Double-loop active-disturbance-rejection trajectory tracking control method for quad-rotor unmanned aerial vehicle

Abstract

The invention discloses a dual-ring active-disturbance-rejection trajectory tracking control method for a four-rotor unmanned aerial vehicle, and the method comprises the steps: representing a dynamic coupling effect of the unmanned aerial vehicle caused by the load mass change and dynamic center-of-gravity shift through a six-degree-of-freedom flight dynamic model containing an asymmetric inertia matrix; an outer ring position control loop and an inner ring attitude control loop are constructed on the basis of an active disturbance rejection controller, and the following control cycles are executed in real time on the basis of integrating an extended state observer: the outer ring position control loop generates a position control signal and an attitude expected angle according to an expected track; the inner loop attitude control loop generates a control quantity for driving the unmanned aerial vehicle according to the attitude expected angle, and the two loops cooperatively work through decoupling logic to realize expected trajectory tracking; estimating and compensating unmodeled dynamics and external interference of the unmanned aerial vehicle in real time through the extended state observer; and dynamically adjusting parameters of the active-disturbance-rejection controller by adopting a particle swarm optimization algorithm.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of unmanned aerial vehicle control, and particularly relates to a double-loop active disturbance rejection trajectory tracking control method and system for a variable-load quadrotor unmanned aerial vehicle, which is suitable for high-precision trajectory tracking of an unmanned aerial vehicle carrying a dynamically changing load in a complex environment. BACKGROUND

[0002] In recent years, quadrotor unmanned aerial vehicles have been widely used in complex scenarios such as logistics transportation, agricultural plant protection, power inspection, and emergency rescue due to their flexibility and versatility. In these practical applications, unmanned aerial vehicles often need to carry variable loads (such as goods, sensors, and spraying devices), which leads to dynamic changes in their mass, inertia characteristics, and center of gravity position as the task progresses. For example, the mass of an agricultural unmanned aerial vehicle continuously decreases during pesticide spraying, and the center of gravity of a logistics unmanned aerial vehicle may suddenly change when loading or unloading goods. Such changes directly change the dynamic model parameters of the unmanned aerial vehicle (such as the increase in the asymmetry of the inertia matrix and the intensification of the coupling effect), posing a serious challenge to the precision and robustness of trajectory tracking control.

[0003] Traditional control methods have significant limitations when dealing with the above problems. Proportional-integral-derivative (PID) control is still the mainstream solution in the industry due to its simple structure and ease of implementation, but its parameter tuning relies on experience and is fixed, making it difficult to adapt to the time-varying dynamics caused by variable loads, and it is prone to problems such as large overshoot and response lag. Model predictive control (MPC) achieves dynamic adjustment through rolling optimization, but its performance is highly dependent on an accurate mathematical model, and when the load changes cause model mismatch, the control accuracy will decrease significantly, and the complex online calculation requirement is difficult to meet the real-time requirements of unmanned aerial vehicles. Although sliding mode control (SMC) has a certain robustness to parameter perturbations, the selection of switching gain needs to balance tracking accuracy and control chattering, and in the variable load scenario, chattering can easily cause wear and tear of the actuator, affecting system stability.

[0004] As an advanced control method independent of accurate model, active disturbance rejection control (ADRC) shows potential in nonlinear and strongly coupled systems by estimating and compensating internal and external disturbances in real time through extended state observer (ESO). However, the parameter tuning problem of ADRC has long restricted its engineering application: the double-loop control architecture (position loop + attitude loop) of quadrotor UAV usually involves multiple parameters such as observer bandwidth, controller bandwidth, and nonlinear gain, and it is difficult to achieve global optimization of multiple parameters by traditional trial-and-error method or empirical method; although existing intelligent optimization algorithms (such as genetic algorithm and particle swarm optimization) can be used for parameter optimization, they are mostly in offline tuning mode and cannot dynamically respond to rapid changes in dynamic characteristics caused by sudden load changes. In addition, in the presence of coupling effects caused by asymmetric inertia matrix of UAV in variable load scenarios, existing ADRC schemes mostly use the assumption of simplified symmetric inertia matrix, which cannot accurately represent the system dynamics, further limiting the improvement of tracking accuracy.

[0005] In summary, the existing control methods generally face problems such as insufficient model adaptability, disturbance compensation lag, and complex parameter tuning in the trajectory tracking task of variable load quadrotor UAV. How to construct a control strategy that can adapt to load changes in real time and efficiently suppress dynamic coupling and external disturbances has become a technical bottleneck that needs to be solved in the current UAV control field. SUMMARY

[0006] In view of the defects and deficiencies of the prior art, the present application provides a double-loop active disturbance rejection trajectory tracking control method and system for variable load quadrotor UAV, aiming to solve the problems of dynamic characteristic changes and insufficient control accuracy caused by load mass changes and dynamic center of gravity shifts of UAV in variable load scenarios, and to overcome the bottlenecks of weak anti-coupling and anti-disturbance ability of traditional control methods and difficult parameter tuning of active disturbance rejection controller (ADRC).

[0007] The present application first quantifies the dynamic coupling effects caused by load mass changes and dynamic center of gravity shifts through a six-degree-of-freedom flight dynamic model containing an asymmetric inertia matrix, providing a dynamic basis for subsequent control that fits actual working conditions; under the support of this model, a double-loop active disturbance rejection control architecture is constructed, in which the outer loop position control loop generates a position control signal and an attitude expectation angle according to the desired trajectory, and the inner loop attitude control loop generates a control amount to drive the UAV according to the attitude expectation angle, the two loops work cooperatively through decoupling logic to independently adjust the position and attitude of the UAV, effectively avoiding dynamic coupling interference between position and attitude. At the same time, the double-loop active disturbance rejection control architecture integrates an extended state observer, which can estimate and compensate unmodeled dynamics and external disturbances in the flight of the UAV in real time, further enhancing the anti-disturbance robustness of the system.

[0008] To adapt to the sudden load change, the particle swarm optimization algorithm is used to perform online self-tuning on the parameters of the dual-loop active disturbance rejection controller: a control performance evaluation index is formed by the trajectory tracking error with time weight and the standard deviation of the attitude angle (the longer the time, the greater the influence of the tracking error on the index, and the standard deviation of the attitude angle is used to quantify the stability of the flight attitude), and the parameter adjustment is started by combining the preset minimum trigger interval and the error threshold event trigger mechanism, which avoids Zeno phenomenon and ensures the real-time performance of parameter adaptation; during the parameter tuning process, the bandwidth method is used to simplify the operation, and a large number of parameters to be tuned are simplified to a small number of key bandwidth parameters (two for each of the attitude control loop and the position control loop), thereby reducing the parameter tuning complexity. In addition, the outer loop position control loop integrates the tracking differentiator to smooth the desired trajectory and suppress the command mutation impact, and the inner loop attitude control loop omits the tracking differentiator of the roll channel and the pitch channel, thereby simplifying the control complexity while ensuring the attitude response speed.

[0009] Through the above design, the unmanned aerial vehicle can realize high-precision trajectory tracking under variable load conditions, so that the position error of the unmanned aerial vehicle converges to a tolerance range of not more than 0.3 m within a preset time, thereby providing a stable and reliable control scheme for the four-rotor unmanned aerial vehicle operation under complex environment and carrying dynamic load.

[0010] The application specifically adopts the following technical solutions:

[0011] A dual-loop active disturbance rejection trajectory tracking control method for a four-rotor unmanned aerial vehicle, which represents the dynamic coupling effect of the unmanned aerial vehicle caused by the change of load mass and the dynamic offset of the center of gravity through a six-degree-of-freedom flight dynamic model with an asymmetric inertia matrix, and constructs an outer loop position control loop and an inner loop attitude control loop based on an active disturbance rejection controller, and realizes the following control cycle in real time on the basis of integrating an extended state observer:

[0012] (1) The outer loop position control loop generates a position control signal and an attitude expected angle according to the desired trajectory, the inner loop attitude control loop generates a control amount for driving the unmanned aerial vehicle according to the attitude expected angle, and the two loops work cooperatively through decoupling logic to realize desired trajectory tracking;

[0013] (2) The unmodeled dynamics and external disturbances of the unmanned aerial vehicle are estimated and compensated in real time through the extended state observer;

[0014] (3) The particle swarm optimization algorithm is used to dynamically adjust the parameters of the active disturbance rejection controller according to the control performance index formed by the trajectory tracking error with time weight of the unmanned aerial vehicle and the standard deviation of the attitude angle, and the event trigger mechanism with the preset minimum trigger interval and the error threshold, so as to adaptively compensate for the change of the dynamic characteristics caused by the change of the load mass and the dynamic offset of the center of gravity.

[0015] Further, the decoupling logic of the outer loop position control loop and the inner loop attitude control loop is that the outer loop position control loop is only responsible for adjusting the displacement of the UAV in the X-axis, Y-axis and Z-axis directions, and the inner loop attitude control loop is only responsible for adjusting the roll angle, pitch angle and yaw angle of the UAV, and the two loops are coordinated through the signal transmission path of the outer loop output attitude expected angle and the inner loop received and executed to avoid dynamic coupling interference of position and attitude.

[0016] Further, the disturbance compensation process of the extended state observer is that the position, speed and total disturbance containing unmodeled dynamics and external disturbance of the UAV are collectively estimated as state variables in real time, wherein the total disturbance is reconstructed as an independent observation variable through a state expansion mechanism, and the observation variable is fed forward to the control end of the dual-loop active disturbance rejection controller to realize active compensation of the disturbance.

[0017] Further, in the control performance index composed of the trajectory tracking error with time weight and the attitude angle standard deviation, the time weight has the following mechanism: the longer the tracking error lasts, the greater the influence on the calculation result of the control performance index; and the attitude angle standard deviation is the sum of the roll angle standard deviation, the pitch angle standard deviation and the yaw angle standard deviation, which is used to quantify the stability of the flight attitude of the UAV.

[0018] Further, the preset minimum triggering interval of the event triggering mechanism is a cooling time to avoid Zeno phenomenon, and the error threshold is a pre-set maximum allowed value of the trajectory tracking error; when the real-time trajectory tracking error of the UAV does not exceed the error threshold, or the time interval of two parameter adjustments does not reach the preset minimum triggering interval, the particle swarm optimization algorithm does not start parameter adjustment.

[0019] Further, the outer loop position control loop integrates a tracking differentiator, which smoothes the input expected trajectory signal and generates a corresponding speed differential signal to suppress the impact on the UAV when the expected trajectory instruction suddenly changes; and the inner loop attitude control loop omits the tracking differentiators of the roll channel and the pitch channel, and directly generates an attitude control signal according to the attitude expected angle.

[0020] Further, the particle swarm optimization algorithm adopts a linearly decreasing inertia weight strategy to adjust the search ability of particles: the inertia weight decreases linearly from an initial value to a final value with the number of iterations of the algorithm.

[0021] Further, the inner loop attitude control loop generates driving control amounts corresponding to the adjustment signals of the roll angle, the pitch angle and the yaw angle of the unmanned aerial vehicle respectively; and the outer loop position control loop calculates a total virtual control amount by combining the control amount in the Z-axis direction and the trigonometric function values of the current roll angle and the pitch angle of the unmanned aerial vehicle, so as to obtain the execution logic of the position control and the attitude control.

[0022] Further, when the particle swarm optimization algorithm adjusts the parameters of the active disturbance rejection controller, a bandwidth method is used to simplify the parameter setting process: the observer gain and the observer bandwidth of the active disturbance rejection controller are respectively established in linear, quadratic and cubic correlation, and the controller gain and the controller bandwidth are respectively established in quadratic and linear correlation, so that the number of parameters to be set is simplified from 36 to 2 bandwidth parameters of the attitude control loop and 2 bandwidth parameters of the position control loop.

[0023] Further, the particle swarm optimization algorithm adjusts the parameters of the active disturbance rejection controller, and a bandwidth method is used to simplify the parameter setting process: the observer gain and the observer bandwidth of the active disturbance rejection controller are respectively established in linear, quadratic and cubic correlation, and the controller gain and the controller bandwidth are respectively established in quadratic and linear correlation, so that the number of parameters to be set is simplified from 36 to 2 bandwidth parameters of the attitude control loop and 2 bandwidth parameters of the position control loop.

[0024] The model representation module is configured to represent the dynamic coupling effect of the unmanned aerial vehicle caused by the change of the load mass and the dynamic offset of the center of gravity through a six-degree-of-freedom flight dynamic model containing an asymmetric inertia matrix.

[0025] The double-loop active disturbance rejection control module includes an outer loop position control submodule and an inner loop attitude control submodule. The double-loop active disturbance rejection control module receives the six-degree-of-freedom flight dynamic model parameters output by the model representation module. The outer loop position control submodule is configured to generate a position control signal and an attitude expectation angle according to the expected trajectory. The inner loop attitude control submodule is configured to generate a control amount for driving the unmanned aerial vehicle according to the attitude expectation angle. The two submodules work cooperatively through a decoupling logic to achieve the expected trajectory tracking.

[0026] The disturbance compensation module integrates an extended state observer and is configured to receive the control feedback signal of the double-loop active disturbance rejection control module, estimate and compensate the unmodeled dynamics and external disturbances of the unmanned aerial vehicle in real time, and feed back the disturbance compensation signal to the double-loop active disturbance rejection control module.

[0027] The parameter self-setting module is configured to receive the control performance feedback signal of the double-loop active disturbance rejection control module, adopt a particle swarm optimization algorithm, combine a preset minimum trigger interval and an error threshold event trigger mechanism according to a control performance index composed of the trajectory tracking error of the unmanned aerial vehicle containing a time weight and the standard deviation of the attitude angle, and dynamically adjust the parameters of the active disturbance rejection controller in the double-loop active disturbance rejection control module, so as to adaptively compensate for the change of the kinetic characteristics caused by the change of the load mass and the dynamic offset of the center of gravity.

[0028] And an electronic device comprising a memory, a processor and a computer program stored on the memory and loadable on the processor, characterized in that the processor implements the steps of the method as described above when executing the program.

[0029] A non-transitory computer-readable storage medium having stored thereon a computer program, characterized in that the computer program, when executed by a processor, implements the steps of the method as described above.

[0030] Compared with the prior art, the application and its preferred schemes first, through the six-degree-of-freedom flight dynamic model containing the asymmetric inertia matrix, the application can accurately quantify the dynamic coupling effect caused by the change of load mass and the dynamic offset of the center of gravity, solve the problem that the conventional model in the prior art is difficult to adapt to the variable load working condition and the inaccurate dynamics characterization, provide a reliable foundation for the subsequent control which fits the actual flight characteristics, and ensure that the control strategy can respond to the dynamics change caused by the variable load; secondly, the double-loop active disturbance rejection control architecture realizes the independent operation of position regulation and attitude regulation through the decoupling cooperation of the outer loop position control loop and the inner loop attitude control loop, effectively avoids the problems of tracking accuracy decline and attitude oscillation caused by the mutual coupling interference of position and attitude in the prior art, and further improves the stability of trajectory tracking and the accuracy of attitude control relying on the characteristics of the active disturbance rejection control; thirdly, the integrated application of the extended state observer can estimate and compensate the unmodeled dynamics and external disturbances in the flight process of the unmanned aerial vehicle in real time, overcome the defects of weak anti-interference ability and difficulty in dealing with complex environmental disturbances of the traditional control method, and significantly enhance the robustness of the system under variable working conditions; in addition, the parameter online self-tuning driven by the particle swarm optimization algorithm can dynamically respond to sudden load changes, automatically adjust the parameters of the double-loop active disturbance rejection controller, solve the bottleneck that the existing active disturbance rejection controller relies on manual parameter setting and fixed parameters cannot adapt to variable load working conditions, and greatly improve the adaptability of the system to dynamic load.

[0031] Meanwhile, the differential configuration of the tracking differentiator (the outer loop position control loop integrates the tracking differentiator, and the inner loop attitude control loop omits the tracking differentiator of the roll and pitch channels) realizes the smooth processing of the expected trajectory and suppresses the command mutation impact through the tracking differentiator, reduces the control complexity through the simplification of the inner loop structure, and guarantees the response speed of the attitude control; the application of the bandwidth method in parameter setting simplifies a large number of to-be-set parameters into a small number of key bandwidth parameters, reduces the operation difficulty of parameter setting in engineering practice, and improves the efficiency of parameter setting; the introduction of the event triggering mechanism effectively avoids the Zeno phenomenon through the preset minimum triggering interval and error threshold, ensures the real-time of parameter setting, avoids unnecessary parameter adjustment operation, and reduces resource consumption; the adoption of the linear decreasing inertia weight strategy in the particle swarm optimization algorithm further balances the global exploration ability and local convergence ability in the parameter search process, is helpful to quickly find better control parameters, and provides support for the stable performance of the controller. BRIEF DESCRIPTION OF DRAWINGS

[0032] The application will be described in further detail below with reference to the drawings and specific embodiments:

[0033] Figure 1 A double-loop active disturbance rejection control framework of a variable load four-rotor wing of an embodiment of the application;

[0034] Figure 2 A structure block diagram of an active disturbance rejection controller of a variable load four-rotor wing of an embodiment of the application;

[0035] Figure 3 A controller parameter online self-setting flowchart of a variable load four-rotor wing of an embodiment of the application;

[0036] Figure 4 A general control flowchart of an embodiment of the application. DETAILED DESCRIPTION

[0037] In the following, specific embodiments of the present application will be described in detail with reference to the drawings, and according to these detailed descriptions, those skilled in the art can clearly understand the present application and implement the present application. The features in each different embodiment can be combined to obtain new implementation modes, or replace some features in some embodiments to obtain other preferred implementation modes, without departing from the principles of the present application.

[0038] In order to make the features and advantages of the present application more obvious and easy to understand, the following specific embodiments are described in detail below, and the drawings are used for description as follows:

[0039] The application is directed to the trajectory tracking control problem of variable load quadrotor unmanned aerial vehicle, and a design scheme based on a double-loop active disturbance rejection controller is proposed. The outer loop position controller and the inner loop attitude controller work cooperatively through a cascade structure to realize accurate control of the position and attitude of the unmanned aerial vehicle. In addition, a particle swarm optimization algorithm is introduced to perform online self-tuning of the controller parameters, further improving the dynamic performance and adaptability of the system. Through theoretical analysis and simulation verification, the method proposed in the application shows good trajectory tracking performance and anti-interference ability under no load and variable load conditions, providing an effective solution for unmanned aerial vehicle control in complex environments.

[0040] For the physical structure and motion characteristics of the quadrotor unmanned aerial vehicle, the displacement control of the unmanned aerial vehicle in the x-axis and y-axis directions needs to be realized by adjusting the attitude angle of the unmanned aerial vehicle. Based on this characteristic, the application proposes a cascade active disturbance rejection control scheme, and the controller of the variable load quadrotor unmanned aerial vehicle is composed of an outer loop position controller and an inner loop attitude controller. The basic idea is to calculate the control variable and the desired angle , through the position controller, then give the control variable , , through the attitude controller, and finally calculate the lift provided by the four motors. In the design of the controller, the application also introduces a particle swarm optimization algorithm to realize online self-tuning of the controller parameters, achieving better control effect. The active disturbance rejection controller mainly consists of three parts: tracking differentiator, extended state observer, and state error feedback. The tracking differentiator is the processing link of the target signal, which can arrange the transition process and obtain the differential signal. The extended state observer observes and compensates the total disturbance in real time, and through parameter tuning, it can achieve good observation and compensation effect. The state error feedback obtains the error signal, which is combined using a function to calculate the control variable and obtain good control effect. The active disturbance rejection controller does not need the accurate model of the controlled object, simplifying the complex control problem, so that the controlled system can stably complete the target in a complex and variable environment.

[0041] As Figure 4As shown, the present application discloses a design scheme of a high-performance dual-loop active disturbance rejection control (ADRC) system for trajectory tracking of variable-load quadrotor unmanned aerial vehicles (UAVs), aiming to solve the control challenges caused by changes in load mass and dynamic shifts in the center of gravity in actual tasks, and is suitable for high-precision trajectory tracking of unmanned aerial vehicles carrying dynamically changing loads in complex environments. The system establishes a precise six-degree-of-freedom flight dynamic model, fully considering the influence of variable loads on the dynamic characteristics of quadrotors, and introduces an asymmetric inertia matrix to quantify the complex coupling effects caused by load changes. The control system adopts a dual-loop structure, including an outer loop position controller and an inner loop attitude controller, which respectively realize accurate adjustment of the position and attitude of the quadrotor through decoupling, to ensure stability and tracking accuracy in complex dynamic environments. To adapt to sudden load changes, the present application proposes an online parameter self-tuning method based on a particle swarm optimization algorithm, which can adjust control parameters in real time and significantly improve the adaptability and robustness of the system. In addition, the system integrates an extended state observer for real-time estimation and compensation of unmodeled dynamics and external disturbances, further enhancing control performance.

[0042] The composition of the system includes:

[0043] a) a six-degree-of-freedom flight dynamic model that considers changes in load mass and dynamic shifts in the center of gravity, wherein an asymmetric inertia matrix is used to quantify the coupling effects caused by variable loads; the six-degree-of-freedom flight dynamic model considers the dynamic coupling effects caused by variable loads through an asymmetric inertia matrix, which can accurately characterize the dynamic behavior of quadrotors under different load conditions.

[0044] b) an outer loop position controller configured to adjust the position of the quadrotor; the outer loop position controller generates a position control signal based on real-time position feedback to maintain stable trajectory tracking.

[0045] c) an inner loop attitude controller configured to adjust the attitude of the quadrotor, wherein the outer loop and inner loop controllers operate in a decoupled manner to control the dynamic characteristics of the quadrotor; the inner loop attitude controller generates an attitude control signal based on real-time attitude feedback to ensure accurate attitude control.

[0046] d) a particle swarm optimization algorithm configured to adjust the control parameters of the outer loop and inner loop controllers online to adapt to sudden load changes. The particle swarm optimization algorithm dynamically adjusts the control parameters of the ADRC system in response to changes in load mass or the center of gravity, optimizing control performance under dynamic conditions.

[0047] The above system also includes an extended state observer integrated into the dual-loop ADRC framework, configured to estimate and compensate for unmodeled dynamics and external disturbances in real time.

[0048] The corresponding control process includes the following steps:

[0049] a) Establish a six-degree-of-freedom flight dynamic model that takes into account load mass variation and center of gravity dynamic shift, and uses an asymmetric inertia matrix;

[0050] b) Implement an outer loop position controller to regulate the quadrotor's position;

[0051] c) Implement an inner loop attitude controller to regulate the quadrotor's attitude;

[0052] d) Decouple the position and attitude control loops to independently manage the quadrotor's dynamic characteristics;

[0053] e) Apply a particle swarm optimization algorithm for online adjustment of control parameters to adapt to sudden load changes.

[0054] Wherein the particle swarm optimization algorithm continuously monitors load changes and adjusts control parameters to maintain stable trajectory tracking under dynamic load conditions.

[0055] Also included is the step of using an extended state observer to estimate and compensate for unmodeled dynamics and external disturbances to enhance the robustness of the control system.

[0056] The above control system ensures stable trajectory tracking performance under sudden load changes.

[0057] The embodiments of the present application provide a control method for wireless charging of a patrol unmanned aerial vehicle, which is used to ensure that the unmanned aerial vehicle enters the effective charging distance of a charging base station within a preset time, and has strong robustness.

[0058] The method of the present application can be applied to servers, terminals or other devices with logic processing capability, and the present application is not limited in this regard. For convenience of description, the following will be described by taking the server as an example.

[0059] The first aspect of the embodiments of the present application provides a position controller design for a quadrotor unmanned aerial vehicle, comprising:

[0060] An active disturbance rejection controller (ADRC) based on real-time disturbance dynamic compensation is proposed, which solves the control accuracy and robustness problem of the controlled object under unknown disturbance through a multi-module collaborative mechanism. The core structure contains three-stage closed loop: trajectory preprocessing and state tracking, tracking differentiator (TD) to the expected trajectory signal Smooth and differential processing to generate a transition trajectory and speed signal without overshoot, to suppress the impact of command mutation; the extended state observer (ESO) collects the output x of the controlled object in real time, reconstructs the system state quantity , and the key extended disturbance term , which integrates external disturbances Integration with model uncertainty; State Error Feedback (SEF) module , and , Generate basic control variables The disturbance dynamic compensator will Feedforward to the control terminal to output the final control quantity .

[0061] The second aspect of this application provides an attitude controller design for a quadcopter unmanned aerial vehicle, including:

[0062] The actual displacement change of the quadcopter is closely related to the attitude angle change of the UAV. To simplify the complexity of the inner loop controller and improve overall control efficiency, this application omits the tracking differentiator in the roll and pitch channels. Channels and In the channel, control variables U2 and U3 are determined by the expected angle value. and This is derived from the controller.

[0063] The third aspect of this application provides online self-tuning of controller parameters based on particle swarm optimization, including:

[0064] Particle swarm optimization, bandwidth optimization, fitness function, and event triggering mechanism;

[0065] In the particle swarm optimization algorithm, each particle has a position and a velocity V. They fly through the search space and continuously update their position and velocity with each iteration. By tracking the optimal position of an individual particle and the global optimal position, the particles can gradually approach the optimal solution to the problem.

[0066] Each active disturbance rejection control (ADRC) has six important parameters to be tuned [k1,k2,β1,β2,β3,b], therefore the entire inner-loop attitude and outer-loop position control system has a total of 36 adjustable parameters. To simplify the parameter tuning process, this invention employs the bandwidth method, introducing two key bandwidth concepts: observer bandwidth ω. o and controller bandwidth ω c .

[0067] In the particle swarm algorithm, the fitness function is an important performance index for evaluating the mass of particles, which directly determines the search direction and the final optimization result of the algorithm. A proper fitness function can guide the particles to fly in the direction of better solutions, so as to quickly find the optimal solution. The present application designs a fitness function according to the trajectory tracking error and the standard deviation of attitude stability. In the design of this fitness function, time is regarded as the weight of error. The longer the time is, the greater the influence of tracking error on the calculation result of the fitness function is. Under this design of the fitness function, the smaller the calculated value is, the more effective the controller is.

[0068] The fourth aspect of the embodiments of the present application provides controller stability analysis, including convergence analysis of the extended state observer and stability analysis of the controller.

[0069] The convergence analysis of the extended state observer takes the z channel as an example, and the state equation thereof can be expressed as:

[0070] (Formula 1)

[0071] Let represent the observation of , and the extended state observer of the system can be constructed as:

[0072] (Formula 2)

[0073] Define , , and the following can be obtained:

[0074] (Formula 3)

[0075] After arrangement, the following is obtained:

[0076] (Formula 4)

[0077] According to assumption 1: the total disturbance and its derivative are bounded, and , and theorem 1: there is a constant , there is a finite time , and , , the following can be obtained: The maximum value thereof is related to the upper bound of the initial error and the total disturbance of the system. When ω o is large enough, the observer error can converge to approximately 0.

[0078] The stability analysis of the controller first gives assumption 2: the given reference signal v and its derivative and are bounded.

[0079] are represented by variables v1, v2 and v3 respectively, and , the controller can be represented as:

[0080] (Equation 5)

[0081] Introduce Lemma 1: Consider a system in the form of where , N is a real matrix. When N is a Hurwitz matrix and , it can be obtained that . .

[0082] According to assumptions 1, 2 and Lemma 1, the stability under the closed loop system can be proved, according to the bandwidth method, when the bandwidth is large enough, the error is approximately tends to 0.

[0083] The control scheme embodiment in the scheme of the present application will be further described below in combination with the drawings.

[0084] Please refer to Figure 1 , one embodiment of the control method for wireless charging of the inspection unmanned aerial vehicle in the present application comprises:

[0085] The basic idea is that the control variable U1 and the expected angle θd, ϕd are calculated by the position controller, then the control variables U2, U3, U4 are given by the attitude controller, and finally the lift provided by the four motors is calculated. In the design of the controller, the particle swarm optimization algorithm is introduced to realize the online self-tuning of the controller parameters, so as to achieve better control effect. The active disturbance rejection controller mainly consists of three parts: tracking differentiator, extended state observer and state error feedback. The tracking differentiator is the processing link of the target signal, which can arrange the transition process and obtain the differential signal. The extended state observer observes and compensates the total disturbance in real time, and through parameter tuning, better observation and compensation effect can be achieved. The state error feedback obtains the error signal, which is combined by using a function, and the control variable is calculated to obtain good control effect.

[0086] Please refer to Figure 2 , one embodiment of the control device for wireless charging of the inspection unmanned aerial vehicle in the present application comprises:

[0087] In the design of the outer loop position controller, taking the displacement x channel as an example, according to the expected displacement x d , the virtual control variable is calculated. By combining the displacement motion equation and the rotation motion equation, the control model of the variable load quadrotor unmanned aerial vehicle can be obtained as:

[0088]

[0089] ( Formula 6 )

[0090] According to the UAV model, the channel is a second-order system, and its dynamic model can be rewritten as:

[0091] ( Formula 7 )

[0092] wherein , , are state variables of the system, defined as:

[0093] The tracking error of the traditional control system is written in a way that is prone to overshoot. In order to reasonably obtain the error signal, a reasonable transition process should be arranged , and the error is taken as:

[0094] ( Formula 8 )

[0095] This can reduce overshoot while minimizing the impact on rapidity, effectively improving the robustness of the system. In this application, the tracking differentiator is designed as follows:

[0096] ( Formula 9 )

[0097] wherein is the trajectory tracking signal, is the trajectory differential signal, r and h are two positive adjustable parameters, r is a speed factor, h is a filter factor, is a "fast control optimal synthesis function", and its expression is:

[0098]

[0099] ( Formula 10 )

[0100] The tracking differentiator is essentially a signal processing link that can filter the input signal and effectively suppress overshoot during the transition process and adjust the tracking speed. The overshoot during the transition process and the tracking speed can be adjusted by adjusting the parameters r and h.

[0101] To obtain better tracking performance, an extended state observer is designed to realize real-time estimation of the system state:

[0102] ( Formula 11 )

[0103] wherein is the observation error of the extended state controller, , 、 are the observation values of 、 、 respectively, 、 、 are the observer gains to be tuned, are the controller gains to be tuned.

[0104] Through the design of the extended state observer, the controller realizes real-time accurate estimation of the system state (including position, speed, acceleration) and unknown disturbance, thereby significantly improving the tracking performance and anti-interference ability of the system.

[0105] The virtual control variable U x is designed as:

[0106] (Formula 12)

[0107] where k x1 , k x2 are the controller gains to be tuned.

[0108] The control design of the y channel in the position controller and the z channel is the same as that of the x channel. In these two channels, the control variables U y and U z are obtained by the controller calculation from the corresponding displacement expectation values y d and z d respectively. The controller design is as follows:

[0109] (Formula 13)

[0110] (Formula 14)

[0111] Through the control variable U z obtained in the z channel, the controller can calculate the virtual control variable U1 according to the following formula:

[0112] (Formula 15)

[0113] Please refer to Figure 3 , another embodiment of the control device for wireless charging of the inspection unmanned aerial vehicle in the embodiment of the application includes:

[0114] Based on the particle swarm algorithm, the application proposes a parameter online self-tuning active disturbance rejection control strategy, which simplifies the parameter tuning process. The research design adopts the bandwidth method and introduces two key bandwidth concepts: observer bandwidth and controller bandwidth Based on the bandwidth method, the systematic tuning of parameters can be achieved by the following relationship:

[0115] (Formula 16)

[0116] Suppose the parameter b = 0.5, in the attitude control loop, the control objectives of the three attitude controllers have similar orders of magnitude; similarly, the orders of magnitude of the three displacement controllers in the position control loop also remain consistent. Therefore, the same parameter set can be used in each control loop. After simplification, the total number of parameters to be tuned is reduced from 36 to 4, of which 2 are used for attitude control and 2 are used for position control, which can guarantee the effectiveness of control. The parameters that need to be adjusted are In the process of parameter tuning of the quadrotor controller, the six controllers work as a coupled system to the controlled target, and the control amount obtained by the unmanned aerial vehicle is closely linked, and the control amount cannot be directly obtained from the target trajectory. In view of the inherent strong coupling of the quadrotor aircraft, the translational displacement is directly affected by the rotational attitude, thereby restricting the tuning of a single sub-controller, resulting in the inability to independently tune the parameters of the six sub-controllers. For example, when the yaw angle is insufficient due to the poor performance of the attitude controller, which leads to a decline in the tracking performance of the x-axis. At this time, optimizing the parameters of the x-channel controller alone cannot effectively improve the tracking performance.

[0117] In the particle swarm algorithm, each particle updates its position and velocity according to the following rules:

[0118] (Formula 17)

[0119] where i represents the i-th particle in the particle swarm, represents the velocity of the i-th particle at the t-th iteration, represents the position at the t-th iteration, represents the historical best position of a single particle, represents the global best position, w represents the inertia weight, c1 and c2 represent the acceleration constant, and r1 and r2 are random numbers between 0 and 1. Through iterative updating of the particle position and velocity, the particle swarm optimization algorithm can effectively explore the solution space. The inertia weight w is an important parameter that determines the inertia influence of the historical velocity of the particle when updating the speed. Generally speaking, a larger w value is beneficial to global search, enabling the particle to search a larger range in the search space. In contrast, a smaller w value is more biased towards local search, which helps the particle converge to the optimal solution more quickly. Therefore, in the particle swarm algorithm, by adjusting the value of the inertia weight w, a proper solution can be obtained in different optimization problems. In the particle swarm search process of the present application, a linear decreasing inertia weight (LDW) strategy is introduced to dynamically adjust w. The updating method of w is:

[0120] (18)

[0121] where, and are the initial and final values of the inertia weight, respectively, t represents the current iteration number, t max represents the maximum number of iterations. This algorithm dynamically adjusts the parameter w through linearly decreasing method, maintaining the balance between global exploration and local exploration throughout the optimization process.

[0122] In the particle swarm algorithm, the fitness function is an important performance indicator for evaluating the quality of particles, which directly determines the search direction and final optimization result of the algorithm. A proper fitness function can guide the particles to fly in the direction of better solutions, thereby quickly finding the optimal solution. In addition, the complexity and computational efficiency of the fitness function also need to be considered. Too complex a function will increase the amount of calculation and affect the efficiency of particle swarm optimization. Therefore, when selecting the fitness function, the complexity and optimization effect need to be balanced.

[0123] Under the target of trajectory tracking, the tracking error of the target trajectory by the UAV is the key to evaluate the control performance. Through the monitoring and analysis of the tracking error, the control effect of the parameters can be effectively evaluated. For the quadrotor system running in three-dimensional space, the trajectory tracking error is defined as:

[0124] (19)

[0125] In addition to the trajectory tracking error, the stability of the attitude is also very important during the flight of the quadrotor. In order to quantify the degree of swing of the attitude angle, the standard deviation is introduced as an indicator. Taking the roll angle as an example, the standard deviation expression is:

[0126] (20)

[0127] where n is the number of samples, is the sample mean.

[0128] In summary, the fitness function designed in this application is:

[0129] (21)

[0130] where a and b are positive parameters , and The standard deviations of roll angle, pitch angle and yaw angle, respectively. In the design of the fitness function, time is regarded as the weight of error. The longer the time, the greater the influence of tracking error on the result of fitness function calculation. Under the design of this fitness function, the smaller the calculated Y value, the more effective the controller.

[0131] In order to realize the online self-tuning of the particle swarm algorithm in the controller, the following event trigger mechanism is designed:

[0132] (Formula 22)

[0133] Where, t p Indicates the triggering moment of parameter adaptation, t p+1 Indicates the next triggering moment; t indicates the minimum triggering interval, which is equivalent to the cooling time, Is the fixed threshold of the designed event trigger. Due to the existence of the triggering interval, the event trigger will not cause Zeno phenomenon.

[0134] The above control scheme provided by the embodiment of the application estimates the mass mutation disturbance, center of gravity offset coupling effect and unmodeled dynamic characteristics faced by the variable load four-rotor unmanned aerial vehicle in complex flight environment in real time by using the cascade extended state observer. In a large range of logistics transportation tasks, in the face of the scene of real-time interference of load dynamic change (such as cargo delivery) and external wind disturbance, the designed double-loop active disturbance rejection controller can still guarantee the full-time trajectory tracking accuracy and attitude stability of the unmanned aerial vehicle system, and can realize that the unmanned aerial vehicle converges to the tolerance range of the target trajectory (position error ≤0.3m) within the preset time, and has strong robustness.

[0135] Based on the same inventive concept, the application further provides a computer device, which comprises one or more processors and a memory for storing one or more computer programs; the program comprises program instructions, and the processor is configured to execute the program instructions stored in the memory. The processor can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc., which are the computing core and control core of the terminal, and are configured to implement one or more instructions, and are specifically configured to load and execute one or more instructions in the computer storage medium to implement the above method.

[0136] It should be further explained that based on the same inventive concept, the present application also provides a computer storage medium, which stores a computer program, and the computer program is executed by a processor to perform the above method. The storage medium can adopt any combination of one or more computer readable media. The computer readable medium can be a computer readable signal medium or a computer readable storage medium. The computer readable storage medium may, for example, but is not limited to, an electrical, magnetic, optical, infrared, or semiconductor system, device or apparatus, or any combination of the above. More specific examples (non-exhaustive list) of the computer readable storage medium include: electrical connections having one or more wires, portable computer disks, hard drives, random access memories (RAM), read-only memories (ROM), erasable programmable read-only memories (EPROM or flash memory), optical fibers, portable compact disk read-only memories (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the above. In the present application, the computer readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, device or apparatus.

[0137] In the description of the present application, the description of the terms "one embodiment", "an example", "a specific example" and the like means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present disclosure. In the present specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0138] The basic principles, main features and advantages of the present disclosure are shown and described above. It should be understood by those skilled in the art that the present disclosure is not limited by the above embodiments, and the above embodiments and descriptions in the specification are only to illustrate the principles of the present disclosure. Without departing from the spirit and scope of the present disclosure, various changes and improvements can be made to the present disclosure, and these changes and improvements all fall within the scope of the claimed present disclosure.

[0139] The present application is not limited to the above best mode, and anyone can derive other various forms of a quadrotor unmanned aerial vehicle double-loop active disturbance rejection trajectory tracking control method under the inspiration of the present application. Any equivalent changes and modifications made within the scope of the present application application patent range shall fall within the scope of the present application.

Claims

1. A quadrotor UAV dual-loop active disturbance rejection trajectory tracking control method, characterized in that: a six-degree-of-freedom flight dynamic model containing an asymmetric inertia matrix is used to represent the dynamic coupling effect of the UAV caused by changes in load mass and dynamic shifts in the center of gravity, and an outer loop position control loop and an inner loop attitude control loop based on an active disturbance rejection controller are constructed, and the following control cycle is executed in real time on the basis of integrating an extended state observer: the outer loop position control loop generates a position control signal and an attitude expected angle according to an expected trajectory, the inner loop attitude control loop generates a control amount for driving the UAV according to the attitude expected angle, and the two loops work cooperatively through decoupling logic to achieve expected trajectory tracking; the extended state observer is used to estimate and compensate for unmodeled dynamics and external disturbances of the UAV in real time; a particle swarm optimization algorithm is used to dynamically adjust the parameters of the active disturbance rejection controller according to a control performance index composed of a time-weighted trajectory tracking error of the UAV and a standard deviation of an attitude angle, combined with an event-triggered mechanism with a preset minimum trigger interval and an error threshold, to adaptively compensate for changes in the kinetic characteristics caused by changes in load mass and dynamic shifts in the center of gravity. 2.The dual-loop active disturbance rejection trajectory tracking control method for a quadrotor UAV according to claim 1, wherein: The decoupling logic of the outer loop position control loop and the inner loop attitude control loop is that the outer loop position control loop is only responsible for adjusting the displacement of the UAV in the X-axis, Y-axis and Z-axis directions, and the inner loop attitude control loop is only responsible for adjusting the roll angle, pitch angle and yaw angle of the UAV, and the two loops cooperatively avoid dynamic coupling interference of position and attitude through a signal transmission path in which the outer loop outputs an attitude expected angle and the inner loop receives and executes. 3.The dual-loop active disturbance rejection trajectory tracking control method for a quadrotor UAV according to claim 1, characterized in that: The disturbance compensation process of the extended state observer is that the position, velocity and total disturbance containing unmodeled dynamics and external disturbances of the UAV are collectively estimated in real time as state variables, the total disturbance is reconstructed as an independent observation variable through a state expansion mechanism, and the observation variable is fed forward to the control end of the dual-loop active disturbance rejection controller to achieve active compensation of the disturbance.

4. The dual-loop active disturbance rejection trajectory tracking control method for a quadrotor UAV of claim 1, wherein: In the control performance index composed of the time-weighted trajectory tracking error and the standard deviation of the attitude angle, the mechanism of the time weight is that the longer the tracking error lasts, the greater its influence on the calculation result of the control performance index; and the standard deviation of the attitude angle is the sum of the standard deviation of the roll angle, the standard deviation of the pitch angle and the standard deviation of the yaw angle, which is used to quantify the stability of the flight attitude of the UAV.

5. The dual-loop active disturbance rejection trajectory tracking control method for a quadrotor UAV of claim 1, wherein: The preset minimum trigger interval of the event-triggered mechanism is a cooling time to avoid Zeno's paradox, and the error threshold is a pre-set maximum allowed value of the trajectory tracking error; when the real-time trajectory tracking error of the UAV does not exceed the error threshold, or the time interval between two parameter adjustments does not reach the preset minimum trigger interval, the particle swarm optimization algorithm does not start parameter adjustment.

6. The dual-loop active disturbance rejection trajectory tracking control method for a quadcopter unmanned aerial vehicle according to claim 2, wherein: The outer loop position control loop integrates a tracking differentiator which smoothes the input desired trajectory signal and generates a corresponding speed differential signal to suppress the impact on the UAV when the desired trajectory command is suddenly changed; the inner loop attitude control loop omits the tracking differentiator of the roll channel and the pitch channel, and directly generates the attitude control signal according to the desired attitude angle.

7. The dual-loop active disturbance rejection trajectory tracking control method for a quadcopter unmanned aerial vehicle according to claim 1, wherein: The particle swarm optimization algorithm adopts a linearly decreasing inertia weight strategy to adjust the search ability of the particles: the inertia weight linearly decreases from an initial value to a final value with the iteration number of the algorithm.

8. The dual-loop active disturbance rejection trajectory tracking control method for a quadcopter unmanned aerial vehicle according to claim 1, wherein: The driving control quantity generated by the inner loop attitude control loop corresponds to the adjustment signal of the roll angle, the pitch angle and the yaw angle of the UAV respectively; the total virtual control quantity for the execution logic of the position control and the attitude control is calculated by the control quantity in the Z-axis direction, combined with the trigonometric values of the current roll angle and the pitch angle of the UAV. 9.The dual-loop active disturbance rejection trajectory tracking control method for quadrotor UAV according to claim 1, wherein: When the particle swarm optimization algorithm adjusts the parameters of the active disturbance rejection controller, the bandwidth method is used to simplify the parameter tuning process: the observer gain and the observer bandwidth of the active disturbance rejection controller are respectively established in linear, quadratic and cubic correlation, and the controller gain and the controller bandwidth are respectively established in quadratic and linear correlation, so that the number of parameters to be tuned is simplified from 36 to 2 bandwidth parameters of the attitude control loop and 2 bandwidth parameters of the position control loop.

10. A dual-ring self-disturbance rejection trajectory tracking control system for a quadcopter unmanned aerial vehicle, characterized in that, It comprises: a model representation module configured to represent the dynamic coupling effect of the UAV caused by the change of the load mass and the dynamic offset of the center of gravity through a six-degree-of-freedom flight dynamic model containing an asymmetric inertia matrix; a double-loop active disturbance rejection control module comprising an outer loop position control submodule and an inner loop attitude control submodule, the double-loop active disturbance rejection control module receiving the six-degree-of-freedom flight dynamic model parameters output by the model representation module, the outer loop position control submodule being configured to generate a position control signal and an attitude desired angle according to a desired trajectory, and the inner loop attitude control submodule being configured to generate a control quantity for driving the UAV according to the attitude desired angle, and the two submodules working cooperatively through decoupling logic to achieve desired trajectory tracking; a disturbance compensation module integrating an extended state observer configured to receive the control feedback signal of the double-loop active disturbance rejection control module, estimate and compensate the unmodeled dynamics and external disturbances of the UAV in real time, and feed the disturbance compensation signal back to the double-loop active disturbance rejection control module; a parameter self-tuning module configured to receive the control performance feedback signal of the double-loop active disturbance rejection control module, adopt a particle swarm optimization algorithm, and dynamically adjust the parameters of the active disturbance rejection controller in the double-loop active disturbance rejection control module according to the control performance index composed of the trajectory tracking error of the UAV with time weight and the standard deviation of the attitude angle, combined with the event triggering mechanism of the preset minimum triggering interval and the error threshold, to adaptively compensate for the change of the kinetic characteristics caused by the change of the load mass and the dynamic offset of the center of gravity.

Images