Unmanned aerial vehicle fault-tolerant prediction control method simulating pigeon flock bifurcation optimization

Abstract

The invention discloses an optimized unmanned aerial vehicle fault-tolerant prediction control method simulating pigeon flock bifurcation. The method comprises the following steps: step 1, modeling an unmanned aerial vehicle flight control system; 2, designing a pigeon flock bifurcation simulation optimization algorithm; 3, designing a prediction performance function; 4, a correction mechanism based on fault bifurcation prediction information is introduced, and control input is adjusted through an optimization algorithm; and 5, designing a state feedback controller and realizing global information cooperative adjustment. The method is easy to implement, control input can be rapidly adjusted when a fault occurs, and the stability and task execution of an unmanned aerial vehicle system are ensured. The fault response speed is optimized, the communication load is effectively reduced through bifurcation control, and the control convergence is improved. According to the method, a high-robustness and quick-response solution is provided for cooperative control of multiple unmanned aerial vehicles in a communication limited environment, and dynamic topology switching and system fault challenges are effectively handled.

Description

Technical Field

[0001] The present invention relates to a method for fault-tolerant predictive control of unmanned aerial vehicles optimized by pigeon flock bifurcation, belonging to the technical field of cooperative control of unmanned aerial vehicle clusters. Background Art

[0002] With the continuous progress of unmanned aerial vehicle technology, especially the wide application of multi-unmanned aerial vehicle cooperation systems in various complex tasks, how to improve the fault tolerance of the system, especially to ensure the stable execution of the overall task in the case of local unmanned aerial vehicle failures, has become the key to current research. In a multi-unmanned aerial vehicle cooperation system, the high degree of dependence between individuals makes a single fault may lead to a significant decline in system performance or mission failure. Traditional fault-tolerant control methods usually rely on technologies such as redundant configuration, real-time fault diagnosis and state estimation. Although they can ensure the normal operation of the system to a certain extent, these methods are often limited by their high computational complexity, long response delay and waste of redundant resources in the face of complex and dynamically changing operating environments. Therefore, how to not only detect and repair faults in real time when the system encounters faults, but also improve the flexibility and robustness of the system has become the core problem to be solved urgently.

[0003] In order to overcome the limitations of traditional fault-tolerant control strategies, the bifurcation optimization control algorithm has shown its unique advantages in multi-unmanned aerial vehicle cooperation systems. The basic principle of the bifurcation optimization control algorithm lies in deeply analyzing the dynamic characteristics of the system, identifying key bifurcation points, and guiding the system to change from one stable state to another when the system state changes. Specifically, the bifurcation optimization control algorithm can not only achieve precise control of system behavior, but also avoid the system from entering an unstable state or falling into a local out-of-control situation by appropriately adjusting control parameters when the system fails or undergoes a state mutation. Different from traditional fault-tolerant control methods that rely on redundant mechanisms or post-fault repair strategies, the bifurcation optimization control algorithm can predict and actively adjust control strategies at the precursor stage of fault occurrence by mastering the evolution trend of the system state in real time, thereby effectively avoiding a sharp decline in system performance and ensuring the smooth progress of the mission.

[0004] In the application of multi - UAV cooperative systems, the bifurcation optimization control algorithm can precisely regulate the flight trajectories and task execution strategies of each UAV by making full use of swarm intelligence and cooperation mechanisms, achieving global optimization. When a certain UAV fails, the bifurcation control can quickly sense and guide other UAVs to reorganize tasks through collaborative work, thus ensuring that the cluster can quickly adapt to changes and avoid the impact of local failures on overall task execution. This strategy ensures that when there are failures in the UAV sensors or power systems, the flight strategy can be quickly adjusted and the task can be continuously executed through dynamic optimization of the flight trajectory and task allocation. Compared with traditional fault - tolerance methods that rely on redundant configurations or complex state estimation, the bifurcation optimization control algorithm can quickly find the optimal control strategy through flexible dynamic adjustment when a failure occurs, avoiding the delays and computational complexities in the traditional methods' fault responses. Therefore, through effective control of the system bifurcation phenomenon, the bifurcation optimization control algorithm not only significantly improves the fault - tolerance ability of the UAV system but also shows higher response speed and computational efficiency when dealing with dynamic environments, possessing stronger practical application potential. Summary of the Invention

[0005] The purpose of the present invention is to provide a UAV fault - tolerance predictive control method based on pigeon - flock - like bifurcation optimization, aiming to solve the fault - tolerance control problem in multi - UAV cooperative systems and improve the robustness of the system and the stability of task execution.

[0006] Aiming at the cooperative control problem of UAV clusters in the environment of actuator hybrid faults, the present invention invents a method combining pigeon - flock - like bifurcation optimization and fault - tolerance control strategies. The algorithm implementation flowchart is as Figure 1 shown, including several parts such as flight control system modeling, pigeon - flock - like optimization algorithm design, predictive performance function design, correction mechanism of fault bifurcation prediction information, multi - UAV cooperative control framework, etc. The specific implementation steps are as follows:

[0007] Step 1: UAV flight control system modeling. Among them, considering the uncertain nonlinear function of the research object caused by external disturbances and parameter uncertainties, multiplicative faults and additive faults are added, and by designing appropriate control inputs, the state of the entire system can reach the target position;

[0008] Specifically:

[0009] Establish a dynamic system composed of N UAVs. The dynamic equation of each UAV is:

[0010]

[0011] where, X i (t) = [x i , y i , z i , φi , θ i , ψ i ∈ R is the state vector of the i-th unmanned aerial vehicle, u i (t) ∈ R m is the control input, where R represents the one-dimensional scalar space, and R m represents the m-dimensional vector space. f i (·) is an uncertain nonlinear function that describes the object under study caused by external disturbances and parameter uncertainties. The design objective of the control input is to make the state of the entire system reach the target position or make the state error decay to zero through an appropriate control input u i (t).

[0012] For the control input of each unmanned aerial vehicle, it is assumed that the actuator of the unmanned aerial vehicle may be affected by additive and multiplicative faults, and the effects of the faults are represented by η i (t) and ψ i (t) as follows:

[0013]

[0014] In the formula, is the ideal control input, η i (t) represents the multiplicative fault, and ψ i (t) represents the additive fault. The fault terms η i (t) and ψ i (t) are unknown time-varying functions that reflect the errors of the unmanned aerial vehicle actuator during flight.

[0015] Step 2: Design of the pigeon flock bifurcation optimization algorithm.

[0016] Specifically: The pigeon flock bifurcation optimization algorithm simulates the group behavior of pigeon flocks. Through the coordination and information exchange among individuals, it optimizes the control strategy of each unmanned aerial vehicle, constrains the magnitude of the control input, and realizes global objective optimization. In the pigeon flock bifurcation optimization algorithm, each unmanned aerial vehicle not only reacts according to its own state, but also adjusts according to the states of neighboring unmanned aerial vehicles, forming group wisdom. The key feature of pigeon flock optimization is the bifurcation behavior, that is, the system can jump out of the local optimal solution and explore the global optimal solution by adjusting the individual behavior at the appropriate time.

[0017] The goal of pigeon flock bifurcation optimization is to dynamically adjust the control input u i (t) of each unmanned aerial vehicle, and the expression is as follows:

[0018]

[0019] In the formula, N iLet \(\mathcal{N}_i\) denote the set of UAVs adjacent to the \(i\)-th UAV, and \(\lambda\) be the regularization parameter. The objective of this optimization problem is to minimize the distance between each UAV and its neighboring UAVs while constraining the magnitude of the control input, thereby achieving global objective optimization.

[0020] Step 3: Design of the prediction performance function. Introduce the prediction performance function to measure the system state error, design the definition of the global error level, minimize the system state error, and ensure the effectiveness of fault-tolerant control.

[0021] Specifically: To ensure the fault-tolerant ability of the system and control the error decay, the present invention introduces a prediction performance function to measure the system state error. The prediction performance function \(P(t)\) reflects the global error level of the system and is defined as:

[0022]

[0023] where is the target trajectory of each UAV. The objective of the function \(P(t)\) is to minimize the system state error and ensure that the state error can still decay when a fault occurs.

[0024] To ensure the stability and fault tolerance of the system, the error decay rate is designed as:

[0025]

[0026] where is the state error of the \(i\)-th UAV, and

[0027] e i (t)=[e i1 (t),e i2 (t),e i3 (t),e i4 (t),e i5 (t),e i6 (t)], \(\Delta e\) i (t) is the correction term based on the fault estimation. By adjusting the decay factor \(\sigma\) i , it can be ensured that the state error of each UAV gradually decays to the target range when a fault occurs.

[0028] To ensure the effectiveness of the fault-tolerant control, the following conditions need to be satisfied:

[0029]

[0030] Step 4: First, introduce a correction mechanism based on the fault bifurcation prediction information and adjust the control input through an optimization algorithm. Design an adaptive adjustment strategy, real-time updated fault information, and the performance function \(P(t)\). By evaluating the prediction performance function in real time, determine whether the current control strategy meets the desired system performance.

[0031] Specifically, to adjust the control input according to the estimated result of the fault, we introduce a fault correction term Δu i (t), which depends on the fault detection result and the prediction performance function. The design of the correction term can be based on the following dynamic model:

[0032]

[0033] where is the amount of information based on fault prediction, reflecting the change in system error caused by the fault; α i and β i are adjustment coefficients, determining the sensitivity of the correction term to the error and the prediction correction amount.

[0034] The fault correction term Δu i (t) affects the control input u i (t) of the UAV. Through dynamic adjustment, the system can return to the stable state as soon as possible after the fault occurs. The update formula of the control input is as follows:

[0035]

[0036] where, is the ideal control input without faults, and the fault correction term Δu i (t) is the part dynamically adjusted according to the fault correction information. To ensure the stability and constraint conditions of the control input during the correction process, a weighting factor γ i can be introduced to balance the relationship between Δu i (t) and :

[0037]

[0038] where, γ i ∈[0,1] is a regulation coefficient, depending on the severity of the fault. Assume that the severity of the system fault is S, and its value range is defined as [0,1]. Among them, when the value of S is between [0, 0.3], it represents a minor fault, and the correction of the control input is relatively conservative, γ(S) = γ min . When the value of S is between [0.3, 0.7], it represents a medium fault, and the correction of the control input needs to be moderate, γ i = γ(S) = γ min +(S - 0.3)(γ max - γ min ) / 0.4. When the value of S is between [0.7, 1], it represents a severe fault, and the correction amplitude of the control input needs to be increased, γ(S) = γ max, where γ(S) is a regulation function to quantify the degree of correction of the control input, and γ min and γ max represent the minimum and maximum regulation coefficients for fault correction, respectively. During the whole process, the adaptive adjustment strategy is optimized through real-time feedback control. When the system detects a fault and predicts the fault impact, the bifurcation fault-tolerant control strategy automatically adjusts the control input of each UAV according to the real-time updated fault information and the performance function P(t). This adjustment not only considers the local control strategy (such as minimizing the state error), but also optimizes according to the requirements of swarm cooperation to ensure the collaborative performance of the entire UAV swarm and the continuity of task execution. Specifically, the control input of each UAV will be adaptively adjusted under the following circumstances: according to the severity of the fault detection, the correction amplitude and method of the control input are adjusted in real time; by real-time evaluating the predicted performance function, it is judged whether the current control strategy meets the expected system performance; through the bifurcation optimization pigeon flock algorithm, the control strategy is adjusted according to the states and behaviors of neighboring UAVs to achieve swarm cooperation.

[0039] Step Five: Design a state feedback controller and implement global information collaborative regulation, and adjust the latest system control input optimization target u i (t) according to the states, distances, and flight target information of neighboring UAVs to ensure the overall stability of the swarm. By dynamically adjusting the flight paths and control strategies of each UAV, the fault tolerance ability is further enhanced to ensure the recovery speed and execution efficiency of the system after a fault occurs.

[0040] Specifically: Based on Steps Three and Four, this step combines state feedback and global information collaborative regulation to achieve rapid recovery and high-efficiency fault-tolerant control of the system. The key to the cooperative control of a multi-UAV system lies in optimizing the control input of each UAV while ensuring the unity and stability of the swarm behavior. By establishing an information transmission and coordination mechanism among each UAV, the present invention proposes a cooperative control framework based on bifurcation control and network topology structure optimization. Assume that each UAV not only adjusts its control input according to its own state, but also adjusts according to the states, distances, and flight target information of neighboring UAVs.

[0041] For each UAV, the optimization target of the system control input can be modified as:

[0042]

[0043] where N i represents the set of UAVs neighboring the i-th UAV, and κ is a newly introduced cooperative adjustment parameter, representing the control coordination weight between adjacent UAVs. Represents the distance metric between the i-th and j-th drones. By adjusting this parameter, the cooperative control and information interaction between drones can be enhanced to a certain extent, thereby improving the robustness of the system. Through the above optimization objectives, the system can quickly adjust the control input when a drone fails and achieve state consistency with the surrounding drones to ensure the overall stability of the cluster.

[0044] 3. Advantages and effects:

[0045] The present invention proposes a fault-tolerant predictive control method for drones optimized by simulating pigeon flock bifurcation, providing an innovative fault-tolerant control strategy for drones, aiming to improve the robustness and fault recovery ability of multi-drone systems. Traditional fault-tolerant control methods usually rely on redundancy mechanisms and real-time fault diagnosis, but in complex dynamic environments, there are often problems such as high computational complexity and large response delays. The present invention, by introducing the bifurcation-optimized pigeon flock algorithm, can monitor and predict the impact of faults in real time when a fault occurs in the system, thereby optimizing the control strategy to ensure that the drone system can maintain stable task execution. The bifurcation-optimized pigeon flock algorithm simulates the collective behavior of pigeon flocks. By means of information exchange and cooperation between individuals, it jumps out of the local optimal solution through bifurcation behavior to explore the global optimal solution, significantly improving the cluster cooperation and task execution efficiency of the multi-drone system. When a drone fails, the bifurcation-optimized pigeon flock algorithm can optimize the control input of each drone and reduce the distance between adjacent drones, thereby enhancing the cooperative ability of the cluster and the stability of task execution.

[0046] In addition, while ensuring the fault tolerance ability, the fault-tolerant predictive control method of the present invention optimizes the computational efficiency and response speed. By introducing a fault prediction mechanism, the system can accurately predict in the early stage of a fault and adjust the control input according to the prediction result, avoiding the problems of hysteresis and overcorrection in traditional fault-tolerant control methods. At the same time, the present invention designs an error attenuation mechanism to ensure that the drone system can quickly recover after a fault, and the state error is attenuated in time and finally stabilized within the set target range. The introduction of the prediction performance function enables the system to adjust the control strategy in real time, avoiding problems such as overcorrection or insufficient control input, and ensuring that it can quickly respond and resume normal task execution when a fault occurs. This new method based on the bifurcation pigeon flock optimization algorithm and the fault-tolerant control strategy can not only improve the fault tolerance ability of the system, but also reduce the computational complexity, improve the response speed and adaptability of the system. Generally speaking, the present invention has significant technical advantages in the field of multi-drone cluster control and can greatly improve the robustness, cooperative operation ability and task execution efficiency of multi-drone systems. Description of the drawings

[0047] Figure 1 It is the flowchart of the fault-tolerant predictive control of drones optimized by simulating pigeon flock bifurcation.

[0048] Figure 2 It is a spatial graph of the consistent collaborative flight trajectories of three unmanned aerial vehicles optimized by imitating pigeon flock bifurcation.

[0049] Figure 3 It is a curve graph of the x-axis component error of three unmanned aerial vehicles optimized by imitating pigeon flock bifurcation.

[0050] Figure 4 It is a curve graph of the y-axis component error of three unmanned aerial vehicles optimized by imitating pigeon flock bifurcation.

[0051] Figure 5 It is a curve graph of the z-axis component error of three unmanned aerial vehicles optimized by imitating pigeon flock bifurcation.

[0052] Figure 6 It is a curve graph of the roll angle error of three unmanned aerial vehicles optimized by imitating pigeon flock bifurcation.

[0053] Figure 7 It is a curve graph of the pitch angle error of three unmanned aerial vehicles optimized by imitating pigeon flock bifurcation.

[0054] Figure 8 It is a curve graph of the heading angle error of three unmanned aerial vehicles optimized by imitating pigeon flock bifurcation.

[0055] The descriptions of the labels and symbols in the figure are as follows:

[0056] t - Simulation time

[0057] x - Horizontal direction component of the i-th unmanned aerial vehicle

[0058] y - Vertical direction component of the i-th unmanned aerial vehicle

[0059] z - Altitude of the i-th unmanned aerial vehicle

[0060] e i1 - Component error of the i-th unmanned aerial vehicle on the X-axis

[0061] e i2 - Component error of the i-th unmanned aerial vehicle on the Y-axis

[0062] e i3 - Component error of the i-th unmanned aerial vehicle on the Z-axis

[0063] e i4 - Component error of the roll angle of the i-th unmanned aerial vehicle

[0064] e i5 - Component error of the pitch angle of the i-th unmanned aerial vehicle

[0065] e i6 - Component error of the heading angle of the i-th unmanned aerial vehicle Specific implementation manner

[0066] See the simulation process framework inFigure 1 , the effectiveness of the UAV fault-tolerant predictive control method based on pigeon flock bifurcation optimization proposed by the present invention is verified by using Matlab programming below. The specific steps of this simulation verification process are as follows:

[0067] This simulation uses three UAVs numbered 1, 2, and 3, assuming that each UAV flies in three-dimensional space. The state of each UAV consists of position and velocity, and a total of 6 state variables are included:

[0068] [x i , y i , z i , φ i , θ i , ψ i . The simulation starts from the initial time step. The initial position of each UAV is X 1 (t) = (7, 18, 0, 0, 0, 0), X 2 (t) = (-4, -3, 0, 0, 0, 0), X 3 (t) = (12, 0, 0, 0, 0, 0), and the desired position The initial position is:

[0069] The goal of the system is to make each UAV reach the target position in a fixed flight shape (triangle), as shown in Figure 2 , and maintain collaborative work during flight. To this end, we adopt a method based on pigeon flock bifurcation optimization to optimize the control input of each UAV, so as to realize the cooperation of multiple UAVs.

[0070] This algorithm optimizes the control by simulating the group behavior of pigeon flocks. At each time step, all UAVs adjust their control inputs according to the current state and the states of neighboring UAVs. During the simulation process of the robust compensation parameters of the position controller, the regularization parameter λ = 0.01 and the performance function weight γ = 0.6 are used to ensure that both the rationality of the control input and the overall performance of the system are considered during the optimization process. To evaluate the effectiveness of the control strategy, we introduce three-axis state errors ex(t), e y (t), e z (t), and this error represents the gap between the current position and the target position of each UAV. The decay rate of the state error is adjusted by the decay factor σ i = 0.3. By decaying the error, it is ensured that when a fault occurs, the state of the system can gradually converge to the target position, thus ensuring the completion of the flight mission. During the simulation, through the update of each time step, the state error gradually decreases and finally approaches zero, as shown in Figures 3 - 5 .

[0071] In the simulation, we set different types of faults to test the fault tolerance of the system. Specifically, multiplicative faults and additive faults affect different UAVs at different time intervals. The multiplicative fault is manifested as the scaling of the control input, reflecting the performance degradation of the actuator; the additive fault is manifested as the deviation of the control input, simulating the damage or malfunction of the actuator. The specific parameters are as follows: the additive fault is η i (t) = 2*sin(0.1t), and the multiplicative fault is ψ i (t) = 1 + 0.1cos(0.1t). These fault injections occur in the following time intervals respectively: the first UAV has a 30% multiplicative fault between 1.5 - 2 seconds, then γ i = 0.3. These faults will affect the control input of the UAV, thus causing errors in the system state, as shown in Figure 6 . Among them, the adjustment coefficients (α 1 , β 1 ) = (0.2, 0.4), (α 2 , β 2 ) = (0.25, 0.72), (α 3 , β 3 ) = (0.35, 0.71). During the simulation process, the fault tolerance control strategy will adjust the control input of each UAV according to the real-time monitoring data and fault prediction results to ensure that the system can still operate stably when a fault occurs.

[0072] Within each simulation time step, through the optimization of the control input and the simulation of the fault impact, the system state is continuously updated. When a fault occurs, the fault tolerance control algorithm automatically adjusts the control input to compensate for the fault and ensures that the system continues to move stably towards the target direction. We introduce the pitch angle and heading angle state errors, which represent the gap between the current angle and the target angle of each UAV. The decay rate of the state error is adjusted by the decay factor ε. By decaying the error, it is ensured that when a fault occurs, the state of the system can gradually converge to the target position, thus ensuring the completion of the flight mission. In the simulation, through the update of each time step, the state error gradually decreases and finally approaches zero, as shown in Figures 7 - 8 .

[0073] The state (position and angle) of each UAV is calculated and compared with the target position to evaluate the effectiveness of the control strategy. At the end of the simulation, by plotting the trajectory diagrams of each UAV, the system behavior under the control input and fault impact can be clearly seen.

Claims

1. A fault-tolerant predictive control method for unmanned aerial vehicles based on pigeon flock bifurcation optimization, characterized in that: The method comprises the following steps: Step 1: Modeling the UAV flight control system; in which, the uncertain nonlinear function of the research object caused by external disturbances and parameter uncertainty is considered, multiplicative faults and additive faults are added, and the state of the entire system reaches the target position by designing the control input; Step 2: Design a pigeon-like bifurcation optimization algorithm to simulate the group behavior of pigeons. Through coordination and information exchange between individuals, optimize the control strategy of each drone, constrain the size of the control input, and achieve global goal optimization. Step 3: Design of prediction performance function: introduce the prediction performance function to measure the system state error, design the global error level definition, minimize the system state error, and ensure the effectiveness of fault-tolerant control; Step 4: First, introduce a correction mechanism based on fault bifurcation prediction information and adjust the control input through the optimization algorithm; adaptively adjust the strategy, design real-time updated fault information and performance function P(t), and judge whether the current control strategy meets the expected system performance through real-time evaluation of the predicted performance function; Step 5: Design a state feedback controller and implement global information coordinated adjustment, adjust and design the latest system control input optimization target u according to the state, distance and flight target information of the nearby drones i (t) to ensure the overall stability of the cluster; by dynamically adjusting the flight path and control strategy of each drone, the fault tolerance capability is further enhanced to ensure the system’s recovery speed and execution efficiency after a failure occurs.

2. The method according to claim 1, characterized in that: The specific process of step one is as follows: Establish a dynamic system consisting of N drones. The dynamic equation of each drone is: Where, X i (t) = [x i ,y i ,z i ,φ i ,θ i ,ψ i ]∈R is the state vector of the i-th drone, u i (t)∈R m is the control input, where R represents a one-dimensional scalar space, R m represents an m-dimensional vector space; f i (·) is an uncertain nonlinear function describing the research object caused by external disturbances and parameter uncertainty; the design goal of the control input is to obtain the control input u i (t) Make the state of the entire system reach the target position, or make the state error decay to zero; For each UAV’s control input, assume that the UAV’s actuator may be affected by additive and multiplicative faults, and the impact of the fault is represented by η i (t) and ψ i (t) to represent: In the formula, is the ideal control input, η i (t) represents multiplicative fault, ψ i (t) represents additive fault; the fault term η i (t) and ψ i (t) is an unknown time-varying function, which reflects the error of the UAV actuator during the flight.

3. The method according to claim 1, characterized in that: The specific process of step 2 is as follows: The goal of pigeon swarm bifurcation optimization is to dynamically adjust the control input u of each drone i (t), the expression is as follows: Where N i represents the set of UAVs adjacent to the i-th UAV, and λ is the regularization parameter. The purpose of this optimization problem is to minimize the distance between each UAV and its neighboring UAVs and constrain the size of the control input at the same time, so as to achieve global objective optimization.

4. The method according to claim 1, characterized in that: The specific process of step 3 is as follows: The prediction performance function P(t) reflects the global error level of the system and is defined as: in, is the target trajectory of each UAV; the goal of this function P(t) is to minimize the system state error and ensure that the state error can still be attenuated when a fault occurs; In order to ensure the stability and fault tolerance of the system, the error attenuation rate is designed to be: in, is the state error of the i-th UAV, and e i (t) = [e i1 (t),e i2 (t),e i3 (t),e i4 (t),e i5 (t),e i6 (t)], Δe i (t) is a correction term based on fault estimation; by adjusting the attenuation factor σ i ,ensuring that the state error of each drone gradually decays to the target range when a failure occurs; To ensure the effectiveness of fault-tolerant control, the following conditions must be met:

5. The method according to claim 1, characterized in that: The specific process of step 4 is as follows: Fault correction term Δu i The design of (t) is based on the following dynamic model: in It is the amount of information based on fault prediction, reflecting the change of system error caused by the fault; α i and β i To adjust the coefficient, determine the sensitivity of the correction term to the error and forecast correction; Fault correction term Δu i (t) The control input u that affects the UAV i (t), through dynamic adjustment, the system can return to a stable state as soon as possible after a fault occurs; the update formula of the control input is as follows: in, is the ideal control input in the absence of faults, and the fault correction term Δu i (t) is the part that is dynamically adjusted according to the fault correction information; in order to ensure the stability and constraints of the control input during the correction process, a weighting factor γ is introduced i To balance Δu i (t) The relationship between: Among them, γ i ∈[0,1] is an adjustment coefficient, which depends on the severity of the fault. In the whole process, the adaptive adjustment strategy is optimized by real-time feedback control. When the system detects a fault and predicts the impact of the fault, the bifurcation fault-tolerant control strategy automatically adjusts the control input of each UAV according to the real-time updated fault information and performance function P(t). Specifically, the control input of each UAV will be adaptively adjusted in the following cases: according to the severity of the fault detection, the correction amplitude and method of the control input are adjusted in real time. Through the real-time evaluation of the predicted performance function, it is judged whether the current control strategy meets the expected system performance. Through the bifurcation optimization pigeon flock algorithm, the control strategy is adjusted according to the status and behavior of the neighboring UAVs to achieve group collaboration.

6. The method according to claim 5, characterized in that: Assume that the fault severity of the system is S, and its value range is defined as [0,1]. When the value of S is between [0,0.3], it indicates a minor fault, and the correction of the control input is more conservative, γ(S) = γ min ; When the value of S is between [0.3,0.7], it indicates a moderate fault and the correction of the control input needs to be moderate, γ i =γ(S)=γ min +(S-0.3)(γ max -γ min ) / 0.4; when the value of S is between [0.7,1], it indicates a serious fault and the correction amplitude of the control input needs to be increased, γ(S)=γ max , where γ(S) is the adjustment function to quantify the degree of correction of the control input, γ min and γ max Respectively represent the minimum and maximum adjustment coefficients for fault correction.

7. The method according to claim 1, characterized in that: The specific process of step five is as follows: each UAV not only adjusts the control input according to its own state, but also adjusts it according to the state, distance and flight target information of the neighboring UAVs; For each UAV, the optimization objective of the system control input is modified as: Among them, N i represents the set of UAVs adjacent to the i-th UAV, κ is the newly introduced collaborative adjustment parameter, which represents the control coordination weight between adjacent UAVs, Represents the distance metric between the i-th and j-th drones.

Images