring feedback module connects with various functional modules to collect operational data. It has functions such as hierarchical permission management, data encryption, and log management. It builds a visual monitoring panel, triggers alarms and switches to backup communication and positioning methods when anomalies occur, and outputs optimization instructions to the path planning module and the multi-source perception fusion module.

[0013] Furthermore, the carrier platform of the vehicle-mounted mobile base station module feeds back the power supply status of the UPS power supply and the deployment status of the take-off and landing platform to the safety management and monitoring feedback module in real time through the localized communication network. At the same time, it receives the scheduling instructions from the safety management and monitoring feedback module to adjust the deployment priority of the drones and robot dogs.

[0014] Furthermore, the positioning data from the satellite positioning module is transmitted to the environmental modeling module via a dual-link connection from the vehicle-mounted mobile base station module. This is used to calibrate the spatial coordinates of the three-dimensional dense point cloud map and to provide the path planning module with real-time position references for the drone and the robot dog, ensuring the accuracy of path adjustments.

[0015] Furthermore, when the sensing dataset of the multi-source sensing fusion module is transmitted to the environment modeling module via the vehicle-mounted mobile base station module, the real-time monitoring of the safety control and monitoring feedback module is triggered simultaneously. When there is an anomaly in the sensing data, the safety control and monitoring feedback module immediately instructs the multi-source sensing fusion module to re-collect the data.

[0016] Furthermore, in the environmental model generated by the environmental modeling module, the safe distance threshold data is synchronized to the path planning module and the multi-device collaborative scheduling module. The path planning module plans obstacle avoidance paths based on the threshold, and the multi-device collaborative scheduling module manages the movement boundaries of the drone and the robot dog based on the threshold.

[0017] Furthermore, the multi-scenario strategy switching command of the path planning module is issued by the vehicle-mounted mobile base station module. After switching, the path planning module immediately feeds back the strategy execution status to the safety management and monitoring feedback module, and simultaneously synchronizes it to the multi-device collaborative scheduling module to adjust the device action mode. The improved fast random tree algorithm of the path planning module expands the search tree in segments with adjacent towers as temporary target points. When selecting random state nodes, it combines the maximum turning angle constraint of the UAV and determines the node selection range according to the angle relationship between the node and the obstacle, thereby improving the feasibility and efficiency of global path planning. In the improved artificial potential field method of the path planning module, the repulsive potential field function introduces the distance factor between the current position of the UAV and the target point. By coordinating forces, it avoids the local optimum problem caused by the mutual cancellation of attraction and repulsion, and optimizes the smoothness and safety of the local obstacle avoidance trajectory.

[0018] Furthermore, when transmitting control commands, the multi-device collaborative scheduling module ensures the continuity of command transmission through the dual-link redundancy design of the vehicle-mounted mobile base station module, and at the same time transmits the movement trajectories of the drone and the robot dog back to the environment modeling module in real time for updating the environment model.

[0019] Furthermore, the access control function of the security management and monitoring feedback module is linked with the multi-device collaborative scheduling module. Only authorized personnel can issue task instructions through the command and control platform of the vehicle-mounted mobile base station module, and the operation records are synchronously stored in the log management module.

[0020] Furthermore, the system's standardized API interface connects with the existing company system through the vehicle-mounted mobile base station module, enabling bidirectional interaction of inspection task data, equipment status data, and obstacle avoidance result data. The interactive data is encrypted by the safety management and monitoring feedback module.

[0021] Furthermore, the command and control platform of the vehicle-mounted mobile base station module displays the equipment movement trajectory, satellite positioning accuracy, and obstacle distance data transmitted by the safety management and monitoring feedback module in real time through the external display screen. Operators can issue adjustment commands through the remote control handle, which are then transmitted to the drone and robot dog through the multi-device collaborative scheduling module.

[0022] Compared with the prior art, the beneficial effects of the present invention are:

[0023] By building a localized communication network and control center based on vehicle-mounted mobile base station modules, drones and robot dogs can be quickly deployed, tasks distributed, collaboratively operated, and their status monitored without relying on external communication signals. This completely solves the problem of construction control interruption in the absence of signal and supports intelligent operations in complex construction sites such as deep mountains and remote areas.

[0024] The vehicle-mounted all-terrain platform is suitable for mountainous and complex road conditions, and can quickly reach the construction site; it integrates dual drone take-off and landing platforms and robot dog storage and charging units, supporting one-click take-off and landing, automatic charging and return to the center for recovery, which greatly shortens the deployment time; the UPS power module provides continuous energy supply, adapting to construction sites without power supply conditions, and ensuring the continuous operation of the control system and operating equipment.

[0025] Through the vehicle-mounted local communication network, the control commands, operation data, and equipment status of the drones can be exchanged in real time, clarifying the division of tasks and action logic, and avoiding equipment coordination conflicts. The dual-link communication design ensures uninterrupted command transmission, supports multiple people and multiple drones to be online for simultaneous management, adapts to the multi-task and all-round operation needs of construction sites, and improves construction efficiency.

[0026] The safety management and monitoring feedback module constructs a localized safety management system, adopting hierarchical permission management and dual login verification to prevent unauthorized operations; it monitors the equipment's operating trajectory, satellite positioning accuracy, and obstacle distance in real time, triggering audible and visual alarms and switching to backup positioning or communication methods in case of anomalies, thus avoiding risks such as equipment collisions and deviations from the operating range in a timely manner; and it stores the operation logs locally for easy subsequent traceability and safety analysis.

[0027] The path planning module features pre-defined strategies specific to each construction scenario, dynamically optimizing the operational paths of drones and robotic dogs based on construction progress and terrain changes. The multi-device collaborative scheduling module supports dynamic task adjustments, enabling rapid response to temporary needs at the construction site (such as re-surveying of key areas and avoidance of sudden obstacles). The system supports private deployment and standardized API interfaces, allowing integration with existing management systems at the construction site to link construction data with control commands, thereby improving overall construction management efficiency. The vehicle-mounted mobile base station module integrates equipment storage, charging, debugging, and control functions, simplifying equipment maintenance processes and reducing the number of on-site operators. Localized data storage and interaction mechanisms avoid dependence on external communication and power supply, reducing external costs for construction management. The modular design facilitates equipment maintenance and upgrades, adapts to the management needs of different types of construction sites, and enhances system reusability. Attached Figure Description

[0028] Figure 1 This is the framework of the unmanned aerial vehicle (UAV) automated inspection system of the present invention;

[0029] Figure 2 This is a schematic diagram of the forces acting on the artificial potential field method of the present invention;

[0030] Figure 3 This is a dense point cloud map of the present invention;

[0031] Figure 4 This is the occupancy grid map of the present invention;

[0032] Figure 5 This is a schematic diagram of the path planning of the present invention. Detailed Implementation

[0033] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0034] Please see Figure 1 —5. The present invention provides an autonomous inspection and obstacle avoidance system for unmanned aerial vehicles based on satellite positioning, comprising a vehicle-mounted mobile base station module, a satellite positioning module, a multi-source perception fusion module, an environmental modeling module, a path planning module, a multi-device collaborative scheduling module, and a safety management and monitoring feedback module:

[0035] The vehicle-mounted mobile base station module, serving as the system deployment and communication hub, integrates an all-terrain carrying platform, a dual-link communication unit, and a UPS power supply. It builds a local communication network in scenarios without communication signals, providing support for mobile deployment, one-click take-off and landing, automatic charging, and return-to-base recovery for drones and robot dogs. At the same time, it enables collaborative management and data transmission of multiple devices.

[0036] After the all-terrain carrying platform arrives at the work site, the operator can start the take-off and landing platform deployment procedure with one click through the control interface of the command and control platform. The pull-out lifting platform can be deployed in a short time to form a stable drone take-off and landing area. The load accessory storage area allows for the direct removal of spare drones, robot dogs, gimbals, and other equipment to quickly complete equipment debugging.

[0037] In the absence of external communication signals, the dual-link communication unit automatically starts to build a local wireless communication network. After the drone or robot dog starts, it automatically connects to the network to achieve bidirectional data transmission with the vehicle-mounted command and control platform, ensuring real-time interaction of control commands and inspection data.

[0038] The UPS power module automatically connects to the vehicle's power supply system and can also be supplemented with power through solar panels, providing continuous energy for the vehicle-mounted command and control platform, communication unit, take-off and landing platform, and equipment charging. Its built-in battery management system monitors the power supply status in real time, and automatically switches the power supply mode and sends alarm information to the safety management module when an abnormality is detected. The command and control platform supports two people working simultaneously. Operators can view the equipment status in real time through the external display screen and remotely control the drone and robot dog using a single-grip or double-grip joystick. They can switch shooting angles and adjust inspection routes to achieve precise inspection operations.

[0039] The satellite positioning module interfaces with the vehicle-mounted mobile base station module, adapts to multi-mode satellite positioning signals, and uses RTK carrier phase differential technology to obtain centimeter-level positioning and attitude information. In satellite denial scenarios, it automatically switches to a backup positioning method that combines laser, inertial odometry and vision, inertial odometry. The positioning data is synchronized to the environment modeling module and the path planning module.

[0040] After startup, the module automatically scans and adapts to multi-mode satellite signals such as GPS, GLONASS, and BeiDou. It optimizes the positioning data using RTK carrier phase differential technology to generate 3D position and attitude information for the drone and robot dog, ensuring positioning accuracy during the inspection process. When satellite signals are blocked in the work area, the module automatically triggers a backup positioning mechanism, switching to a positioning method that combines laser-inertial odometry and vision-inertial odometry. By fusing inertial measurement data and visual perception data, the positioning information is supplemented, avoiding positioning interruption. The positioning data is transmitted in real time to the environmental modeling module and the path planning module via dual links of the vehicle-mounted mobile base station, providing spatial coordinate references for map construction and real-time location data for equipment adjustments, ensuring accurate matching between the path planning and the actual location of the equipment.

[0041] The multi-source perception fusion module transmits data through the communication link of the vehicle-mounted mobile base station module, integrates orthogonal lidar, binocular vision sensor and millimeter-wave radar, collects environmental perception data, and generates a perception dataset after noise reduction and weighted fusion, and feeds it back to the environmental modeling module.

[0042] After the orthogonal lidar is activated, it performs a full-range scan of obstacles in both horizontal and vertical directions, capturing the outline and position information of the obstacles. A binocular vision sensor simultaneously acquires near-range environmental images, and through feature point extraction and disparity calculation, obtains the three-dimensional coordinates of near-range obstacles. A millimeter-wave radar, designed for complex weather conditions such as rain, fog, and smoke, penetrates interference to detect distant obstacles. All three sensors acquire data simultaneously, achieving full coverage of the sensing range. The acquired raw data is transmitted to the module processing unit via a localized communication network. First, a Kalman filter algorithm is used to reduce noise and remove invalid data caused by environmental interference. Then, a weighted fusion algorithm integrates the effective data from the three sensors to generate a unified multi-dimensional obstacle perception dataset, ensuring data consistency and integrity. The fused perception dataset is fed back to the environmental modeling module in real time, providing the latest obstacle information for map updates. Simultaneously, the safety control module triggers real-time monitoring. When abnormal perception data is detected, the safety control module immediately instructs the module to re-acquire data, ensuring the reliability of the perception.

[0043] The environmental modeling module, based on the positioning data from the satellite positioning module, the perception dataset from the multi-source perception fusion module, and prior information on power transmission lines, towers, and mountain terrain, uses SLAM technology to construct a 3D dense point cloud map and an occupied grid map, dynamically sets a safe distance threshold to generate an environmental model, and outputs it to the path planning module.

[0044] After receiving the position data from the satellite positioning module and the perception dataset from the multi-source perception fusion module, the module starts SLAM technology. First, it calibrates the spatial coordinates using the fusion data of laser-inertial odometry and vision-inertial odometry. Then, it constructs a three-dimensional dense point cloud map to intuitively present the spatial distribution of environmental obstacles. At the same time, it generates an occupancy grid map, which uses the grid cell status to characterize the obstacle occupancy and clearly defines the free areas and dangerous areas.

[0045] The module pre-imports prior information such as power transmission lines, tower distribution, and mountainous terrain, and integrates this information with the real-time constructed map to identify key landmarks and dangerous boundaries in the inspection area, thus avoiding the omission of core obstacle information in the map.

[0046] The safety distance threshold is dynamically adjusted according to the type of obstacle. The threshold data is synchronized to the path planning module and the multi-device collaborative scheduling module to provide obstacle avoidance basis for path planning and define the safety boundary for device movement.

[0047] The path planning module receives the environmental model from the environmental modeling module and the inspection tasks issued by the vehicle-mounted mobile base station module. It pre-sets multiple scenario strategies and plans the global path and local trajectory using an improved fast random tree algorithm and an improved artificial potential field method. The planning results are transmitted to the multi-device collaborative scheduling module. The module pre-sets four core scenario strategies: routine inspection, fault alarm, weather response, and mountain terrain adaptation. Operators can select the corresponding scenario through the vehicle-mounted command and control platform, or the system can automatically trigger strategy switching based on perception data. Global path planning uses an improved fast random tree algorithm, expanding the search tree segment by segment using adjacent towers and other key landmarks as temporary target points. It combines the maximum turning angle constraint of the UAV to select nodes, generating an efficient and feasible global inspection route. Local path planning uses an improved artificial potential field method, introducing distance factors and coordination forces to avoid the mutual cancellation of attraction and repulsion, generating a smooth local obstacle avoidance trajectory. When a sudden obstacle is detected, the local trajectory is adjusted in real time, and the global path is updated synchronously. The planned path instructions are transmitted to the multi-device collaborative scheduling module through the local communication network, clarifying the inspection routes, task division, and obstacle avoidance requirements for the UAV and the robot dog.

[0048] The multi-scenario strategy switching command of the path planning module is issued by the vehicle-mounted mobile base station module. After switching, the path planning module immediately feeds back the strategy execution status to the safety control and monitoring feedback module, and simultaneously synchronizes it to the multi-device collaborative scheduling module to adjust the device action mode. The improved fast random tree algorithm of the path planning module expands the search tree in segments with adjacent towers as temporary target points. When selecting random state nodes, it combines the maximum turning angle constraint of the UAV and determines the node selection range according to the angle relationship between the node and the obstacle, thereby improving the feasibility and efficiency of global path planning. In the improved artificial potential field method of the path planning module, the repulsive potential field function introduces the distance factor between the current position of the UAV and the target point. By coordinating forces, it avoids the local optimum problem caused by the mutual cancellation of attraction and repulsion, and optimizes the smoothness and safety of the local obstacle avoidance trajectory.

[0049] The multi-device collaborative scheduling module, relying on the localized communication network of the vehicle-mounted mobile base station module, receives path instructions from the path planning module, establishes a collaborative action mechanism between the drone and the robot dog, and achieves real-time synchronization of task distribution and movement trajectory through dual-link transmission of control instructions, while feeding back the device status to the safety management and monitoring feedback module. After receiving the path instructions from the path planning module, the module assigns specific tasks to the drone and robot dog based on the inspection task requirements, clarifying the operation sequence and collaborative logic. Control instructions are preferentially transmitted through the 2.4G link, and when the link transmission quality is detected to be substandard, it automatically switches to the 433M backup link to ensure uninterrupted instruction transmission. At the same time, the movement trajectory and operation status of the drone and robot dog are transmitted back to the vehicle-mounted command and control platform and the environmental modeling module in real time through dual links for map updates and status monitoring. The module coordinates the operating range of the drone and robot dog in real time to avoid collisions between devices. When one device detects a sudden obstacle, it immediately sends an obstacle avoidance collaborative instruction to the other device to ensure the continuity of the overall inspection task.

[0050] The security management and monitoring feedback module connects with various functional modules to collect operational data. It features hierarchical access control, data encryption, and log management, and constructs a visual monitoring panel. In case of anomalies, it triggers alarms and switches to backup communication and positioning methods, while simultaneously outputting optimization instructions to the path planning module and the multi-source perception fusion module. A hierarchical authorization mechanism is adopted, assigning different permissions to operators based on department and role, ensuring only authorized personnel can issue task instructions. It supports dual login verification using passwords and verification codes, and all operational behaviors are automatically recorded in the log management module for traceability. All transmitted and stored data is encrypted to meet system security requirements. The visual monitoring panel provides real-time updates. The system currently collects operational data from each module, including equipment movement trajectory, satellite positioning accuracy, obstacle distance, link transmission quality, and equipment status. When anomalies, sensor malfunctions, or link interruptions are detected, an audible and visual alarm is immediately triggered, and the system automatically switches to a backup positioning or communication method while simultaneously recording an anomaly log. The module supports private deployment and can interface with existing company systems through standardized API interfaces to achieve bidirectional interaction of inspection task data, equipment status data, and obstacle avoidance result data. It also has data analysis capabilities, outputting optimization instructions to the path planning module and multi-source perception fusion module through statistical analysis of operational data, continuously improving system operational efficiency and safety.

[0051] The vehicle-mounted mobile base station module's carrying platform uses a localized communication network to feed back the power supply status of the UPS and the deployment status of the take-off and landing platform to the safety management and monitoring feedback module in real time. At the same time, it receives scheduling instructions from the safety management and monitoring feedback module to adjust the deployment priority of drones and robot dogs.

[0052] The positioning data from the satellite positioning module is transmitted to the environmental modeling module via a dual-link connection from the vehicle-mounted mobile base station module. This data is used to calibrate the spatial coordinates of the 3D dense point cloud map and to provide the path planning module with real-time position references for the drone and robot dog, ensuring the accuracy of path adjustments.

[0053] When the sensing dataset from the multi-source sensing fusion module is transmitted to the environment modeling module via the vehicle-mounted mobile base station module, it simultaneously triggers real-time monitoring by the safety control and monitoring feedback module. When there is an anomaly in the sensing data, the safety control and monitoring feedback module immediately instructs the multi-source sensing fusion module to re-collect the data.

[0054] In the environmental model generated by the environmental modeling module, the safe distance threshold data is synchronized to the path planning module and the multi-device collaborative scheduling module. The path planning module plans obstacle avoidance paths based on the threshold, and the multi-device collaborative scheduling module manages the movement boundaries of the drone and the robot dog based on the threshold.

[0055] When transmitting control commands, the multi-device collaborative scheduling module ensures the continuity of command transmission through the dual-link redundancy design of the vehicle-mounted mobile base station module. At the same time, it transmits the movement trajectories of the drone and robot dog back to the environment modeling module in real time for updating the environment model.

[0056] The access control function of the security management and monitoring feedback module is linked with the multi-device collaborative scheduling module. Only authorized personnel can issue task instructions through the command and control platform of the vehicle-mounted mobile base station module, and the operation records are synchronously stored in the log management module.

[0057] The system's standardized API interface connects with the existing company system through the vehicle-mounted mobile base station module, enabling two-way interaction of inspection task data, equipment status data, and obstacle avoidance result data. The interactive data is encrypted by the safety management and monitoring feedback module.

[0058] The command and control platform of the vehicle-mounted mobile base station module displays the equipment movement trajectory, satellite positioning accuracy, and obstacle distance data transmitted by the safety management and monitoring feedback module in real time through the external display screen. Operators can issue adjustment commands through the remote control handle, which are then transmitted to the drone and robot dog through the multi-device collaborative scheduling module.

[0059] The vehicle-mounted all-terrain platform arrives at the work site. The operator starts the vehicle-mounted mobile base station module, deploys the take-off and landing platform, sets up the local communication network, takes out the drone and robot dog, and completes the equipment startup and network access; the UPS power module starts and enters continuous power supply state.

[0060] The satellite positioning module starts up and acquires initial positioning data. The multi-source sensing fusion module starts up simultaneously to collect environmental data, which is then processed and transmitted to the environmental modeling module. The environmental modeling module combines prior information to construct a three-dimensional environmental model and a safe distance threshold, and outputs it to the path planning module.

[0061] The path planning module selects the corresponding scenario strategy based on the inspection task requirements and the environmental model, generates global and local path instructions, and transmits them to the multi-device collaborative scheduling module; the module distributes the task and path instructions to the drones and robot dogs.

[0062] The drone and robot dog perform inspection tasks according to instructions. The satellite positioning module provides real-time location information, the multi-source perception fusion module continuously collects environmental data, and if an obstacle is detected, the path planning module adjusts the path in real time, and the collaborative scheduling module coordinates the equipment's obstacle avoidance actions to ensure operational safety.

[0063] The safety management module monitors the operation status in real time and triggers alarms and emergency handling when abnormalities occur. After the inspection is completed, the system completes the acceptance operation through three-dimensional digital means, uploads the inspection data to the existing company system, and the operator retrieves the equipment, shuts down the vehicle-mounted mobile base station module, and completes the operation.

Claims

1. A satellite-positioning-based autonomous unmanned aerial vehicle (UAV) inspection and obstacle avoidance system, characterized in that, It includes a vehicle-mounted mobile base station module, a satellite positioning module, a multi-source sensing fusion module, an environmental modeling module, a path planning module, a multi-device collaborative scheduling module, and a safety management and monitoring feedback module. The vehicle-mounted mobile base station module, serving as the system deployment and communication hub, integrates an all-terrain carrying platform, a dual-link communication unit, and a UPS power supply. It builds a local communication network in scenarios without communication signals, providing support for mobile deployment, one-click take-off and landing, automatic charging, and return-to-base recovery for drones and robot dogs. At the same time, it enables collaborative management and data transmission of multiple devices. The satellite positioning module interfaces with the vehicle-mounted mobile base station module, adapts to multi-mode satellite positioning signals, and uses RTK carrier phase differential technology to obtain centimeter-level positioning and attitude information. In satellite denial scenarios, it automatically switches to a backup positioning method that combines laser, inertial odometry and vision, inertial odometry. The positioning data is synchronized to the environment modeling module and the path planning module. The multi-source perception fusion module transmits data through the communication link of the vehicle-mounted mobile base station module, integrates orthogonal lidar, binocular vision sensor and millimeter-wave radar, collects environmental perception data, and generates a perception dataset after noise reduction and weighted fusion, and feeds it back to the environmental modeling module. The environmental modeling module, based on the positioning data from the satellite positioning module, the perception dataset from the multi-source perception fusion module, and prior information on power transmission lines, towers, and mountain terrain, uses SLAM technology to construct a 3D dense point cloud map and an occupied grid map, dynamically sets a safe distance threshold to generate an environmental model, and outputs it to the path planning module. The path planning module receives the environmental model from the environmental modeling module and the inspection task issued by the vehicle-mounted mobile base station module. It presets multi-scenario strategies and plans the global path and local trajectory through the improved fast random tree algorithm and the improved artificial potential field method. The planning results are transmitted to the multi-device collaborative scheduling module. The multi-device collaborative scheduling module relies on the localized communication network of the vehicle-mounted mobile base station module to receive path instructions from the path planning module, establish an action collaboration mechanism between the drone and the robot dog, transmit control instructions through dual links to achieve real-time synchronization of task distribution and action trajectory, and feed back the device status to the safety management and monitoring feedback module. The security management and monitoring feedback module connects with various functional modules to collect operational data. It has functions such as hierarchical permission management, data encryption, and log management. It builds a visual monitoring panel, triggers alarms and switches to backup communication and positioning methods when anomalies occur, and outputs optimization instructions to the path planning module and the multi-source perception fusion module.

2. The satellite positioning-based autonomous inspection and obstacle avoidance system for unmanned aerial vehicles according to claim 1, characterized in that, The vehicle-mounted mobile base station module's carrying platform feeds back the UPS power supply status and the deployment status of the take-off and landing platform to the safety management and monitoring feedback module in real time through a localized communication network. At the same time, it receives scheduling instructions from the safety management and monitoring feedback module to adjust the deployment priority of drones and robot dogs.

3. The satellite positioning-based autonomous inspection and obstacle avoidance system for unmanned aerial vehicles according to claim 1, characterized in that, The positioning data from the satellite positioning module is transmitted to the environmental modeling module via a dual-link connection from the vehicle-mounted mobile base station module. This data is used to calibrate the spatial coordinates of the 3D dense point cloud map and to provide the path planning module with real-time position references for the drone and robot dog, ensuring the accuracy of path adjustments.

4. The satellite positioning-based autonomous inspection and obstacle avoidance system for unmanned aerial vehicles according to claim 1, characterized in that, When the sensing dataset of the multi-source sensing fusion module is transmitted to the environment modeling module via the vehicle-mounted mobile base station module, the real-time monitoring of the safety control and monitoring feedback module is triggered simultaneously. When there is an anomaly in the sensing data, the safety control and monitoring feedback module immediately instructs the multi-source sensing fusion module to re-collect the data.

5. The satellite positioning-based unmanned aerial vehicle (UAV) autonomous inspection and obstacle avoidance system according to claim 1, characterized in that, In the environment model generated by the environment modeling module, the safe distance threshold data is synchronized to the path planning module and the multi-device collaborative scheduling module. The path planning module plans obstacle avoidance paths based on the threshold, and the multi-device collaborative scheduling module manages the movement boundaries of the drone and the robot dog based on the threshold.

6. The satellite positioning-based autonomous inspection and obstacle avoidance system for unmanned aerial vehicles according to claim 1, characterized in that, The multi-scenario strategy switching command of the path planning module is issued by the vehicle-mounted mobile base station module. After switching, the path planning module immediately feeds back the strategy execution status to the safety management and monitoring feedback module, and simultaneously synchronizes it to the multi-device collaborative scheduling module to adjust the device action mode. The improved fast random tree algorithm of the path planning module expands the search tree in segments with adjacent towers as temporary target points. When selecting random state nodes, it combines the maximum turning angle constraint of the UAV and determines the node selection range according to the angle relationship between the node and the obstacle, thereby improving the feasibility and efficiency of global path planning. In the improved artificial potential field method of the path planning module, the repulsive potential field function introduces the distance factor between the current position of the UAV and the target point. By coordinating forces, it avoids the local optimum problem caused by the mutual cancellation of attraction and repulsion, and optimizes the smoothness and safety of the local obstacle avoidance trajectory.

7. The satellite positioning-based autonomous inspection and obstacle avoidance system for unmanned aerial vehicles according to claim 1, characterized in that, When transmitting control commands, the multi-device collaborative scheduling module ensures the continuity of command transmission through the dual-link redundancy design of the vehicle-mounted mobile base station module. At the same time, it transmits the movement trajectories of the drone and the robot dog back to the environment modeling module in real time for updating the environment model.

8. The satellite positioning-based autonomous inspection and obstacle avoidance system for unmanned aerial vehicles according to claim 1, characterized in that, The access control and monitoring feedback module is linked with the multi-device collaborative scheduling module. Only authorized personnel can issue task instructions through the command and control platform of the vehicle-mounted mobile base station module, and the operation records are synchronously stored in the log management module.

9. The satellite positioning-based autonomous inspection and obstacle avoidance system for unmanned aerial vehicles according to claim 1, characterized in that, The system's standardized API interface connects with the existing company system through the vehicle-mounted mobile base station module, enabling two-way interaction of inspection task data, equipment status data, and obstacle avoidance result data. The interactive data is encrypted by the safety management and monitoring feedback module.

10. The satellite positioning-based unmanned aerial vehicle (UAV) autonomous inspection and obstacle avoidance system according to claim 1, characterized in that, The command and control platform of the vehicle-mounted mobile base station module displays the equipment movement trajectory, satellite positioning accuracy, and obstacle distance data transmitted by the safety management and monitoring feedback module in real time through the external display screen. Operators can issue adjustment commands through the remote control handle, which are transmitted to the drone and robot dog through the multi-device collaborative scheduling module.

Images