Unmanned aerial vehicle autonomous obstacle avoidance and path planning system and method

Abstract

The invention discloses an unmanned aerial vehicle autonomous obstacle avoidance and path planning system, which relates to the technical field of unmanned aerial vehicle systems and covers cooperative work of multiple sensors including environment perception, data fusion processing, path planning, flight control, communication, autonomous decision making, power management, a ground control station, system self-inspection and calibration, a load adaptation and optimization module and an environment perception module. The data fusion processing module is used for precisely fusing data; the path planning module generates an optimized path; the flight control module flies stably; the communication module guarantees communication; the autonomous decision-making module makes a decision intelligently; the power management module optimizes a power supply; the ground control station realizes monitoring and operation; a system self-inspection and calibration module ensures reliability; the load adaptation and optimization module adapts to different loads, and all the modules achieve efficient flight of the unmanned aerial vehicle. The flight safety and operation efficiency of the unmanned aerial vehicle are improved, accurate obstacle avoidance and intelligent path planning are achieved, the stability and reliability of the system are enhanced, and wide application and development of the unmanned aerial vehicle technology in multiple fields are promoted.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle systems, and in particular to an unmanned aerial vehicle autonomous obstacle avoidance and path planning system and method. Background Art

[0002] In today's era of rapid technological development, drone technology has been widely used in many fields, such as aerial photography, logistics distribution, agricultural plant protection, surveying and mapping, etc. However, drones face many challenges during flight, among which the lack of autonomous obstacle avoidance and path planning capabilities has become a key factor restricting their further development and widespread application.

[0003] As the application scenarios of drones continue to expand, their flight environments are becoming increasingly complex and diverse. In urban environments, where tall buildings are everywhere and electromagnetic interference is strong, drones need to accurately sense the locations of obstacles such as buildings and utility poles to avoid collision accidents. In the wild, the terrain is complex and changeable, such as mountainous areas and forests. Drones not only have to deal with undulating terrain, but may also encounter severe weather conditions, such as strong winds, heavy rains, and heavy fog, which places higher demands on their obstacle avoidance and path planning capabilities.

[0004] Existing obstacle avoidance technologies for drones mainly rely on a single sensor or a simple combination of sensors, which has obvious limitations. For example, ultrasonic sensors alone have limited detection range and are easily interfered with in complex environments, making it impossible to accurately identify long-distance and small obstacles. Although lidar has high accuracy, its performance will drop significantly in rainy and foggy weather, and it is expensive. Binocular cameras are greatly affected by lighting conditions. In low-light or strong direct light environments, the accuracy of image recognition is reduced, making it difficult to effectively detect obstacles.

[0005] In terms of path planning, traditional algorithms often do not fully consider real-time environmental changes and the performance limitations of the drone itself. Some simple path planning methods cannot adjust the path in time when facing dynamic obstacles or emergencies, causing the drone to fall into a dangerous situation. At the same time, most existing systems lack the ability to adapt to load changes. When the drone is equipped with different mission equipment, the flight performance and endurance change, but the path planning and flight control are not optimized accordingly, affecting the efficiency of mission execution.

[0006] In addition, in multi-machine collaborative operation scenarios, there is a lack of efficient communication and coordination mechanisms between drones, which easily leads to flight conflicts and reduces overall operation efficiency. Moreover, the reliability and stability of current drone systems need to be improved, and system failures or sensor errors may lead to serious consequences. In summary, there is an urgent need for a more advanced and intelligent drone autonomous obstacle avoidance and path planning system and method to meet the increasingly complex application requirements and ensure the safety and efficiency of drone flights. Summary of the invention

[0007] The present invention proposes an autonomous obstacle avoidance and path planning system and method for a UAV to solve the problems mentioned in the above-mentioned prior art.

[0008] In order to achieve the above-mentioned purpose, the present invention adopts the following technical solution: an autonomous obstacle avoidance and path planning system for unmanned aerial vehicles, comprising:

[0009] Environmental perception module: equipped with laser radar, binocular camera, millimeter wave radar and ultrasonic sensor. The laser radar adopts a multi-line rotating structure to construct an accurate three-dimensional point cloud map. The binocular camera obtains environmental depth information, and the ultrasonic sensor detects obstacles. The accuracy formula of the laser radar point cloud data is as follows: Among them C correct is the exact number of point cloud data, C total The total number of point cloud data, requiring PLA ≥ 95%;

[0010] Data fusion processing module: Use Kalman filtering algorithm to fuse lidar and millimeter wave radar data, feature matching algorithm to align binocular camera and lidar data, and integrate ultrasonic sensor data according to timestamp and spatial coordinates. Suppose the data integrity formula after fusion is Where D complete is the number of complete fusion data, D total is the total number of fused data, the expected DFI is ≥ 98%, and dynamic weights are assigned to different sensor data. The weight calculation formula is: Where i represents the sensor type and n is the number of sensors;

[0011] Path planning module: A* algorithm combined with Dijkstra algorithm is used to generate the initial path. The path smoothness formula is: where θ i is the tangent direction of the node on the path, m is the number of path nodes, PS is required to be ≤ 10°, and the flight altitude is planned in combination with the digital elevation model DEM data;

[0012] Flight control module: receives path instructions and uses PID control algorithm to adjust flight attitude. The flight attitude control error formula is: where α j，actual is the jth actual flight attitude angle, α j，target is the jth target flight attitude angle, k is the number of control times, FCE is required to be ≤ 0.5°, and the control algorithm is used to automatically adjust the control parameters according to the changes in the flight environment;

[0013] Communication module: 2.4GHz / 5GHz dual-band communication is adopted, and encrypted communication protocol is used to ensure security. The communication packet loss rate formula is: Where L lost is the number of lost packets, Ltotal The total number of packets sent, requiring CLR ≤ 0.1%;

[0014] Autonomous decision-making module: Based on the environment and path planning results, the risk is assessed through fuzzy logic algorithm. The risk assessment accuracy formula is: Among them A correct is the number of correctly assessed risks, A total is the total number of risk assessments, with an expected RAA ≥ 90%;

[0015] Power management module: real-time monitoring of battery power, voltage and temperature, using a dynamic power allocation strategy, assuming that the battery remaining power estimation error formula is Among them B estimated To estimate the remaining power, B actual The actual remaining power, requiring BRE ≤ 5%, and warning of insufficient power 10-30 minutes in advance;

[0016] Ground control station: including display unit, operation unit and data storage unit. Assume that the data storage integrity formula is Where S complete is the amount of data stored completely, S total For the total amount of data that should be stored, DSI ≥ 99% is expected, and the display unit switches between different viewing angles and modes;

[0017] System self-check and calibration module: Perform system self-check before the drone takes off and during flight, check sensor status, verify control algorithm parameters, monitor communication link quality, and set the system reliability improvement formula as where F after F is the number of normal operation of the system after self-check calibration, before The number of normal operation of the system before self-check calibration, the expected SRI ≥ 15%;

[0018] Load adaptation and optimization module: According to the different mission loads carried by the drone, the system parameters are automatically adjusted, the load weight and center of gravity are calculated, and the flight control parameters are adjusted to adapt to the load changes; the load adaptability scoring formula is set as Where P i，adapted is the actual fitness value of the i-th performance index, P i，ideal is the ideal value of the i-th performance indicator, n is the number of performance indicators, and improves load adaptation capability and operation efficiency.

[0019] Furthermore, the laser radar in the environmental perception module emits a laser wavelength of 905nm or 1550nm. The data fusion processing module removes noise and outliers when fusing data, adopts a median filtering algorithm, and the filter window size is dynamically adjusted according to data fluctuations. When the path planning module plans the path, the no-fly zone data is updated in real time.

[0020] Furthermore, when controlling the flight of the UAV, the flight control module uses a model predictive control algorithm to predict the future flight status. The communication module uses frequency hopping communication technology to improve the communication anti-interference capability in complex electromagnetic environments, and the frequency hopping rate is automatically adjusted according to the interference intensity. The fuzzy logic algorithm membership function in the autonomous decision-making module is adjusted according to environmental characteristics and flight mission requirements.

[0021] Furthermore, the power management module adopts an intelligent charging management system, and the display unit of the ground control station supports virtual reality VR or augmented reality AR display technology.

[0022] Furthermore, the self-check cycle in the system self-check and calibration module is 1-5 minutes before takeoff, and is adjusted according to the flight duration and environmental changes during the flight, and is no longer than 30 minutes. The calibration methods include automatic calibration and manual calibration.

[0023] Furthermore, the load adaptation and optimization module uses a pressure sensor and an inertial measurement unit (IMU) when calculating the load weight and the center of gravity position, and uses an adaptive adjustment algorithm when adjusting the flight control parameters according to the load adaptation.

[0024] Furthermore, the laser radar is located on the top of the drone, the binocular cameras are distributed on the front and sides of the fuselage, the millimeter-wave radar is installed on the bottom and sides of the fuselage, and the ultrasonic sensors are evenly distributed around the fuselage to ensure all-round perception of the environment without blind spots.

[0025] Furthermore, after fusing the data, the data fusion processing module verifies the fusion result in real time, and uses a cross-validation method to compare the fused data with the original sensor data. If the error exceeds a preset threshold, the fusion process is performed again.

[0026] Furthermore, the path planning module considers the real-time performance status of the UAV when generating the path, and dynamically adjusts the parameters of the path planning algorithm according to the performance status.

[0027] Further, the following steps are included:

[0028] Environmental perception steps: Use a variety of sensors to collect information about the surrounding environment, including obstacle location, shape, speed, terrain features, and weather conditions;

[0029] Data fusion step: Transmitting data from different sensors to the data fusion processing module, processing according to the above fusion algorithm and weight adjustment method to obtain a fusion model;

[0030] Path planning steps: Based on the fusion model, the flight path is generated by using the path planning algorithm combined with terrain data;

[0031] Flight control steps: The flight control module receives instructions, controls the flight of the drone according to the control algorithm, and adjusts the attitude and speed in real time;

[0032] Autonomous decision-making steps: The autonomous decision-making module monitors in real time and handles any anomalies according to risk assessment and decision-making algorithms;

[0033] Communication and interaction steps: The communication module implements data transmission and command interaction to meet the communication packet loss rate and delay requirements;

[0034] Power management steps: The power management module continuously monitors the power supply, allocates power according to the strategy, estimates the power consumption and issues warnings;

[0035] Data recording and analysis steps: The ground control station records the data and stores it according to the data storage integrity requirements for evaluation and optimization;

[0036] System self-check and calibration steps: Perform self-check and calibration before takeoff and in flight to check sensors, algorithm parameters and communication links, and improve system stability according to system reliability improvement requirements;

[0037] Load adaptation and optimization steps: adjust system parameters according to the load, calculate load characteristics and optimize flight control and path planning, and improve operational efficiency according to load adaptability scores.

[0038] Compared with the prior art, the present invention has the following beneficial effects:

[0039] First, the all-round and accurate environmental perception capability greatly improves flight safety. Multiple sensors work together, and through optimized data fusion algorithms, they can accurately identify various obstacles in complex environments, effectively avoid collision accidents, ensure the safety of personnel and property, and reduce the risk of loss.

[0040] Secondly, efficient and intelligent path planning significantly improves operational efficiency. The smooth and optimized path generated by comprehensively considering multiple factors reduces flight energy consumption and time costs, allowing drones to reach their destination faster and more stably, especially in complex terrain and multi-task scenarios.

[0041] Furthermore, adaptive flight control and autonomous decision-making enhance system flexibility. The drone can adjust its flight attitude and strategy in real time according to environmental changes and mission requirements, respond to emergencies more calmly, improve mission success rate, and reduce the need for manual intervention.

[0042] In addition, stable and reliable communication and power management ensure stable operation of the system. Low packet loss rate and low latency communication ensure timely transmission of instructions, and intelligent power management strategy extends flight time and provides power warning to avoid flight accidents caused by power exhaustion.

[0043] Finally, the powerful system self-check and calibration and load adaptation functions improve the overall system performance. Timely discover and solve system problems, optimize flight parameters according to load, improve the adaptability and reliability of drones in different mission scenarios, and expand their application range. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 A schematic block diagram of an autonomous obstacle avoidance and path planning system for a UAV proposed by the present invention;

[0045] Figure 2 This is a schematic block diagram of an autonomous obstacle avoidance and path planning method for a UAV proposed by the present invention. DETAILED DESCRIPTION

[0046] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0047] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the present invention.

[0048] In addition, the terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined. In addition, the terms "installed", "connected" and "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, and it can be the internal connection of two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances. The present invention will be further described in detail below in conjunction with the accompanying drawings.

[0049] Reference Figure 1-2 : An autonomous obstacle avoidance and path planning system for unmanned aerial vehicles, comprising:

[0050] Environmental perception module: equipped with laser radar, binocular camera, millimeter wave radar and ultrasonic sensor. The laser radar adopts a multi-line rotating structure with 16-64 lines, a detection range of 0.1-150 meters, an angular resolution of 0.1°-1°, and an update frequency of 10-20Hz, which is used to construct accurate three-dimensional point cloud maps; the binocular camera has a horizontal field of view of 80°-120° and a vertical field of view of 50°-70° to obtain environmental depth information; the millimeter wave radar has an effective detection distance of 5-100 meters and can penetrate rain, fog, sand and dust; the ultrasonic sensor has a close-range accuracy of ±0.01-±0.1 meters and can detect close-range obstacles. Assume that the accuracy formula of the laser radar point cloud data is Among them C correct is the exact number of point cloud data, C total The total number of point cloud data is required, and PLA ≥ 95% is required to ensure the accuracy of environmental perception.

[0051] Data fusion processing module: It mainly uses advanced Kalman filtering algorithm to skillfully fuse the data from LiDAR and millimeter-wave radar. LiDAR can measure the distance and shape of surrounding objects with extremely high accuracy, while millimeter-wave radar has strong penetration and stability in adverse weather conditions. The fusion of the two data can complement each other and provide more comprehensive and accurate environmental information.

[0052] In order to further improve the accuracy and consistency of the data, the module also uses a feature matching algorithm to align the data of the binocular camera and the lidar. The binocular camera can obtain rich visual information. Through feature matching, the image features captured by the camera can be accurately matched with the data scanned by the lidar, thereby constructing a more three-dimensional and realistic environment model.

[0053] In addition, the data fusion processing module also cleverly integrates the data of the ultrasonic sensor based on the timestamp and spatial coordinates. The ultrasonic sensor can quickly detect obstacles at close range and provide timely feedback when the drone approaches an object. The time sequence of the data is ensured by the timestamp, and the position relationship of the data is determined by the spatial coordinates, so that the data of different sensors can be fused in a unified space-time framework. Suppose the data integrity formula after fusion is Where D complete is the number of complete fusion data, D total is the total number of fused data, and the expected DFI is ≥ 98%, so as to obtain a comprehensive and accurate environmental model. Dynamic weights are assigned to different sensor data, and the weights are adjusted in real time according to the sensor reliability (value 0-1), environmental condition influencing factor E (value 0-1) and data quality score Q (value 0-100). The weight calculation formula is: Among them, i represents the sensor type, n represents the number of sensors, and the sensor reliability reflects the performance and stability of the sensor itself; the environmental condition influencing factor takes into account the impact of the current flight environment on the accuracy of sensor data. For example, in harsh environments such as strong light, rain and snow, the data of some sensors may be greatly disturbed, and their influencing factors will be adjusted accordingly; the data quality score is a comprehensive evaluation of the quality of real-time data collected by the sensor, including data accuracy, completeness, consistency, etc. Through such a dynamic weight allocation mechanism, the data of each sensor can be used more reasonably, their advantages can be fully utilized, and data errors can be effectively reduced, thereby greatly improving the accuracy of fused data and providing solid data support for the safe flight and efficient operation of drones in complex environments.

[0054] Path planning module: First, this module uses a combination of the advanced A* algorithm and the classic Dijkstra algorithm to generate the initial path. The A* algorithm is an efficient path search algorithm that can quickly find a better path from the starting point to the target point by comprehensively considering the actual cost and estimated cost of the path; while the Dijkstra algorithm uses its rigorous calculation method to ensure that the shortest path between two points can be found. The powerful combination of these two algorithms makes the generated initial path both more efficient and can guarantee the optimality of the path to a certain extent.

[0055] Next, in order to make the drone's flight smoother, the path planning module uses the spline curve interpolation algorithm to smooth and optimize the initial path. The spline curve interpolation algorithm can generate a continuous and smooth curve based on the given path nodes, thereby greatly reducing the sharp turns in the path, allowing the drone to change direction more smoothly during flight, reducing flight energy consumption, and improving flight stability and comfort. Suppose the path smoothness formula is where θ iis the tangent direction of the node on the path, m is the number of path nodes, PS is required to be ≤ 10°, and the path turns are reduced. Combined with the digital elevation model (DEM) data, the resolution is 1-10 meters. Through the analysis and processing of these data, the path planning module can plan a reasonable flight altitude for the drone, so that it can cleverly avoid terrain obstacles such as mountains, tall buildings, trees, etc., to ensure the safety of the drone during flight. For example, when the drone needs to cross a mountainous area, the path planning module will set the flight altitude to a position higher than the mountain and to ensure a safe distance based on the elevation information of the mountainous area in the DEM data. At the same time, combined with the requirements of path smoothness, a safe and smooth flight path is generated, so that the drone can successfully complete the task. In this way, the path planning module creates a safe, efficient and smooth flight path for the drone through a series of advanced algorithms and precise data processing, so that it can fly freely in various complex environments and complete various tasks.

[0056] Flight control module: receives path instructions, uses PID control algorithm to adjust flight attitude, control accuracy ±0.1°-±1°, and adjusts motor speed to control speed (0-20 m / s).

[0057] Assume that the flight attitude control error formula is where α j，actual is the jth actual flight attitude angle, α j，target is the jth target flight attitude angle, k is the control number, and FCE≤0.5° is required to ensure stable flight. Adaptive control algorithm is used to automatically adjust control parameters according to flight environment changes to enhance anti-interference ability.

[0058] Communication module: 2.4GHz / 5GHz dual-band communication is adopted, the communication distance is 5-10 kilometers, the transmission rate is not less than 100Mbps, and the encryption communication protocol is used to ensure security. Assume that the communication packet loss rate formula is Where L lost is the number of lost packets, L total The total number of packets sent requires CLR ≤ 0.1%. Support 4G / 5G network communication, with a transmission delay of less than 100 milliseconds to ensure the timeliness of control instructions.

[0059] Autonomous decision-making module: Based on the environment and path planning results, the risk is assessed through fuzzy logic algorithm and the decision is made within 10-100 milliseconds. Suppose the risk assessment accuracy formula is Among them A correct is the number of correctly assessed risks, A total The total number of risk assessments, with an expected RAA ≥ 90%. Deep reinforcement learning algorithm training is used to improve decision-making accuracy and efficiency.

[0060] Power management module: The power management module has the function of real-time monitoring of battery power, and its monitoring accuracy can reach ±1%-±5%, which means that it can grasp the real-time changes of battery power extremely accurately. At the same time, it also monitors the voltage and temperature of the battery in real time, because the stability of the voltage directly affects the power output of the drone, and the temperature is an important indicator of the battery's health and safety performance. Once the voltage fluctuates abnormally or the temperature exceeds the normal range, the power management module will respond quickly and take corresponding protective measures to ensure the safety of the battery and the drone.

[0061] This module uses an advanced dynamic power allocation strategy, which can intelligently allocate battery power according to the real-time power requirements of each system and module of the drone to achieve the best energy utilization efficiency. For example, when the flight control system of the drone needs more power to cope with complex flight environments, the power management module will prioritize allocating power to it to ensure the accuracy and stability of flight control; and when the communication module needs to send important data, it will also receive sufficient power support in time to ensure smooth communication. Assume that the battery remaining power estimation error formula is Among them B estimated To estimate the remaining power, B actual The actual remaining power, BRE ≤ 5%. Support hot swap of batteries, shortening the mission preparation time. It will warn of low power 10-30 minutes in advance. When the battery is about to run out, it will send out a warning signal in time to remind the operator or ground control station to take corresponding measures, such as adjusting the flight path, finding a suitable landing place, etc., so as to ensure the flight safety of the drone and avoid accidents caused by running out of power.

[0062] Ground control station: includes display unit, operation unit and data storage unit. The display unit displays the flight status in a three-dimensional map, the operation unit implements manual control and parameter setting, and the data storage unit stores data (with a capacity of no less than 1TB). Suppose the data storage integrity formula is Where S complete is the amount of data stored completely, S total The total amount of data to be stored is expected to be DSI ≥ 99%, which is convenient for data analysis and troubleshooting. The display unit can switch between different viewing angles and modes to facilitate operators to understand the flight situation.

[0063] System self-check and calibration module: Perform system self-checks regularly before the drone takes off and during flight. Check the sensor status. If the laser radar transmission power is lower than the preset threshold (such as 80% of the normal power), an alarm is issued; verify the control algorithm parameters. If the PID control parameters are out of the normal range, automatic calibration is performed; monitor the quality of the communication link. If the signal strength is lower than the set value, try to switch the frequency band or enhance the signal. Through self-check and calibration, the system reliability and stability are improved. The system reliability improvement formula is: where F after F is the number of normal operation of the system after self-check calibration, before It is the number of times the system operates normally before self-check calibration. The expected SRI is ≥ 15%.

[0064] Load adaptation and optimization module: Automatically adjust system parameters according to the different mission loads carried by the drone (such as photographic equipment, cargo transport devices, etc.). Calculate the load weight and center of gravity position, adjust the flight control parameters to adapt to the load changes; optimize path planning, and select a more suitable flight path based on the impact of the load on endurance and maneuverability. Suppose the load adaptability scoring formula is Where P i，adapted is the actual fitness value of the i-th performance index, P i，ideal is the ideal value of the i-th performance indicator, n is the number of performance indicators, and improves load adaptation capability and operation efficiency.

[0065] In the present invention, the laser radar of the environment perception module plays a vital role. Among them, the laser wavelength emitted by the laser radar is 905nm or 1550nm. The selection of these two specific wavelengths is not random, but carefully considered. The laser radar with a wavelength of 905nm has the advantages of relatively low cost and mature technology. In some scenarios where the detection distance and accuracy requirements are not particularly stringent, it can effectively detect objects in the surrounding environment and provide basic environmental perception information for the drone; while the laser radar with a wavelength of 1550nm has stronger penetration and anti-interference capabilities, especially in complex environments, such as smoke, dust, etc., it can better penetrate these interference factors, improve the detection effect of objects in different environments, ensure that the drone can clearly and accurately perceive the surrounding environment, and provide a reliable basis for subsequent flight decisions.

[0066] The data fusion processing module also undertakes key tasks in the entire system. When it fuses the data from various sensors, it will first preprocess the data, which is a crucial step. Because in the actual environment, the data collected by the sensor is often interfered by various factors, resulting in noise and outliers. If these noises and outliers are not processed, they will seriously affect the subsequent data fusion quality and the flight decision of the drone. Therefore, this module uses the median filtering algorithm to remove these noises and outliers. The median filtering algorithm has a good filtering effect. It can effectively smooth the data and remove noise interference while retaining the important features of the data. Moreover, in order to further improve the filtering effect, the filter window size will be dynamically adjusted according to the data fluctuation. For example, when the data fluctuates greatly, it means that the environment may be more complex. At this time, the filter window will be appropriately increased to better capture the overall trend of the data and remove outliers; when the data fluctuates less, the window will be reduced accordingly to improve the real-time and accuracy of data processing. Through such a preprocessing method, the data quality is greatly improved, laying a solid foundation for subsequent data fusion and precise flight of drones.

[0067] When planning the flight path of the drone, the path planning module fully considers the important factor of no-fly zone information. No-fly zones refer to areas where drones are prohibited from flying due to security, privacy, military and other reasons. The information of these areas is crucial to ensure the safety and compliance of drone flights. In order to ensure the accuracy and timeliness of no-fly zone information, no-fly zone data will be updated in real time, and its sources are also multi-channel, including official releases, geographic information systems and real-time monitoring. The official no-fly zone information is authoritative and accurate, and is one of the important sources of no-fly zone data; the geographic information system can provide detailed geospatial information to help accurately define the scope and location of the no-fly zone; real-time monitoring can promptly discover some temporary no-fly situations, such as sudden military activity areas, large-scale activity airspace, etc. By integrating these multi-channel information sources and updating the no-fly zone data in real time, the path planning module can plan an efficient, safe and compliant flight path for the drone, avoid the drone from mistakenly entering the no-fly zone, and ensure the smooth progress of the flight mission and flight safety.

[0068] In the present invention, when controlling the flight of the UAV, the flight control module adopts a model predictive control algorithm to predict the future flight state, adjust the control input in advance, and enhance the flight stability and tracking accuracy; the communication module adopts frequency hopping communication technology to improve the communication anti-interference ability in a complex electromagnetic environment, and the frequency hopping rate is automatically adjusted according to the interference intensity, ranging from 100 to 1000 times per second; the fuzzy logic algorithm membership function in the autonomous decision-making module is dynamically adjusted according to environmental characteristics and flight mission requirements, thereby improving the flexibility and accuracy of risk assessment.

[0069] In the present invention, the power management module innovatively adopts an intelligent charging management system, which can accurately and automatically select the best charging mode according to the type and real-time status of the battery. In terms of battery type, different types of batteries (such as lithium batteries, nickel-metal hydride batteries, etc.) have different electrochemical characteristics and charging requirements. The intelligent charging management system can accurately identify the battery type. For example, for lithium batteries, it is well aware of their sensitivity to charging voltage and current. According to the characteristics of lithium batteries, it will adopt a constant current charging mode in the initial charging stage to quickly inject energy into the battery with a relatively stable current to improve charging efficiency; when the battery voltage is close to full, the system will automatically switch to a constant voltage charging mode to keep the charging voltage constant, so that the battery can be fully charged safely and smoothly, avoiding damage to the battery caused by overcharging, thereby effectively extending the battery life.

[0070] As for the battery status, the intelligent charging management system also always maintains a high level of attention and accurate judgment. When the battery is in different health states (such as new batteries, batteries that have been used for a period of time, aging batteries, etc.), its charging requirements are also different. For example, for new batteries, the system may adopt a milder charging strategy to gradually activate the battery performance; and for batteries whose performance has declined after a period of use, the system will dynamically adjust the charging mode according to changes in parameters such as its internal resistance, and may increase the frequency and intensity of pulse charging, and use pulse current to alleviate the polarization phenomenon inside the battery, restore some of the battery's performance, and further extend the battery's service life.

[0071] At the same time, the display unit of the ground control station has also undergone a major upgrade. It supports advanced virtual reality (VR) or augmented reality (AR) display technology, bringing operators unprecedented operating experience and improved control accuracy.

[0072] When operators use a display unit that supports VR technology, they feel as if they are in the perspective of the drone and can get an immersive monitoring experience. After they put on the VR device, they see a 360-degree panoramic view of the drone, as if they are soaring in the air and feeling the drone's environment in real time. This immersive feeling allows operators to be more sensitive to surrounding obstacles, terrain changes and other information, so as to make more accurate flight control decisions. For example, when performing tasks in a complex urban environment, through VR display, operators can clearly see potential dangers such as narrow passages and wires between high-rise buildings, and adjust the drone's flight path in advance to avoid collision accidents.

[0073] The display unit that supports AR technology perfectly integrates virtual information with real images. When the operator monitors the flight of the drone, various useful virtual information will be superimposed on the real picture in front of him in real time, such as the real-time flight parameters of the drone (speed, altitude, heading, etc.), detailed data of the surrounding environment (wind speed, wind direction, no-fly zone prompts, etc.) and task-related prompts. These virtual information are presented to the operator in an intuitive and easy-to-understand way, without the need for them to frequently switch their sights to view different dashboards or display screens, greatly improving the convenience of operation and control accuracy. For example, when the drone approaches the boundary of the no-fly zone, the AR display unit will immediately highlight the boundary line and warning sign of the no-fly zone in the picture, reminding the operator to adjust the flight direction in time to ensure flight safety and compliance.

[0074] Through the intelligent charging management system of the power management module and the advanced display technology of the ground control station display unit, the present invention provides all-round guarantee and support for the efficient and safe operation of the UAV, greatly improving the overall performance and user experience of the UAV system.

[0075] In the present invention, regarding the self-check cycle, before the drone takes off, the self-check cycle of the system self-check and calibration module is set between 1 and 5 minutes. The setting of this time period is carefully considered to ensure that the drone can conduct a comprehensive and detailed inspection of various key systems and components before taking off, and not miss any potential problems that may affect flight safety. For example, in these short minutes, the self-check system will quickly detect the flight control system of the drone, check whether the working status of each sensor (such as gyroscope, accelerometer, barometer, etc.) is normal, and ensure that they can accurately perceive the attitude, speed, height and other information of the drone; at the same time, the power system of the drone will also be tested, including motors, electric regulators, batteries, etc., to check whether the motor runs smoothly, whether the control signal of the electric regulator is accurate, and whether the battery power and voltage meet the take-off requirements. Through this series of strict self-checks before take-off, it is like giving the drone a comprehensive "physical examination". Only when all the test items are qualified will the drone be allowed to take off, thus laying a solid foundation for flight safety.

[0076] During the flight, the self-check cycle will be dynamically adjusted according to the flight duration and environmental changes, and the longest time will not exceed 30 minutes. This is because as the flight time goes by, the various components of the drone may experience performance changes or potential failures due to long-term work, and changes in the flight environment (such as changes in temperature, humidity, air pressure, and electromagnetic interference) may also affect the drone system. For example, after flying for a period of time in a high-temperature environment, the electronic components of the drone may overheat, which may affect its performance and stability. At this time, shortening the self-check cycle can detect these problems more promptly; for example, when the drone enters an area with strong electromagnetic interference from an area with a relatively stable electromagnetic environment, the flight control system may be disturbed, resulting in deviations in the flight attitude. By dynamically adjusting the self-check cycle, these abnormal situations can be detected more quickly and timely measures can be taken. This method of flexibly adjusting the self-check cycle according to actual conditions fully reflects the intelligence and adaptability of the system, and maximizes the safety and reliability of the drone during flight.

[0077] In addition, calibration is also an important part of the system self-check and calibration module, which includes automatic calibration and manual calibration. Automatic calibration is the system automatically calibrating the relevant parameters and systems of the drone according to preset algorithms and procedures. For example, when the compass of the drone is deviated by the interference of the surrounding magnetic field, the automatic calibration function will use the built-in calibration algorithm of the drone and combine the data during the flight to automatically calibrate the compass so that it can accurately indicate the direction and ensure that the flight direction of the drone is accurately controlled. Manual calibration provides operators with more flexibility and control, and it can be remotely operated by the ground control station, which is very useful in some special situations. For example, when a drone is performing a long-distance mission, the ground operator finds through the monitoring screen of the ground control station that the shooting angle of the drone's camera is deviated, affecting the execution of the mission. At this time, the operator can remotely send calibration instructions through the operation interface of the ground control station to manually calibrate the camera angle to restore it to normal working condition. In addition, manual calibration can also be performed after the drone lands. For example, when the drone completes a flight mission, ground maintenance personnel can conduct a comprehensive inspection of the drone. If it is found that certain components (such as the installation position of the mechanical structure, etc.) need to be calibrated, manual calibration can be performed on the ground to ensure that the drone is in the best condition before the next flight.

[0078] Through this perfect self-checking and calibration mechanism, the UAV system of the present invention can always maintain a good operating state, effectively deal with various complex situations and potential risks, and provide a strong guarantee for the safe flight and efficient operation of the UAV.

[0079] In the present invention, two advanced measuring devices, high-precision pressure sensors and inertial measurement units (IMUs), are used in the calculation of load weight and center of gravity position. High-precision pressure sensors are like sensitive "weight sensors" that can extremely accurately sense the changes in pressure exerted by the load on the drone, thereby accurately calculating the weight of the load. Its weight measurement accuracy can reach ±0.1kg, which means that both small load changes and large weight differences can be accurately measured. For example, when a drone needs to carry goods of different weights for delivery, even if the weight of the goods differs by only a few dozen grams, the sensor can accurately capture this change and provide accurate weight data for subsequent flight control.

[0080] The inertial measurement unit (IMU) is like the "balance perception center" of the drone. It can measure the acceleration, angular velocity and other information of the drone in three-dimensional space in real time, and then accurately calculate the center of gravity of the load. The center of gravity position measurement accuracy can reach ±0.05m. Such high accuracy enables the drone to clearly know the center of gravity distribution of the load. For example, when the drone is equipped with an irregularly shaped or offset center of gravity device, the IMU can accurately determine the center of gravity position, which is crucial to maintaining the flight balance and stability of the drone.

[0081] Moreover, when the flight control parameters are adjusted according to the load adaptation, the load adaptation and optimization module adopts an adaptive adjustment algorithm, which is an intelligent and efficient algorithm. It can quickly and automatically adjust the flight control parameters such as motor speed, rudder angle, flight attitude, etc. according to the real-time changes in the load to ensure that the drone can maintain good flight performance under different loads. Moreover, this adaptive adjustment is extremely fast, and the adjustment time does not exceed 1 second. For example, when the drone suddenly unloads part of the load during flight, the adaptive adjustment algorithm will instantly sense this change and immediately adjust the output power of the motor and the control angle of the rudder, so that the drone can quickly adapt to the new load state and avoid problems such as loss of control of flight attitude or reduced flight efficiency due to load changes.

[0082] By adopting high-precision measurement equipment and advanced adaptive adjustment algorithms, the load adaptation and optimization module provides strong support for the UAV's load management and flight control, greatly improving the adaptability and reliability of the UAV in various complex mission scenarios, and ensuring that the UAV can complete various tasks safely, stably and efficiently.

[0083] In the present invention, the sensor installation position in the environmental perception module is optimized and designed, and the laser radar is cleverly placed on the top of the drone. As a high-precision ranging sensor, the laser radar can quickly scan the surrounding environment in the form of a laser beam, generate detailed three-dimensional point cloud data, and provide accurate distance and space information for the drone. Installing it on the top is like putting a "smart hat" on the drone, allowing it to detect the environmental conditions above and around it at a long distance from a high position. For example, when a drone passes through woods or urban areas with tall buildings, the laser radar on the top can detect obstacles such as branches, wires, and building tops above in advance, providing an important basis for the drone to plan a safe flight path and avoid collision accidents.

[0084] The binocular cameras are distributed on the front and sides of the fuselage, like the "visual sentinels" of the drone. The binocular cameras on the front are like the "forward-looking eyes" of the drone, which can clearly capture the scene in front, provide accurate visual information for the direction of the drone's advance, and help the drone identify the road ahead, obstacles, target objects, etc. The binocular cameras on both sides are like the "side-view guards" of the drone, which can monitor the environmental dynamics on both sides of the drone in real time, whether it is birds flying from the side, sudden protrusions on the side of the building, or changes in the terrain on both sides, they can be detected in time. This distribution method enables the drone to obtain rich visual information from multiple angles during flight, providing comprehensive support for flight decisions.

[0085] The millimeter-wave radar is installed at the bottom and side of the fuselage, and it is like the "invisible guardian" of the drone. The millimeter-wave radar has strong penetration and anti-interference capabilities, and can work normally even in adverse weather conditions (such as fog, heavy rain, etc.). The millimeter-wave radar installed at the bottom of the fuselage can detect the ground conditions under the drone, such as the ups and downs of the ground, the height of obstacles, etc. This is especially important for drones flying or landing at low altitudes. It can ensure that the drone maintains a safe distance from the ground and avoid accidents caused by ground obstacles or terrain changes. The millimeter-wave radar installed on the side can work in conjunction with the binocular camera on the side to further enhance the drone's perception of the side environment, especially in detecting long-distance moving objects (such as other aircraft, vehicles, etc.). It plays an important role and provides timely information for the drone's avoidance and flight path adjustment.

[0086] Ultrasonic sensors are evenly distributed around the fuselage, like the drone's "bodyguard". Ultrasonic sensors measure distance by emitting and receiving ultrasonic waves, and have the advantages of low cost and fast response. They are evenly distributed around the fuselage and can monitor the close-range environment around the drone in real time. Especially when the drone is close to objects such as buildings, trees, and people, the ultrasonic sensors can quickly sense the distance changes to these objects and issue an alarm in time, allowing the drone to respond quickly, such as adjusting the flight attitude, slowing down or hovering, etc., to ensure the safety of the drone in a close-range environment.

[0087] Through this carefully optimized sensor installation position, each sensor in the environmental perception module performs its own function and cooperates with each other to form a strict environmental perception network, ensuring that the UAV can perceive the surrounding environment in an all-round and no-dead-angle manner during flight, providing a solid guarantee for the safe flight and efficient operation of the UAV.

[0088] In the present invention, when the data fusion processing module completes the data fusion, it does not end there, but immediately verifies the fusion result in real time. This step is like a strict quality control process to ensure that the fused data is accurate and reliable. Specifically, the cross-validation method is adopted, which is a scientific and rigorous data verification method.

[0089] During the cross-validation process, the fused data is carefully compared with the raw sensor data. Raw sensor data is the information initially collected by the drone's sensors without fusion processing, and each has unique advantages and limitations. For example, LiDAR can accurately measure the distance and three-dimensional shape of objects, but it may be affected in certain weather conditions; cameras can provide rich visual information, but have certain requirements for light and distance; and inertial measurement units (IMUs) can perceive information such as the drone's attitude and acceleration in real time, but long-term use may cause drift errors.

[0090] By comparing the fused data with the original sensor data, the accuracy and reliability of the fusion results can be fully evaluated. If the error between the fused data and the original sensor data exceeds the preset threshold (e.g. 5%), this indicates that there may be problems in the fusion process, such as unreasonable weight distribution of data from certain sensors during fusion, or deviations in the data fusion algorithm during processing. At this point, the data fusion processing module will not hesitate to re-process the fusion to ensure that the data ultimately provided to the UAV flight control system is high-quality and high-precision, providing a solid and reliable basis for the UAV's flight decision.

[0091] This real-time verification and strict control mechanism of data fusion results fully reflects the rigor and professionalism of the present invention in UAV data processing, greatly improves the overall performance and reliability of the UAV system, enables it to operate stably in various complex environments, and excellently complete various tasks.

[0092] In the present invention, the path planning module fully considers the real-time performance status of the UAV when generating the path, which is crucial to ensure the smooth flight of the UAV.

[0093] First of all, the remaining power is one of the most critical factors in path planning. The battery power of a drone is like the "energy source" for its flight. The amount of power directly affects the drone's endurance and flight range. For example, when the remaining power of a drone is high, the path planning module may plan a relatively long and complex path so that the drone can complete its tasks more efficiently, such as covering a larger area for aerial photography or environmental monitoring; when the remaining power is low, the module will quickly adjust the strategy and prioritize planning a path that can quickly return to the take-off point or the nearest charging station to ensure that the drone will not crash due to exhaustion. In this process, the path planning algorithm will accurately calculate the drone's range at different flight speeds based on real-time power data, so as to reasonably plan each node and flight distance on the path to avoid insufficient power.

[0094] Flight speed is also an important performance state that affects path planning. Different mission scenarios have different requirements for the flight speed of drones, and the flight speed is closely related to factors such as the energy consumption and maneuverability of drones. For example, when performing express delivery tasks, in order to improve delivery efficiency, drones may need to fly at a higher speed, but too high a speed may increase energy consumption, shorten the flight time, and in the event of emergencies, maneuverability will be affected to a certain extent. At this time, the path planning module will comprehensively consider these factors and dynamically adjust the curvature and node spacing of the path according to the real-time flight speed of the drone. If the current flight speed of the drone is fast, the path planning algorithm will appropriately increase the turning radius of the path to ensure that the drone can turn safely and smoothly when flying at high speed, avoiding the risk of loss of control due to sharp turns; at the same time, it will also reasonably arrange the stop points or deceleration areas on the path so that the drone can appropriately decelerate when necessary to save power or cope with complex environments.

[0095] Maneuverability is about the ability of a drone to flexibly change its flight attitude and direction during flight. The maneuverability of a drone is affected by many factors, such as its own structure and power system. For example, some small drones have high maneuverability and can flexibly shuttle in a small space, while large drones have relatively poor maneuverability. The path planning module will plan a path suitable for its performance characteristics based on the real-time maneuverability data of the drone. For drones with good maneuverability, when encountering obstacles or needing to avoid no-fly zones, the path planning algorithm can generate more complex and flexible paths to give full play to its maneuverability advantages; for drones with poor maneuverability, it will try to plan a smoother path with fewer turns to reduce the difficulty and risk of flight.

[0096] In short, the path planning module monitors the drone's remaining power, flight speed, maneuverability and other performance status in real time, and dynamically adjusts the parameters of the path planning algorithm according to these status, such as the path length, curvature, node position, etc., to ensure that the planned path is not only highly feasible and can give full play to the performance advantages of the drone, but also can maximize the safety of flight, so that the drone can complete the flight mission smoothly and efficiently in various complex environments and mission requirements.

[0097] The present invention comprises the following steps:

[0098] Environmental perception steps: Use a variety of sensors to collect information about the surrounding environment, including obstacle location, shape, speed, terrain features, and weather conditions.

[0099] Data fusion step: Transmit the data from different sensors to the data fusion processing module, process them according to the above fusion algorithm and weight adjustment method, and obtain the fusion model.

[0100] Path planning steps: Based on the fusion model, the path planning algorithm is used in combination with terrain data to generate the optimal flight path, which is optimized according to the path smoothness requirements.

[0101] Flight control steps: The flight control module receives instructions, controls the flight of the drone according to the control algorithm, and adjusts the attitude and speed in real time to meet the attitude control error requirements.

[0102] Autonomous decision-making steps: The autonomous decision-making module monitors in real time and handles any abnormalities according to risk assessment and decision-making algorithms to ensure flight safety.

[0103] Communication and interaction steps: The communication module realizes data transmission and command interaction to meet the communication packet loss rate and delay requirements.

[0104] Power management steps: The power management module continuously monitors the power supply, allocates power according to the strategy, estimates the power consumption and issues warnings.

[0105] Data recording and analysis steps: The ground control station records the data and stores it according to the data storage integrity requirements for evaluation and optimization.

[0106] System self-check and calibration steps: Perform self-check and calibration regularly before takeoff and during flight to check sensors, algorithm parameters and communication links, and improve system stability according to system reliability improvement requirements.

[0107] Load adaptation and optimization steps: adjust system parameters according to the load, calculate load characteristics and optimize flight control and path planning, and improve operational efficiency according to load adaptability scores.

[0108] The above are only preferred specific implementation modes of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical solutions and inventive concepts of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. An autonomous obstacle avoidance and path planning system for unmanned aerial vehicles, characterized in that: include: Environmental perception module: equipped with laser radar, binocular camera, millimeter wave radar and ultrasonic sensor. The laser radar adopts a multi-line rotating structure to construct an accurate three-dimensional point cloud map. The binocular camera obtains environmental depth information, and the ultrasonic sensor detects obstacles. The accuracy formula of the laser radar point cloud data is as follows: Among them C correct is the exact number of point cloud data, C total The total number of point cloud data, requiring PLA ≥ 95%; Data fusion processing module: Use Kalman filtering algorithm to fuse lidar and millimeter wave radar data, feature matching algorithm to align binocular camera and lidar data, and integrate ultrasonic sensor data according to timestamp and spatial coordinates. Suppose the data integrity formula after fusion is Where D complete is the number of complete fusion data, D total is the total number of fused data, the expected DFI is ≥ 98%, and dynamic weights are assigned to different sensor data. The weight calculation formula is: Where i represents the sensor type and n is the number of sensors; Path planning module: A* algorithm combined with Dijkstra algorithm is used to generate the initial path. The path smoothness formula is: where θ i is the tangent direction of the node on the path, m is the number of path nodes, PS is required to be ≤ 10°, and the flight altitude is planned in combination with the digital elevation model DEM data; Flight control module: receives path instructions and uses PID control algorithm to adjust flight attitude. The flight attitude control error formula is: where α j，actual is the jth actual flight attitude angle, α j，target is the jth target flight attitude angle, k is the number of control times, FCE is required to be ≤ 0.5°, and the control algorithm is used to automatically adjust the control parameters according to the changes in the flight environment; Communication module: 2.4GHz / 5GHz dual-band communication is adopted, and encrypted communication protocol is used to ensure security. The communication packet loss rate formula is: Where L lost is the number of lost packets, L total The total number of packets sent, requiring CLR ≤ 0.1%; Autonomous decision-making module: Based on the environment and path planning results, the risk is assessed through fuzzy logic algorithm. The risk assessment accuracy formula is: Among them A correct is the number of correctly assessed risks, A total is the total number of risk assessments, with an expected RAA ≥ 90%; Power management module: real-time monitoring of battery power, voltage and temperature, using a dynamic power allocation strategy, assuming that the battery remaining power estimation error formula is Among them B estimated To estimate the remaining power, B actual The actual remaining power, requiring BRE ≤ 5%, and warning of insufficient power 10-30 minutes in advance; Ground control station: including display unit, operation unit and data storage unit. Assume that the data storage integrity formula is Where S complete is the amount of data stored completely, S total For the total amount of data that should be stored, DSI ≥ 99% is expected, and the display unit switches between different viewing angles and modes; System self-check and calibration module: Perform system self-check before the drone takes off and during flight, check sensor status, verify control algorithm parameters, monitor communication link quality, and set the system reliability improvement formula as where F after F is the number of normal operation of the system after self-check calibration, before The number of normal operation of the system before self-check calibration, the expected SRI ≥ 15%; Load adaptation and optimization module: According to the different mission loads carried by the drone, the system parameters are automatically adjusted, the load weight and center of gravity are calculated, and the flight control parameters are adjusted to adapt to the load changes; the load adaptability scoring formula is set as Where P i，adapted is the actual fitness value of the i-th performance index, P i，ideal is the ideal value of the i-th performance indicator, n is the number of performance indicators, and improves load adaptation capability and operation efficiency.

2. The autonomous obstacle avoidance and path planning system for unmanned aerial vehicles according to claim 1, characterized in that: The laser radar in the environmental perception module emits a laser wavelength of 905nm or 1550nm. The data fusion processing module removes noise and outliers when fusing data. It uses a median filtering algorithm and dynamically adjusts the filter window size based on data fluctuations. When the path planning module plans the path, the no-fly zone data is updated in real time.

3. The autonomous obstacle avoidance and path planning system for unmanned aerial vehicles according to claim 1, characterized in that: When controlling the flight of the UAV, the flight control module uses a model predictive control algorithm to predict the future flight status. The communication module uses frequency hopping communication technology to improve the communication anti-interference capability in complex electromagnetic environments, and the frequency hopping rate is automatically adjusted according to the interference intensity. The fuzzy logic algorithm membership function in the autonomous decision-making module is adjusted according to environmental characteristics and flight mission requirements.

4. The autonomous obstacle avoidance and path planning system for unmanned aerial vehicles according to claim 1, characterized in that: The power management module adopts an intelligent charging management system, and the display unit of the ground control station supports virtual reality VR or augmented reality AR display technology.

5. The autonomous obstacle avoidance and path planning system for unmanned aerial vehicles according to claim 1, characterized in that: The self-check cycle in the system self-check and calibration module is 1-5 minutes before takeoff. It is adjusted according to the flight duration and environmental changes during the flight, and the maximum time is no more than 30 minutes. The calibration methods include automatic calibration and manual calibration.

6. The autonomous obstacle avoidance and path planning system for unmanned aerial vehicles according to claim 1, characterized in that: The load adaptation and optimization module uses pressure sensors and inertial measurement units (IMUs) when calculating load weight and center of gravity position, and uses an adaptive adjustment algorithm when adjusting flight control parameters according to load adaptation.

7. The autonomous obstacle avoidance and path planning system for unmanned aerial vehicles according to claim 1, characterized in that: The laser radar is located on the top of the drone, the binocular cameras are distributed on the front and sides of the fuselage, the millimeter-wave radar is installed on the bottom and sides of the fuselage, and the ultrasonic sensors are evenly distributed around the fuselage.

8. The autonomous obstacle avoidance and path planning system for unmanned aerial vehicles according to claim 1, characterized in that: After fusing the data, the data fusion processing module verifies the fusion result in real time, using the cross-validation method to compare the fused data with the original sensor data. If the error exceeds the preset threshold, the fusion process is performed again.

9. The autonomous obstacle avoidance and path planning system for unmanned aerial vehicles according to claim 1, characterized in that: When generating a path, the path planning module considers the real-time performance status of the UAV and dynamically adjusts the parameters of the path planning algorithm according to the performance status.

10. A method for applying the autonomous obstacle avoidance and path planning system for unmanned aerial vehicles according to any one of claims 1 to 9, characterized in that: The following steps are involved: Environmental perception steps: Use a variety of sensors to collect information about the surrounding environment, including obstacle location, shape, speed, terrain features, and weather conditions; Data fusion step: Transmitting data from different sensors to the data fusion processing module, processing according to the above fusion algorithm and weight adjustment method to obtain a fusion model; Path planning steps: Based on the fusion model, the flight path is generated by using the path planning algorithm combined with terrain data; Flight control steps: The flight control module receives instructions, controls the flight of the drone according to the control algorithm, and adjusts the attitude and speed in real time; Autonomous decision-making steps: The autonomous decision-making module monitors in real time and handles any anomalies according to risk assessment and decision-making algorithms; Communication and interaction steps: The communication module implements data transmission and command interaction to meet the communication packet loss rate and delay requirements; Power management steps: The power management module continuously monitors the power supply, allocates power according to the strategy, estimates the power consumption and issues warnings; Data recording and analysis steps: The ground control station records the data and stores it according to the data storage integrity requirements for evaluation and optimization; System self-check and calibration steps: Perform self-check and calibration before takeoff and in flight to check sensors, algorithm parameters and communication links, and improve system stability according to system reliability improvement requirements; Load adaptation and optimization steps: adjust system parameters according to the load, calculate load characteristics and optimize flight control and path planning, and improve operational efficiency according to load adaptability scores.

Images