Unmanned aerial vehicle formation positioning system and method
By integrating multiple sensors and deep learning technologies, high-precision positioning and autonomous obstacle avoidance of drone formations are achieved, which solves the problems of insufficient positioning and environmental adaptability of traditional systems under signal occlusion and multipath effects, and improves the intelligence of formations. and task execution efficiency.
Patent Information
- Application Number
- CN202411054843.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-02
- Publication Date
- 2024-11-12
AI Technical Summary
Traditional UAV formation positioning systems are difficult to achieve high-precision positioning under the influence of signal occlusion and multipath effect, and their environmental adaptability and dynamic path planning capabilities are insufficient.
Using integrated high-precision GPS receiver, inertial measurement unit, magnetic compass and barometer positioning and navigation unit, combined with lidar, millimeter-wave radar and high-resolution camera perception and obstacle avoidance units, use deep learning and model prediction control to achieve Autonomous obstacle avoidance and dynamic path planning, and improve the self-organization and collaboration capabilities of the formation through cluster intelligence and multi-agent reinforcement learning frameworks.
It improves the positioning accuracy and flight safety of the drone formation, ensures safe and efficient flight in complex environments, and improves the overall intelligence level and mission execution efficiency of the formation.
Smart Images

Figure CN118938279A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of unmanned aerial vehicle formation positioning systems, and in particular to a unmanned aerial vehicle formation positioning system and method. Background Art
[0002] With the development of drone technology, drone formations have been widely used in military reconnaissance, geographic mapping, agricultural spraying, disaster relief and other fields. Drones in formation flight require accurate positioning information to ensure flight safety and mission execution efficiency. Traditional GPS positioning methods are affected by signal shielding and multipath effects and cannot meet the needs of high-precision positioning.
[0003] After searching, the application scheme of Chinese patent application No. 201610838850.8 discloses a UAV positioning system and a UAV take-off and landing method, including a UAV, a wireless positioning base station, a ground station, a wireless positioning tag, and a position correction module. When the UAV enters and exits the range of the wireless positioning base station, the wireless positioning tag receives the positioning radio wave sent by the ground station and replies with a wireless pulse signal. The wireless positioning base station captures the wireless pulse signal and sends the wireless pulse signal and its time point information to the ground station. The ground station processes the wireless pulse signal and its time point information into the three-dimensional coordinate information of the UAV relative to the first base station and sends it to the UAV in real time. The position correction module corrects the position of the UAV in real time through the three-dimensional coordinate information sent by the ground station to achieve accurate positioning. The positioning system in the above patent has the following shortcomings: it mainly relies on wireless positioning base stations and wireless positioning tags for positioning, has general environmental adaptability, and is still lacking in obstacle avoidance and dynamic path planning capabilities. Summary of the invention
[0004] The purpose of the present invention is to solve the shortcomings of the prior art and to propose a UAV formation positioning system and method.
[0005] In order to achieve the above object, the present invention adopts the following technical solutions:
[0006] A UAV formation positioning system, comprising:
[0007] The drone body integrates the positioning and navigation unit, the perception and obstacle avoidance unit, and the communication and collaboration unit;
[0008] The ground control station is responsible for data collection, processing and command transmission;
[0009] Communication link for data transmission between the UAV and the ground control station;
[0010] Among them, the positioning and navigation unit integrates high-precision GPS receivers and inertial measurement units to provide basic position and attitude information; it is equipped with a magnetic compass and a barometer to enhance the measurement accuracy of heading and altitude; the perception and obstacle avoidance unit is equipped with laser radar and millimeter-wave radar to achieve environmental scanning and obstacle detection; it integrates a high-resolution camera and uses computer vision algorithms for real-time image processing and obstacle identification; the communication and collaboration unit uses a high-frequency broadband data link to ensure stable communication with the ground control station and other UAVs; it integrates a network protocol stack to support multiple communication modes and network topologies to achieve flexible formation configuration;
[0011] The ground control station comprises:
[0012] Data processing and management unit: equipped with high-performance servers, running data fusion and positioning algorithms, processing sensor data from drones; setting up a database system to store flight data, environmental maps and mission planning information, and supporting data playback and analysis;
[0013] Mission Planning and Decision-Making Unit: Develop a graphical interface that allows operators to quickly set mission parameters and waypoint planning; integrate an intelligent decision support system to provide mission planning suggestions and risk assessment to assist operators in making decisions;
[0014] Monitoring and control unit: implements a real-time monitoring interface to display drone status, location, video stream and warning information; is equipped with a manual control interface to allow direct remote control operation in certain situations;
[0015] The communication link comprises:
[0016] Signal transmission unit: Use multi-channel, high-bandwidth radio equipment to ensure data transmission stability and low latency; apply advanced modulation and demodulation technology and channel coding strategies to optimize the robustness of signals in complex environments;
[0017] Network management unit: implement dynamic network management protocols, automatically adjust communication resource allocation, and adapt to changing formation topology; introduce self-organizing network technology to support direct communication between drones and improve the flexibility of formation coordination.
[0018] Preferably, the positioning method of the positioning system comprises the following steps:
[0019] S1: Data preprocessing: filtering and calibrating the data collected by each sensor to eliminate noise and system errors;
[0020] S2: Feature extraction: Use the environmental features obtained by LiDAR and camera to match and identify and build a local map;
[0021] S3: State estimation: The position and attitude of the drone are estimated by fusing GPS data, IMU data, and local map information through a Kalman filter or a particle filter.
[0022] S4: Relative positioning: The relative positions between drones are calculated through distance measurement and angle measurement technology, and combined with the state estimation results, the relative positioning of drones in the formation is realized;
[0023] S5: Global optimization: Use graph optimization algorithm to optimize the global consistency of the positioning results of the entire formation to improve positioning accuracy.
[0024] Preferably, in S3, the data of different sensors are fused through the Kalman filter algorithm, and the specific formula is:
[0025]
[0026] P k =(IK k H k ) k-1
[0027] in, is the state estimate at time k, K k is the Kalman gain, z k is the observed value, H k is the observation matrix, P k is the error covariance matrix.
[0028] Preferably, in S3, the particle filter algorithm is used to process nonlinear and non-Gaussian noise, and the specific formula is:
[0029] p(x k |z 1:k )∝p(z k |x k )∫p(x k |x k-1 )p(x k-1 |z 1:k-1 )dx k-1
[0030] Among them, p(x k |z 1:k ) is the posterior probability distribution, p(z k |x k ) is the likelihood function, p(x k |x k-1 ) is the state transition probability.
[0031] Preferably, in S5, global positioning optimization is performed by a graph optimization algorithm, and the specific formula is:
[0032]
[0033] Among them, x is the node state vector, E is the set of edges, and e ij is the error vector of edge (i,j), Ω ij is the information matrix.
[0034] Preferably: also include:
[0035] Machine Learning Module: Use machine learning algorithms to classify and identify environmental features to improve the accuracy and efficiency of feature extraction; use convolutional neural networks to process image data to identify the characteristics of terrain or obstacles; apply reinforcement learning algorithms to optimize the flight path and formation strategy of drones, and learn the optimal behavior strategy through interaction with the environment to improve the efficiency and safety of task execution;
[0036] Adaptive filtering module: Combined with adaptive filtering technology, filter parameters are dynamically adjusted according to actual observation data to better handle nonlinear systems and model uncertainties; deep reinforcement learning is introduced to conduct online learning and adjustment of sensor fusion strategies in complex environments to adapt to environmental changes and sensor performance fluctuations;
[0037] Intelligent decision support module: Integrate expert system and fuzzy logic to provide intelligent decision support for UAV formation; use genetic algorithm or particle swarm optimization algorithm to optimize formation control parameters and automatically find the optimal control strategy to improve the stability and response speed of the formation;
[0038] Autonomous obstacle avoidance and path planning module: It uses semantic map construction technology based on deep learning to achieve a deep understanding of the environment and provide rich contextual information for path planning. It combines model predictive control and deep learning to develop autonomous obstacle avoidance and dynamic path planning algorithms to ensure that drone formations can fly safely and efficiently in complex environments.
[0039] Swarm intelligence and collaboration module: Apply the principles of swarm intelligence to achieve self-organization and collaboration in UAV formations, and improve the adaptability and flexibility of the formation to complex tasks; use the multi-agent reinforcement learning framework to coordinate the behavior of each UAV in the formation, achieve resource sharing and information exchange, and improve the overall intelligence level of the formation.
[0040] Preferably: the machine learning module includes:
[0041] Environmental feature learning unit: Use convolutional neural networks and image processing technology to extract and classify environmental features; combine semantic segmentation and object detection technology to improve the ability to understand complex scenes;
[0042] Flight Strategy Optimization Unit: Apply reinforcement learning algorithms to optimize flight paths and formation strategies through simulation and real flight data training; integrate model predictive control to achieve more accurate and responsive flight control.
[0043] Preferably: the adaptive filtering module comprises:
[0044] State estimation and correction unit: implements extended Kalman filtering and unscented Kalman filtering to improve the accuracy of multi-sensor data fusion; introduces a quality detection mechanism to dynamically adjust the weight of sensor data to deal with sensor failure or performance degradation;
[0045] Environmental adaptability adjustment unit: Adopts adaptive neuro-fuzzy inference system to adjust control parameters according to environmental changes; combines deep learning to pre-process sensor data in complex environments and improve the robustness of the filtering algorithm.
[0046] Preferably: the intelligent decision support module includes:
[0047] Formation control optimization unit: Apply genetic algorithm and particle swarm optimization algorithm to automatically search for optimal formation control parameters; integrate expert system to provide rule-based decision support;
[0048] Risk assessment and management unit: Use Bayesian networks and fuzzy logic to assess and warn of potential risks during task execution; provide real-time risk analysis reports to support operators in making timely decision adjustments.
[0049] Preferably: the autonomous obstacle avoidance and path planning module includes:
[0050] Environmental perception and map construction unit: Use 3D laser radar and stereo vision system to build high-precision environmental maps; apply deep learning to build semantic maps to understand and mark environmental elements;
[0051] Dynamic path planning unit: Combines A-search and RRT algorithms to generate the optimal flight path to avoid obstacles; uses model predictive control for path tracking to ensure smooth and accurate flight;
[0052] The cluster intelligence and collaboration module includes:
[0053] Self-organizing collaborative unit: Apply ant colony algorithm and bird swarm algorithm to realize the self-organization and collaborative flight of drone formations; realize role allocation and task coordination among drones through the multi-agent system framework;
[0054] Swarm intelligence decision-making unit: Utilize swarm intelligence algorithms to optimize the overall task execution efficiency of the formation; integrate deep learning to achieve complex decision-making in the formation.
[0055] The beneficial effects of the present invention are:
[0056] 1. The present invention integrates multiple sensors such as high-precision GPS receivers, inertial measurement units, magnetic compasses and barometers, so that the system can provide accurate position and attitude information and enhance the measurement accuracy of heading and altitude. It is equipped with sensors such as laser radar, millimeter-wave radar and high-resolution cameras, combined with computer vision algorithms, to achieve comprehensive scanning of the environment, obstacle detection and identification, and improve flight safety.
[0057] 2. The present invention combines deep learning and model predictive control technology to achieve autonomous obstacle avoidance and dynamic path planning, ensuring that UAV formations can fly safely and efficiently in complex environments; it applies cluster intelligence principles and a multi-agent reinforcement learning framework to achieve self-organization and collaborative behavior in UAV formations, thereby improving the overall intelligence level of the formation and task execution efficiency.
[0058] 3. The present invention integrates expert system, fuzzy logic, genetic algorithm and particle swarm optimization algorithm to provide intelligent decision support for UAV formation, optimize formation control parameters, improve stability and response speed; and adopts graph optimization algorithm to perform global consistency optimization on the positioning results of the entire formation to further improve positioning accuracy. BRIEF DESCRIPTION OF THE DRAWINGS
[0059] Figure 1 The present invention provides a flow chart of a positioning method for a UAV formation positioning system. DETAILED DESCRIPTION
[0060] The technical solution of the present invention is further described in detail below in conjunction with specific implementation methods.
[0061] Embodiment 1:
[0062] A UAV formation positioning system, comprising:
[0063] The drone body integrates the positioning and navigation unit, the perception and obstacle avoidance unit, and the communication and collaboration unit;
[0064] The ground control station is responsible for data collection, processing and command transmission;
[0065] Communication link for data transmission between the UAV and the ground control station;
[0066] Among them, the positioning and navigation unit integrates a high-precision GPS receiver and inertial measurement unit (IMU) to provide basic position and attitude information; it is equipped with a magnetic compass and a barometer to enhance the measurement accuracy of heading and altitude; the perception and obstacle avoidance unit is equipped with a laser radar (LiDAR) and a millimeter-wave radar to achieve environmental scanning and obstacle detection; it integrates a high-resolution camera and uses computer vision algorithms for real-time image processing and obstacle identification; the communication and collaboration unit uses a high-frequency broadband data link to ensure stable communication with the ground control station and other drones; it integrates a network protocol stack to support multiple communication modes and network topologies to achieve flexible formation configuration.
[0067] Wherein, the ground control station comprises:
[0068] Data processing and management unit: equipped with high-performance servers, running data fusion and positioning algorithms, processing sensor data from drones; setting up a database system to store flight data, environmental maps and mission planning information, and supporting data playback and analysis;
[0069] Mission Planning and Decision-Making Unit: Develop a graphical interface that allows operators to quickly set mission parameters and waypoint planning; integrate an intelligent decision support system to provide mission planning suggestions and risk assessment to assist operators in making decisions;
[0070] Monitoring and control unit: implements a real-time monitoring interface to display drone status, location, video stream and warning information; is equipped with a manual control interface to allow direct remote control operation in specific situations.
[0071] Wherein, the communication link includes:
[0072] Signal transmission unit: Use multi-channel, high-bandwidth radio equipment to ensure data transmission stability and low latency; apply advanced modulation and demodulation technology and channel coding strategies to optimize the robustness of signals in complex environments;
[0073] Network management unit: implement dynamic network management protocols, automatically adjust communication resource allocation, and adapt to changing formation topology; introduce self-organizing network technology to support direct communication between drones and improve the flexibility of formation coordination.
[0074] The positioning method of the positioning system comprises the following steps:
[0075] S1: Data preprocessing: filtering and calibrating the data collected by each sensor to eliminate noise and system errors;
[0076] S2: Feature extraction: Use the environmental features obtained by LiDAR and camera to match and identify and build a local map;
[0077] S3: State estimation: The position and attitude of the drone are estimated by fusing GPS data, IMU data, and local map information through a Kalman filter or a particle filter.
[0078] S4: Relative positioning: The relative positions between drones are calculated through distance measurement and angle measurement technology, and combined with the state estimation results, the relative positioning of drones in the formation is realized;
[0079] S5: Global optimization: Use graph optimization algorithm to optimize the global consistency of the positioning results of the entire formation to improve positioning accuracy.
[0080] Among them, in S3, the data of different sensors are fused through the Kalman filter algorithm, and the specific formula is:
[0081]
[0082] P k =(IK k H k ) k-1
[0083] in, is the state estimate at time k, K k is the Kalman gain, z k is the observed value, H k is the observation matrix, P k is the error covariance matrix.
[0084] Among them, in S3, the particle filter algorithm is used to process nonlinear and non-Gaussian noise, and the specific formula is:
[0085] p(x k |z 1:k )∝p(z k |x k )∫p(x k |x k-1 )p(x k-1 |z 1:k-1 )dx k-1
[0086] Among them, p(x k |z 1:k ) is the posterior probability distribution, p(z k |x k ) is the likelihood function, p(x k |x k-1 ) is the state transition probability.
[0087] Among them, in S5, global positioning optimization is performed through a graph optimization algorithm, and the specific formula is:
[0088]
[0089] Among them, x is the node state vector, E is the set of edges, and e ij is the error vector of edge (i,j), Ω ij is the information matrix.
[0090] Embodiment 2:
[0091] A drone formation positioning system, based on embodiment 1, further includes:
[0092] Machine Learning Module: Use machine learning algorithms to classify and identify environmental features to improve the accuracy and efficiency of feature extraction; for example, use convolutional neural networks (CNNs) to process image data to identify the characteristics of terrain or obstacles; apply reinforcement learning algorithms to optimize the flight path and formation strategy of drones, and learn the optimal behavior strategy through interaction with the environment to improve the efficiency and safety of task execution;
[0093] Adaptive filtering module: Combines adaptive filtering techniques, such as extended Kalman filtering (EKF) or unscented Kalman filtering (UKF), to dynamically adjust filter parameters based on actual observation data to better handle nonlinear systems and model uncertainties; introduces deep reinforcement learning to learn and adjust sensor fusion strategies online in complex environments to adapt to environmental changes and sensor performance fluctuations;
[0094] Intelligent decision support module: Integrates expert systems and fuzzy logic to provide intelligent decision support for UAV formations, such as decision-making for formation, maintenance and disbanding operations; uses genetic algorithms or particle swarm optimization algorithms to optimize formation control parameters and automatically find the optimal control strategy to improve the stability and response speed of the formation;
[0095] Autonomous obstacle avoidance and path planning module: It uses semantic map construction technology based on deep learning to achieve a deep understanding of the environment and provide rich contextual information for path planning. It combines model predictive control (MPC) and deep learning to develop autonomous obstacle avoidance and dynamic path planning algorithms to ensure that drone formations can fly safely and efficiently in complex environments.
[0096] Swarm intelligence and collaboration module: Apply swarm intelligence principles, such as ant colony algorithm or bird colony algorithm, to achieve self-organization and collaboration in UAV formations, and improve the adaptability and flexibility of the formation to complex tasks; use a multi-agent reinforcement learning framework to coordinate the behavior of each UAV in the formation, achieve resource sharing and information exchange, and improve the overall intelligence level of the formation.
[0097] Wherein, the machine learning module includes:
[0098] Environmental feature learning unit: Use convolutional neural network (CNN) and image processing technology to extract and classify environmental features; combine semantic segmentation and object detection technology to improve the ability to understand complex scenes;
[0099] Flight Strategy Optimization Unit: Apply reinforcement learning algorithms to optimize flight paths and formation strategies through simulation and real flight data training; integrate model predictive control (MPC) to achieve more precise and responsive flight control.
[0100] Wherein, the adaptive filtering module comprises:
[0101] State estimation and correction unit: Implements the extended Kalman filter (EKF) and the unscented Kalman filter (UKF) to improve the accuracy of multi-sensor data fusion; introduces a quality detection mechanism to dynamically adjust the weight of sensor data to deal with sensor failure or performance degradation;
[0102] Environmental adaptability adjustment unit: Adopts adaptive neural fuzzy inference system (ANFIS) to adjust control parameters according to environmental changes; combines deep learning to pre-process sensor data in complex environments to improve the robustness of the filtering algorithm.
[0103] Wherein, the intelligent decision support module includes:
[0104] Formation control optimization unit: Apply genetic algorithm and particle swarm optimization algorithm to automatically search for the optimal formation control parameters; integrate expert system to provide rule-based decision support, such as the timing of formation and disbanding;
[0105] Risk assessment and management unit: Use Bayesian networks and fuzzy logic to assess and warn of potential risks during task execution; provide real-time risk analysis reports to support operators in making timely decision adjustments.
[0106] Wherein, the autonomous obstacle avoidance and path planning module includes:
[0107] Environmental perception and map construction unit: Use 3D laser radar and stereo vision system to build high-precision environmental maps; apply deep learning to build semantic maps to understand and mark environmental elements;
[0108] Dynamic Path Planning Unit: Combines A-search and RRT algorithms to generate the optimal flight path to avoid obstacles; uses Model Predictive Control (MPC) for path tracking to ensure smooth and precise flight.
[0109] Wherein, the cluster intelligence and collaboration module includes:
[0110] Self-organizing collaborative unit: The ant colony algorithm and the bird swarm algorithm are applied to realize the self-organization and collaborative flight of the UAV formation; the role allocation and task coordination between UAVs are realized through the multi-agent system (MAS) framework;
[0111] Swarm intelligence decision-making unit: Utilize swarm intelligence algorithms, such as particle swarm optimization (PSO), to optimize the overall task execution efficiency of the formation; integrate deep learning to achieve complex decision-making in the formation, such as dynamic task allocation and environmental adaptation.
[0112] The above description is only a preferred specific implementation manner 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 scheme and inventive concept 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. A UAV formation positioning system, characterized in that: include: The drone body integrates the positioning and navigation unit, the perception and obstacle avoidance unit, and the communication and collaboration unit; The ground control station is responsible for data collection, processing and command transmission; Communication link for data transmission between the UAV and the ground control station; Among them, the positioning and navigation unit integrates high-precision GPS receivers and inertial measurement units to provide basic position and attitude information; it is equipped with a magnetic compass and a barometer to enhance the measurement accuracy of heading and altitude; the perception and obstacle avoidance unit is equipped with laser radar and millimeter-wave radar to achieve environmental scanning and obstacle detection; it integrates a high-resolution camera and uses computer vision algorithms for real-time image processing and obstacle identification; the communication and collaboration unit uses a high-frequency broadband data link to ensure stable communication with the ground control station and other UAVs; it integrates a network protocol stack to support multiple communication modes and network topologies to achieve flexible formation configuration; The ground control station comprises: Data processing and management unit: equipped with high-performance servers, running data fusion and positioning algorithms, processing sensor data from drones; setting up a database system to store flight data, environmental maps and mission planning information, and supporting data playback and analysis; Mission Planning and Decision-Making Unit: Develop a graphical interface that allows operators to quickly set mission parameters and waypoint planning; integrate an intelligent decision support system to provide mission planning suggestions and risk assessment to assist operators in making decisions; Monitoring and control unit: implements a real-time monitoring interface to display drone status, location, video stream and warning information; is equipped with a manual control interface to allow direct remote control operation in certain situations; The communication link comprises: Signal transmission unit: Use multi-channel, high-bandwidth radio equipment to ensure data transmission stability and low latency; apply advanced modulation and demodulation technology and channel coding strategies to optimize the robustness of signals in complex environments; Network management unit: implement dynamic network management protocols, automatically adjust communication resource allocation, and adapt to changing formation topology; introduce self-organizing network technology to support direct communication between drones and improve the flexibility of formation coordination.
2. The UAV formation positioning system according to claim 1, characterized in that: The positioning method of the positioning system comprises the following steps: S1: Data preprocessing: filtering and calibrating the data collected by each sensor to eliminate noise and system errors; S2: Feature extraction: Use the environmental features obtained by LiDAR and camera to match and identify and build a local map; S3: State estimation: The position and attitude of the drone are estimated by fusing GPS data, IMU data, and local map information through a Kalman filter or a particle filter. S4: Relative positioning: The relative positions between drones are calculated through distance measurement and angle measurement technology, and combined with the state estimation results, the relative positioning of drones in the formation is realized; S5: Global optimization: Use graph optimization algorithm to optimize the global consistency of the positioning results of the entire formation to improve positioning accuracy.
3. The UAV formation positioning system according to claim 2, characterized in that: In S3, the data of different sensors are fused through the Kalman filter algorithm. The specific formula is: P k =(I-K k H k )P k-1 in, is the state estimate at time k, K k is the Kalman gain, z k is the observed value, H k is the observation matrix, P k is the error covariance matrix.
4. The UAV formation positioning system according to claim 2, characterized in that: In S3, the particle filter algorithm is used to process nonlinear and non-Gaussian noise. The specific formula is: p(x k |z 1:k )∝p(z k |x k )∫p(x k |x k-1 )p(x k-1 |z 1:k-1 )dx k-1 Among them, p(x k |z 1:k ) is the posterior probability distribution, p(z k |x k ) is the likelihood function, P(x k |x k-1 ) is the state transition probability.
5. The UAV formation positioning system according to claim 2, characterized in that: In S5, global positioning optimization is performed by using a graph optimization algorithm, and the specific formula is: Among them, x is the node state vector, E is the set of edges, and e ij is the error vector of edge (i,j), Ω ij is the information matrix.
6. The UAV formation positioning system according to claim 1, characterized in that: Also includes: Machine Learning Module: Use machine learning algorithms to classify and identify environmental features to improve the accuracy and efficiency of feature extraction; use convolutional neural networks to process image data to identify the characteristics of terrain or obstacles; Apply reinforcement learning algorithms to optimize the flight path and formation strategy of drones, and learn the optimal behavior strategy through interaction with the environment to improve the efficiency and safety of mission execution; Adaptive filtering module: Combined with adaptive filtering technology, filter parameters are dynamically adjusted according to actual observation data to better handle nonlinear systems and model uncertainties; deep reinforcement learning is introduced to conduct online learning and adjustment of sensor fusion strategies in complex environments to adapt to environmental changes and sensor performance fluctuations; Intelligent decision support module: Integrate expert system and fuzzy logic to provide intelligent decision support for UAV formation; use genetic algorithm or particle swarm optimization algorithm to optimize formation control parameters and automatically find the optimal control strategy to improve the stability and response speed of the formation; Autonomous obstacle avoidance and path planning module: It uses semantic map construction technology based on deep learning to achieve a deep understanding of the environment and provide rich contextual information for path planning. It combines model predictive control and deep learning to develop autonomous obstacle avoidance and dynamic path planning algorithms to ensure that drone formations can fly safely and efficiently in complex environments. Swarm intelligence and collaboration module: Apply the principles of swarm intelligence to achieve self-organization and collaboration in UAV formations, and improve the adaptability and flexibility of the formation to complex tasks; use the multi-agent reinforcement learning framework to coordinate the behavior of each UAV in the formation, achieve resource sharing and information exchange, and improve the overall intelligence level of the formation.
7. The UAV formation positioning system according to claim 6, characterized in that: The machine learning module includes: Environmental feature learning unit: Use convolutional neural networks and image processing technology to extract and classify environmental features; combine semantic segmentation and object detection technology to improve the ability to understand complex scenes; Flight Strategy Optimization Unit: Apply reinforcement learning algorithms to optimize flight paths and formation strategies through simulation and real flight data training; integrate model predictive control to achieve more accurate and responsive flight control.
8. The UAV formation positioning system according to claim 7, characterized in that: The adaptive filtering module comprises: State estimation and correction unit: implements extended Kalman filtering and unscented Kalman filtering to improve the accuracy of multi-sensor data fusion; introduces a quality detection mechanism to dynamically adjust the weight of sensor data to deal with sensor failure or performance degradation; Environmental adaptability adjustment unit: Adopts adaptive neuro-fuzzy inference system to adjust control parameters according to environmental changes; combines deep learning to pre-process sensor data in complex environments and improve the robustness of the filtering algorithm.
9. The UAV formation positioning system according to claim 8, characterized in that: The intelligent decision support module includes: Formation control optimization unit: Apply genetic algorithm and particle swarm optimization algorithm to automatically search for optimal formation control parameters; integrate expert system to provide rule-based decision support; Risk assessment and management unit: Use Bayesian networks and fuzzy logic to assess and warn of potential risks during task execution; provide real-time risk analysis reports to support operators in making timely decision adjustments.
10. The UAV formation positioning system according to claim 9, characterized in that: The autonomous obstacle avoidance and path planning module includes: Environmental perception and map construction unit: Use 3D laser radar and stereo vision system to build high-precision environmental maps; apply deep learning to build semantic maps to understand and mark environmental elements; Dynamic path planning unit: Combines A-search and RRT algorithms to generate the optimal flight path to avoid obstacles; uses model predictive control for path tracking to ensure smooth and accurate flight; The cluster intelligence and collaboration module includes: Self-organizing collaborative unit: Apply ant colony algorithm and bird swarm algorithm to realize the self-organization and collaborative flight of drone formations; realize role allocation and task coordination among drones through the multi-agent system framework; Swarm intelligence decision-making unit: Utilize swarm intelligence algorithms to optimize the overall task execution efficiency of the formation; integrate deep learning to achieve complex decision-making in the formation.
Citation Information
Cited By
Self-adaptive navigation system of unmanned aerial vehicle
CN119594989A
An adaptive navigation system for unmanned aerial vehicles
CN119594989B
Unmanned aerial vehicle detection tracking method and system based on combination of laser radar and vision
CN119846652A
An Unmanned Aerial Vehicle Detection and Tracking Method and System Based on the Combination of LiDAR and Vision
CN119846652B
Unmanned aerial vehicle infrared target positioning system fusing multi-model KF algorithm and deep learning
CN119901293A
