Multi-unmanned aerial vehicle cooperative inspection system

Abstract

The invention discloses a multi-unmanned aerial vehicle cooperative inspection system, and relates to the fields of a biological signal identification technology, an unmanned aerial vehicle technology, a bird-like bionic algorithm and multi-unmanned aerial vehicle cooperative management. The system comprises five modules: a biological signal multi-mode acquisition and identification module, an environment perception and multi-machine safety evaluation module, a bird-like bionic path planning and flight control module, a global communication and data processing module, and a ground management and control and cooperative scheduling module. Compared with the prior art, the method has the remarkable advantages that optical limitation is broken through through multi-mode identification, and the omission ratio of biological hidden dangers is reduced; the bird-like bionic algorithm and multiple machines cooperate to improve the inspection efficiency and shorten the period; a dynamic evaluation and collision avoidance mechanism guarantees multi-machine safety; the environmental adaptability is enhanced through three-dimensional meteorological capture and bionic control; the integrated function can be flexibly adapted to multiple industries, and the cross-scene application cost is reduced.

Description

Technical Field

[0001] This invention relates to the fields of biosignal recognition technology, unmanned aerial vehicle (UAV) technology, bird-like bionic algorithms, and multi-UAV collaborative management, specifically a multi-UAV collaborative inspection system. Background Technology

[0002] With the rapid development of the low-altitude economy, low-altitude intelligent inspection has become a core means of infrastructure maintenance and safety monitoring in various industries. Currently, the mainstream inspection technologies mainly include the following three categories: Traditional manual / manned aircraft inspection: relies on human or manned aircraft, and is limited by labor costs, safety risks and environmental adaptability. It is difficult to cover complex terrain (such as mountains and canyons) and dangerous areas (such as high-voltage facilities and toxic environments), and the inspection efficiency is low.

[0003] Single drone inspection: It achieves automated data collection by being equipped with cameras and infrared sensors, but it lacks the ability to recognize biosignals and cannot effectively monitor safety hazards caused by biological activities (such as birds nesting on power lines and wild animals damaging pipelines).

[0004] Existing multi-machine collaborative inspection methods mostly adopt fixed path planning, lack dynamic environmental adaptation capabilities, and are prone to problems such as path redundancy and high collision risk in complex terrain (urban high-rise buildings blocking the view) and severe weather (strong winds, fog and haze). Furthermore, a sound multi-machine operation safety assessment mechanism has not been established. Summary of the Invention

[0005] The purpose of this invention is to provide a multi-UAV collaborative inspection system to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: A multi-UAV collaborative inspection system, the system comprising: The biosignal multimodal acquisition and recognition module is used to acquire and recognize biological information and send the information to the global communication and data processing module; The environmental perception and multi-drone safety assessment module is used to dynamically acquire environmental information and determine the safety information of the UAV status, generate environmental safety assessment information and UAV status safety information and send them to the global communication and data processing module. The bird-like bionic path planning and flight control module is used to generate planned path information based on bird-like bionic algorithms and send the planned path information to the global communication and data processing module, and receive UAV control information sent by the global communication and data processing module; The global communication and data processing module is used to receive information sent by the above modules and UAV control command information sent by the ground control and collaborative scheduling module, generate UAV control information, and send it to the bird-like bionic path planning and flight control module. At the same time, it sends the generated UAV control information to the ground control and collaborative scheduling module. The ground control and collaborative scheduling module is used to generate UAV control command information and send it to the global communication and data processing module. It also receives UAV control information sent by the global communication and data processing module in real time, and generates and outputs inspection data information.

[0007] As a further aspect of the present invention: the biosignal multimodal acquisition and recognition module includes: The sensor configuration unit is used to receive information collected by the camera, microphone, electrode sensor, head-mounted sensor and infrared sensor, and generate a multimodal bio-information array; The signal processing unit is used to perform noise reduction and amplification processing on the above array information; and to extract ECG R wave / ST segment, EEG frequency band energy and speech feature parameter information. The data fusion unit, based on the synchronous acquisition of signals from multiple sensors, is used to identify biological species, identity, and activity status information; The first sending unit is used to send the biological species, identity, and activity status information identified by the data fusion unit to the global communication and data processing module.

[0008] As a further aspect of the present invention: the environmental perception and multi-machine security assessment module includes: A three-dimensional meteorological field construction unit is used to acquire GIS geographic information of low-altitude areas, multi-source sensor scanning data, and real-time meteorological parameter information from UAVs. The three-dimensional meteorological field is constructed by simplifying the model through the Kriging spatial interpolation algorithm and the Navier-Stokes equation, combined with CFD physical simulation. An incremental Gaussian process regression algorithm is used to periodically update the meteorological field information. The safety index calculation unit is used to establish the mapping relationship between environmental parameters and signal strength through a linear regression model. Combined with the predicted trajectory of the UAV and the dwell time in the grid, it calculates the stability index and marks UAVs with a stability index below a threshold as abnormal objects. The formula for calculating the stability index is as follows: In the formula, m is the number of grid combinations traversed by the drone, and t n For the duration of traversing the nth combination, PSP n t represents the percentage of the predicted signal strength for the nth combination. future Predicting the duration for the future; The multi-machine cooperative collision avoidance unit is used to calculate the safe distance based on the real-time velocity of the abnormal object. The calculation formula is as follows: In the formula, SP now At the current speed, SP sta The reference speed is L, and the distance influence coefficient is L. It is also used to predict the three-dimensional trajectories of multiple aircraft, identify objects affected by trajectories that intersect, reduce the flight speed of risky objects and generate new trajectories, iteratively detect secondary conflicts, until the distance between multiple aircraft at all time points is greater than the safe distance; The second transmitting unit is used to send meteorological field information, stability index information, and safe distance information to the global communication and data processing module.

[0009] As a further embodiment of the present invention: the bird-like bionic path planning and flight control module includes: The core unit of the biomimetic algorithm is used for bird-like biomimetic algorithms based on particle swarm optimization. It simulates the foraging behavior of birds, treats each UAV as a particle in a flock of birds, and dynamically adjusts flight parameters through individual optimal paths and global optimal paths to obtain adaptive path planning information. The flight control unit, based on model predictive control methods and combining UAV dynamics models with real-time sensor data, is used to achieve stable flight through PID attitude adjustment; it integrates GPS and / or BeiDou dual-mode positioning, inertial navigation and visual navigation to ensure positioning accuracy. The obstacle avoidance unit is used to detect obstacles using lidar and ultrasonic sensors, and generate obstacle avoidance paths by combining bird-like bionic algorithms. In the event of strong winds and rain, it dynamically adjusts its flight speed and altitude to simulate the headwind flying skills of birds and obtain obstacle avoidance path information. The third sending unit is used to send adaptive path planning information, positioning progress information and obstacle avoidance path information to the global communication and data processing module. The receiving control unit is used to receive UAV control information sent by the global communication and data processing module.

[0010] As a further embodiment of the present invention: the global communication and data processing module includes: The communication unit is used to achieve high-speed data backhaul via 4G or 5G in areas with good signal; and to ensure basic communication via data transmission radio in areas with weak signal. The security unit is used to protect data privacy with AES encryption algorithm and ensure data transmission reliability through channel coding and retransmission mechanism. The data receiving unit is used to receive data information sent by the first sending unit, the second sending unit, the third sending unit, and the task management unit based on the streaming data processing framework. The data mining unit is used to perform real-time data analysis by applying K-Means clustering and LSTM prediction after cleaning, deduplication, and error correction. The fourth transmitting unit is used to send the UAV control information obtained from the above data analysis to the receiving control unit and the ground control and coordination scheduling module.

[0011] As a further embodiment of the present invention: the ground control and collaborative scheduling module includes: The real-time monitoring unit is used to receive UAV control information sent by the fourth transmitting unit and display the UAV flight status, three-dimensional weather field, biometric identification results, and abnormal object alarms. Task Management Unit: Used to send custom inspection information, dynamically allocate tasks across multiple machines, and optimize resource allocation; Data output unit: Used to generate inspection reports from the information acquired by the real-time monitoring unit.

[0012] Compared with the prior art, the beneficial effects of the present invention are: ① Accurate identification of biological hazards: Multimodal biosignal recognition breaks through the limitations of traditional optics, is not affected by light or obstruction, and significantly reduces the false negative rate of biological hazards.

[0013] ② Improve inspection efficiency: Bird-like bionic algorithm dynamically optimizes the path, combined with multi-machine collaborative scheduling, significantly improving the inspection coverage area per unit time and greatly shortening the inspection cycle.

[0014] ③ Ensuring multi-aircraft safety: Dynamic safety assessment and collaborative collision avoidance mechanisms significantly reduce the risk of multi-aircraft collisions and avoid secondary conflicts caused by local avoidance.

[0015] ④ Environmental adaptability: The three-dimensional dynamic meteorological field captures environmental changes in real time, and the bird-like bionic flight control enhances the ability to adapt to severe weather, effectively reducing the interruption rate of inspection tasks.

[0016] ⑤ Expanding multi-scenario adaptation: It integrates biological monitoring, facility inspection, and safety collaboration functions, and can be flexibly adapted to industries such as power, petroleum, and environmental protection, significantly reducing the cost of cross-scenario applications. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of a multi-UAV collaborative inspection system.

[0018] Figure 2 This is a schematic diagram of the structure of a biosignal multimodal acquisition and recognition module in a multi-UAV collaborative inspection system.

[0019] Figure 3 This is a schematic diagram of the structure of an environmental perception and multi-drone safety assessment module in a multi-drone collaborative inspection system.

[0020] Figure 4 This is a schematic diagram of the structure of a bird-like bionic path planning and flight control module in a multi-UAV collaborative inspection system.

[0021] Figure 5 This is a schematic diagram of the structure of the global communication and data processing module in a multi-UAV collaborative inspection system.

[0022] Figure 6 This is a schematic diagram of the ground control and collaborative scheduling module in a multi-UAV collaborative inspection system.

[0023] Figure 7 This is a physical deployment diagram of the ground control center in a multi-UAV collaborative inspection system. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0025] Figure 1 This diagram illustrates the structure of a multi-UAV collaborative inspection system according to an embodiment of the present invention. Figure 7 This is a physical deployment diagram of a ground control center for a multi-UAV collaborative inspection system, the system comprising: The biosignal multimodal acquisition and recognition module 11 is used to acquire and recognize biological information and send the information to the global communication and data processing module; The environmental perception and multi-drone safety assessment module 12 is used to dynamically acquire environmental information and determine the safety information of the UAV status, generate environmental safety assessment information and UAV status safety information and send them to the global communication and data processing module. The bird-like bionic path planning and flight control module 13 is used to generate planned path information based on the bird-like bionic algorithm and send the planned path information to the global communication and data processing module, and receive UAV control information sent by the global communication and data processing module. The global communication and data processing module 14 is used to receive information sent by the above modules and UAV control command information sent by the ground control and collaborative scheduling module, generate UAV control information, and send it to the bird-like bionic path planning and flight control module. At the same time, it sends the generated UAV control information to the ground control and collaborative scheduling module. The ground control and collaborative scheduling module 15 is used to generate UAV control command information and send it to the global communication and data processing module, receive UAV control information sent by the global communication and data processing module in real time, and generate and output inspection data information.

[0026] Figure 2 This diagram illustrates the structural block diagram of a biosignal multimodal acquisition and recognition module 11 provided in an embodiment of the present invention. The biosignal multimodal acquisition and recognition module 11 includes: The sensor configuration unit 111 is used to receive information collected by a CMOS camera (face recognition, iris recognition), MEMS microphone (speech recognition), dry electrode sensor (ECG signal), high-resolution head-mounted sensor (EEG signal), and uncooled infrared sensor (body temperature signal) to generate a multimodal bio-information array. Signal processing unit 112 is used to perform noise reduction and amplification processing on the above array information; and to extract ECG R wave / ST segment, EEG frequency band energy and speech feature parameter information; Processing: The original signal is denoised and amplified using a Butterworth filter and wavelet transform; Feature extraction: Extract core parameters such as ECG R wave / ST segment, EEG frequency band energy, and speech features; Identification and judgment: CNN (face recognition), RNN / LSTM (behavior analysis), and SVM classifiers are used to compare with a biometric template library to identify the species, identity, and activity status of the organism.

[0027] The data fusion unit 113 realizes the synchronous acquisition of signals from multiple sensors based on the time synchronization mechanism, which is used to identify biological species, identity and activity status information. Multimodal data fusion improves the reliability of identification in complex environments and avoids missed detections caused by the failure of a single sensor. The first sending unit 114 is used to send the biological species, identity and activity status information identified by the data fusion unit to the global communication and data processing module.

[0028] Figure 3 This diagram illustrates the structural block diagram of the environmental perception and multi-machine security assessment module 12 provided in an embodiment of the present invention. The environmental perception and multi-machine security assessment module 12 includes: The 3D meteorological field construction unit 121 is used to acquire low-altitude GIS geographic information, multi-source sensor scanning data (terrain elevation, obstacle outlines), and real-time meteorological parameter information (temperature, humidity, wind speed, air pressure) from UAVs. It constructs the 3D meteorological field using the Kriging spatial interpolation algorithm (for processing discrete meteorological data) and the Navier-Stokes equations simplified model (for simulating fluid dynamics), combined with CFD physical simulation. An incremental Gaussian process regression algorithm is used to periodically (every 10 seconds) update the meteorological field information, avoiding the bottleneck of full-scale computation. The safety index calculation unit 122 adopts an adaptive octree structure in the three-dimensional meteorological field, divides the cubic mesh based on the greatest common divisor of the UAV sampling distance, and merges small volume meshes (volume < u³) at the boundary through the SurfaceNets algorithm to reduce redundant calculations; This is used to establish the mapping relationship between environmental parameters (temperature, humidity) and signal strength through a linear regression model. Combined with the predicted trajectory of the UAV and the dwell time in the grid, a stability index is calculated. UAVs with a stability index below a threshold are marked as abnormal objects. The formula for calculating the stability index is as follows: In the formula, m is the number of grid combinations traversed by the drone, and t n For the duration of traversing the nth combination, PSP n t represents the percentage of the predicted signal strength for the nth combination. future Predicting the duration for the future; The multi-machine cooperative collision avoidance unit 123 is used to calculate the safe distance based on the real-time speed of the abnormal object. The calculation formula is as follows: In the formula, SP now At the current speed, SP sta The reference speed is L, and the distance influence coefficient is L. It is also used to predict the three-dimensional trajectories of multiple aircraft, identify objects affected by trajectories that intersect, reduce the flight speed of risky objects and generate new trajectories, iteratively detect secondary conflicts, until the distance between multiple aircraft at all time points is greater than the safe distance; The second transmitting unit 124 is used to transmit meteorological field information, stability index information, and safe distance information to the global communication and data processing module.

[0029] Figure 4 This diagram illustrates the structural block diagram of a bird-like bionic path planning and flight control module 13 provided in an embodiment of the present invention. The bird-like bionic path planning and flight control module 13 includes: The core unit 131 of the biomimetic algorithm is used for bird-like biomimetic algorithms based on particle swarm optimization. It simulates the foraging behavior of birds, treats each UAV as a particle in the flock, and dynamically adjusts flight parameters through the individual optimal path (its own flight experience) and the global optimal path (the flock's flight experience) to obtain adaptive path planning information. The flight control unit 132, based on model predictive control, combines the UAV dynamics model with real-time sensor data to achieve stable flight through PID attitude adjustment; it integrates GPS and / or BeiDou dual-mode positioning, inertial navigation and visual navigation to ensure positioning accuracy (error ≤ 1m); The obstacle avoidance unit 133 is used to detect obstacles using lidar (range accuracy ≤5cm) and ultrasonic sensors, and generate obstacle avoidance paths by combining bird-like bionic algorithms; in response to strong winds and rain, it dynamically adjusts flight speed and altitude, simulates the headwind flying skills of birds, improves environmental adaptability, and obtains obstacle avoidance path information. The third sending unit 134 is used to send adaptive path planning information, positioning progress information and obstacle avoidance path information to the global communication and data processing module. The receiving control unit 135 is used to receive UAV control information sent by the global communication and data processing module.

[0030] Figure 5 This diagram illustrates the structural block diagram of the global communication and data processing module 14 provided in an embodiment of the present invention. The global communication and data processing module 14 includes: The communication unit 141 is used to achieve high-speed data backhaul via 4G or 5G in areas with good signal (such as cities) (supporting real-time transmission of high-definition images and biological signals); and to ensure basic communication via data transmission radio in areas with weak signal (such as mountainous areas). The security unit 142 is used to protect data privacy with AES encryption algorithm, and to ensure data transmission reliability through channel coding (such as LDPC code) and retransmission mechanism to avoid packet loss and tampering. The data receiving unit 143 is used to receive data information sent by the first sending unit, the second sending unit, the third sending unit and the task management unit based on a streaming data processing framework; and to receive data (biological signals, flight parameters, meteorological data) through a streaming data processing framework (such as Flink). The data mining unit 144 is used to perform real-time data analysis and feedback (delay < 2 seconds) by using K-Means clustering (to uncover biological activity patterns) and LSTM prediction (to provide early warning of equipment defects and safety hazards) after cleaning, deduplication, and error correction. The fourth transmitting unit 145 is used to transmit the UAV control information obtained from the above data analysis to the receiving control unit and the ground control and coordination scheduling module.

[0031] Figure 6 This diagram illustrates the structure of a ground control and coordinated scheduling module 15 provided in an embodiment of the present invention. The ground control and coordinated scheduling module 15 includes: The real-time monitoring unit 151 is used to receive the UAV control information sent by the fourth sending unit and display the UAV flight status (position, speed, battery level), three-dimensional weather field, biometric identification results, and abnormal objects are highlighted in red as an alarm. Task Management Unit 152: Used to send customized inspection information (area, duration, key monitoring targets), dynamically allocate tasks across multiple machines, and optimize resource allocation; Data output unit 153: Used to generate inspection reports (including hazard locations, biological activity records, and flight trajectories) from information acquired by the real-time monitoring unit, supporting export to Excel / PDF format and providing decision-making suggestions.

[0032] 1. Hardware selection and setup Hardware integration: A multimodal biosignal sensor group is installed at the front end, a flight controller and high-precision navigation module are deployed in the center, and a dual-mode communication module is configured at the top; the modules are interconnected with the CAN bus via a high-speed data cable (USB3.0), and key components (sensors and controllers) are fixed with shockproof brackets to improve anti-interference capabilities.

[0033] 2. Software programming and algorithm implementation Development architecture: Based on the Linux operating system (Ubuntu 20.04), it adopts a hybrid programming of Python (biosignal processing, machine learning) and C++ (real-time flight control, communication protocol) to ensure computational efficiency (algorithm response time < 1 second).

[0034] Core algorithm deployment: Data interaction protocol: A standardized data format (JSON) is defined to enable real-time interaction between sensors, flight controllers, and ground control centers. Data encryption (AES-256) and verification (CRC32) are supported to ensure data integrity.

[0035] 3. System integration and debugging (1) Integration process Hardware integration testing: Verify the accuracy of sensor data acquisition (e.g., infrared temperature measurement error ≤ 0.3℃), the response speed of flight controller commands (≤ 0.5 seconds), and the transmission stability of communication modules (packet loss rate < 1%).

[0036] Software deployment: Burn customized firmware (adapted to drone models), optimize driver compatibility; complete algorithm model training (such as establishing a biometric template library with a sample size of ≥1000 groups) and parameter calibration (obstacle avoidance threshold, safety index threshold).

[0037] Scenario-based testing: Plains scenario (railway inspection): Verify the efficiency of multi-machine collaborative path planning and ensure no omissions in coverage; Urban scenario (infrastructure inspection): Testing navigation accuracy and obstacle avoidance capabilities under the obstruction of tall buildings; Mountainous scenario (power line inspection): Verify flight stability and biometric accuracy under adverse weather conditions (light rain, level 3 wind).

[0038] (2) Debugging and optimization Data anomalies: Check sensor shielding and grounding, optimize signal filtering algorithms (e.g., add a 50Hz power frequency notch filter), and reduce electromagnetic interference; Flight jitter: Calibrate the IMU sensors (gyroscope, accelerometer), adjust the PID parameters (proportional coefficient 0.8, integral coefficient 0.2, derivative coefficient 0.1) and the stiffness of the shock-absorbing bracket; Recognition error: Expand the biometric dataset (add samples with different lighting and poses), and optimize the SVM kernel function parameters and multimodal fusion weights; Coordination conflict: Adjust the safety distance coefficient L (value 1.2-1.5) and trajectory prediction time (5-10 seconds) to reduce multi-machine scheduling delay.

[0039] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0040] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

[0041] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A multi-UAV cooperative inspection system, characterized in that, The system comprises: a biological signal multi-modal acquisition and identification module for acquiring and identifying biological information and sending the information to a global communication and data processing module; an environment perception and multi-UAV safety evaluation module for dynamically obtaining environment information and judging UAV state safety information, generating environment safety evaluation information and UAV state safety information, and sending the information to the global communication and data processing module; a bird-like bionic path planning and flight control module for generating planning path information based on a bird-like bionic algorithm and sending the planning path information to the global communication and data processing module and receiving UAV control information sent by the global communication and data processing module; the global communication and data processing module for receiving information sent by the above modules and UAV control instruction information sent by a ground control and cooperative scheduling module, generating UAV control information, and sending the UAV control information to the bird-like bionic path planning and flight control module and simultaneously sending the generated UAV control information to the ground control and cooperative scheduling module; the ground control and cooperative scheduling module for generating UAV control instruction information and sending the information to the global communication and data processing module, receiving UAV control information sent by the global communication and data processing module in real time, generating and outputting patrol data information.

2. The multi-UAV cooperative inspection system of claim 1, wherein, The biological signal multi-modal acquisition and identification module comprises: a sensor configuration unit for receiving information collected by a camera, a microphone, an electrode sensor, a head-mounted sensor, and an infrared sensor, and generating a multi-modal biological information array; a signal processing unit for performing noise reduction and amplification processing on the array information, extracting ECG R wave / ST segment, EEG frequency band energy, and speech feature parameter information; a data fusion unit for identifying biological species, identity, and activity state information based on multi-sensor signal synchronous acquisition; a first sending unit for sending biological species, identity, and activity state information identified by the data fusion unit to the global communication and data processing module.

3. The multi-UAV cooperative inspection system of claim 1, wherein, The environment perception and multi-UAV safety evaluation module comprises: a three-dimensional meteorological field construction unit for obtaining low-altitude area GIS geographic information, multi-source sensing scanning data, and real-time meteorological parameter information of a UAV, constructing a three-dimensional meteorological field through a Kriging spatial interpolation algorithm and a Navier-Stokes equation simplified model in combination with CFD physical simulation, and periodically updating meteorological field information by using an incremental Gaussian process regression algorithm; a safety index calculation unit for establishing a mapping relationship between environment parameters and signal strength through a linear regression model, calculating a stability index in combination with a UAV predicted trajectory and grid residence time, and marking a UAV with a stability index lower than a threshold value as an abnormal object, the stability index calculation formula being as follows: In the formula, m is the number of grid combinations that the UAV passes through, t n is the time length of passing through the nth combination, PSP n is the predicted signal strength percentage of the nth combination, t future is the future prediction time length; a multi-UAV cooperative collision avoidance unit for calculating a safety distance based on real-time speed of an abnormal object, the calculation formula being as follows: In the formula, SP now is the current speed, SP sta is the reference speed, and L is the distance influence coefficient. and simultaneously for predicting multi-UAV three-dimensional trajectories, identifying influence objects intersecting the trajectories, reducing flight speed of a risk object, and generating a new trajectory, iteratively detecting secondary conflicts, and stopping until distances between multi-UAVs at all time points are greater than the safety distance; a second sending unit for sending meteorological field information, stability index information, and safety distance information to the global communication and data processing module.

4. The multi-UAV coordinated inspection system of claim 1, wherein, The bird-like bionic path planning and flight control module comprises: a bionic algorithm core unit for a bird-like bionic algorithm based on particle swarm optimization, simulating the foraging behavior of birds, regarding each UAV as a particle in a bird swarm, dynamically adjusting flight parameters through individual optimal path and global optimal path to obtain adaptive path planning information; a flight control unit based on a model predictive control method, combining a UAV dynamics model and real-time sensor data, for realizing stable flight through PID attitude adjustment; integrating GPS and / or Beidou dual-mode positioning, inertial navigation and visual navigation to ensure positioning accuracy information; an obstacle avoidance unit for detecting obstacles through laser radar and ultrasonic sensors, generating an obstacle avoidance path in combination with the bird-like bionic algorithm; dynamically adjusting flight speed and height in strong wind and rain to simulate the bird's upwind flight skills and obtain obstacle avoidance path information; a third sending unit for sending adaptive path planning information, positioning progress information and obstacle avoidance path information to the global communication and data processing module; a receiving control unit for receiving UAV control information sent by the global communication and data processing module.

5. The multi-UAV coordinated inspection system of claim 1, wherein, The global communication and data processing module comprises: a communication unit for realizing high-speed data backhaul through 4G or 5G in signal good areas and ensuring basic communication through data radio in signal weak areas; a security guarantee unit for protecting data privacy by using an AES encryption algorithm and ensuring data transmission reliability through channel coding and retransmission mechanism; a data receiving unit for receiving data information sent by the first sending unit, the second sending unit, the third sending unit and the task management unit based on a stream data processing framework; a data mining unit for performing real-time data analysis by using K-Means clustering and LSTM prediction after cleaning, deduplication and error correction; a fourth sending unit for sending UAV control information obtained through the above data analysis to the receiving control unit and the ground control and collaborative scheduling module.

6. The multi-UAV coordinated inspection system of claim 1, wherein, The ground control and collaborative scheduling module comprises: a real-time monitoring unit for receiving UAV control information sent by the fourth sending unit and displaying UAV flight state, three-dimensional meteorological field, biological recognition result and abnormal object alarm; a task management unit for sending custom inspection information, dynamically allocating multi-UAV tasks and optimizing resource allocation; a data output unit for generating an inspection report based on information obtained by the real-time monitoring unit.

Images