CN113253617A - Online self-adaptive control method for quad-rotor unmanned aerial vehicle - Google Patents

Abstract

The invention provides an online self-adaptive control method for a quadrotor unmanned aerial vehicle. The coupling problem on each loop is processed to obtain the outer loop position subsystem and the inner loop attitude subsystem; for the outer loop position subsystem, according to the first-order tracking error, a robust controller based on the sliding mode control algorithm is designed. The controller introduces the exponential approximation law as a robust term to realize the suppression of external disturbances by the outer loop control system; for the inner loop attitude subsystem, in order to solve the modeling deviation caused by the uncertain part of the mechanism model, a fuzzy compensation strategy is designed. And establish an adaptive fuzzy sliding mode controller based on fuzzy compensation for real-time identification and online adjustment of model parameters; according to the outer ring position subsystem and the inner ring attitude subsystem, the model of the quadrotor UAV is analyzed and proved stability.

Description

Online adaptive control method for quadrotor UAV

technical field

The invention relates to the field of quad-rotor UAV control, in particular to an online adaptive control method for quad-rotor UAV.

Background technique

In recent years, quadrotor UAVs have been widely used in the fields of military, agriculture, search and rescue, fire protection, environmental protection and personal entertainment. The popularity of drones has benefited from their safety and mobility. Compared with helicopters and fixed-wing UAVs, quadrotors have significant advantages such as strong vertical take-off and landing capability, strong maneuverability, simple mechanical structure, and low cost. In addition, the flexible design capabilities of its size and specification allow the quadrotor to be quickly adapted to the requirements of a given industry. Despite these advantages, in the actual flight environment, the ideal trajectory tracking performance of the quadrotor is always affected by uncertain internal and external nonlinear factors. Since the quadrotor system is a typical dynamic nonlinear system, it has the characteristics of strong coupling and underactuation, and the model has uncertainty due to the influence of airframe vibration and air resistance, so it is difficult to obtain an accurate system model.

SUMMARY OF THE INVENTION

The invention provides an online self-adaptive control method for a quadrotor unmanned aerial vehicle, the purpose of which is to solve the problem that the quadrotor unmanned aerial vehicle cannot realize real-time adjustment of control parameters and body state.

In order to achieve the above object, an embodiment of the present invention provides an online adaptive control method for a quadrotor unmanned aerial vehicle, including:

Based on the mechanism model of the quadrotor UAV, analyze the nonlinear coupling characteristics of the model, introduce intermediate variables to deal with the coupling problem of the model in the two loops of position and attitude, and obtain the outer loop position subsystem and the inner loop attitude subsystem;

For the outer-loop position subsystem, according to the first-order tracking error, a robust controller based on the sliding mode control algorithm is designed, and the controller introduces an exponential approximation law as a robust term to achieve the outer-loop control system's suppression of external disturbances;

For the inner loop attitude subsystem, a fuzzy compensation strategy for model uncertainty is designed, and an adaptive fuzzy sliding mode controller based on fuzzy compensation is established for real-time identification and online adjustment of model parameters;

According to the analysis of the outer ring position subsystem and the inner ring attitude subsystem, the stability of the model of the quadrotor UAV is proved.

Wherein, the step 1 specifically includes:

According to Newtonian mechanics and the Newton-Lagrange equation, the dynamic equation of the quadrotor relative to the inertial coordinate system can generally be expressed as:

（1）

(1)

Decouple the outer loop position subsystem by introducing two dummy variables 𝑈𝑥 and 𝑈𝑦;

（2）

(2)

in,

The outer-loop subsystem model with internal and external disturbances applied is:

（3）

(3)

表示系统在x，y和z三个方向上的内部不确定性，

代表实际环境中的阵风、突变因素导致的外部扰动；in,

represents the internal uncertainty of the system in the three directions of x, y and z,

Represents external disturbances caused by gusts and sudden changes in the actual environment;

Decoupling the inner loop attitude subsystem;

According to the model expression in Eq. (1), the equation of the attitude subsystem of the inner ring can be expressed as

（4）

(4)

。in,

.

Wherein, the step 2 specifically includes:

By introducing three dummy variables Ux, Uy and Uz, the translational motion equation can be simplified to:

（5）

(5)

Defining two state variables x ₁ , x ₂ , equation (5) can be rewritten in the following state space form:

（6）

(6)

Among them, 𝑈𝑥 = 𝑈1(𝑐𝑜𝑠𝜙𝑠𝑖𝑛𝜃𝑐𝑜𝑠𝜓 + 𝑠𝑖𝑛𝜙𝑠𝑖𝑛𝜓) , 𝑥 ₁ , 𝑥 ₂ are used to describe the speed and acceleration of the system in the x direction, respectively;

Define a dummy variable 𝛽 and introduce a first-order tracking error,

（7）

(7)

Define two error variables 𝑒 ₁ 𝑥, 𝑒 ₂ 𝑥, and set

（8）

(8)

，根据方程（8）得：Introduce sliding surface

, according to equation (8):

（9）

(9)

Use the exponential approximation law to improve the robustness of the position control system:

（10）

(10)

The final control amount is as follows:

，求导得：Define a Lyapunov function

, find the derivation:

（11）

(11)

使，滑动面s→0，位置控制子系统是渐近稳定的；when satisfied

so that the sliding surface s→0, the position control subsystem is asymptotically stable;

Further control quantities in the y and z directions can be calculated:

（12）。

(12).

Wherein, the step 3 includes:

Use the fuzzy compensation model to solve the modeling deviation caused by the uncertain part of the mechanism model, and establish an adaptive fuzzy sliding mode controller based on fuzzy compensation for real-time identification and online adjustment of model parameters. The adaptive fuzzy sliding mode controller The improved super-twisting algorithm is used to suppress the chattering of the control system on the sliding mode surface. By integrating the fuzzy model and the robust sliding mode controller into a unified framework, the real-time identification of the parameters of the closed-loop control system is realized.

The above-mentioned scheme of the present invention has the following beneficial effects:

The online self-adaptive control method for the quadrotor UAV of the present invention relies on the actual mechanism model, takes into account the internal uncertainty and external disturbance of the model, and has good robustness. According to the nonlinear coupling characteristics of the system, the system is decoupled into inner and outer loop subsystems, and the characteristics and disturbance characteristics of each subsystem are fully considered to design the controller independently, which effectively ensures the stability of the control process. In the design of the inner loop controller, according to the idea of feedback compensation control, the fuzzy modeling method is used for the adaptive compensation of the internal uncertainty of the model to effectively restore the real system dynamics; The model controller adopts an improved super-twist algorithm, which can not only effectively suppress the chattering of the control system on the sliding mode surface, but also improve the robustness of the entire closed-loop control system.

Description of drawings

1 is a schematic flowchart of an online adaptive control method for a quadrotor unmanned aerial vehicle of the present invention;

2 is an overall architecture diagram of an online adaptive control method for a quadrotor unmanned aerial vehicle of the present invention;

Fig. 3 is the framework diagram of the robust sliding mode control method of the outer loop position subsystem of the present invention;

Fig. 4 is the framework diagram of the self-adaptive sliding mode control method of the inner loop attitude subsystem based on fuzzy compensation of the present invention;

Fig. 5 is the designed online fuzzy approximation model structure diagram of the present invention;

Fig. 6 is the control idea schematic diagram of the adaptive robust sliding mode controller of the present invention;

7 is a schematic diagram of the roll angle attitude tracking result and error of the present invention;

Fig. 8 is the pitch angle attitude tracking result and error schematic diagram of the present invention;

9 is a schematic diagram of the heading angle and attitude tracking result and error of the present invention;

10 is a schematic diagram of white Gaussian noise introduced in the simulation of the present invention;

FIG. 11 is a schematic diagram of a trajectory tracking result under random noise according to the present invention.

Detailed ways

In order to make the technical problems, technical solutions and advantages to be solved by the present invention more clear, the following will be described in detail with reference to the accompanying drawings and specific embodiments.

As shown in FIG. 1 and FIG. 2 , an embodiment of the present invention provides an online adaptive control method for a quadrotor UAV, including: for the model features of the outer loop position subsystem, develop a sliding mode control method based on The robust controller of the algorithm; by introducing the integration of the first-order tracking error into the controller, and using the exponential approximation law as the robust term, the outer loop control system can suppress the external disturbance and effectively improve the robust tracking of the system. performance. Secondly, for the attitude subsystem of the inner loop, in order to solve the modeling deviation caused by the uncertain part of the mechanism model, a fuzzy compensation strategy for the uncertainty of the model is designed, and an adaptive fuzzy sliding mode controller based on fuzzy compensation is established. For the real-time identification and online adjustment of model parameters, the inner loop controller adopts an improved superdistortion algorithm, which effectively suppresses the chattering of the control system on the sliding surface. These sub-control systems are integrated into a unified closed-loop system, and the consistent stability of the whole system is theoretically analyzed and rigorously proved, and its superior control performance compared with several commonly used methods is demonstrated.

When controlling the position subsystem, the chattering caused by the control switching of the attitude subsystem is regarded as the external disturbance, and the integral of the first-order tracking error is introduced to improve the robustness to the external disturbance. During attitude control, the chattering caused by the control switching of the position subsystem is regarded as external disturbance, and fuzzy and super-distortion algorithms are introduced to ensure the robustness of the system to disturbances. In this way, jitter caused by control switching is suppressed.

1) Construction of the quadrotor UAV model

According to Newtonian mechanics and the Newton-Lagrange equation, the dynamic equation of the quadrotor relative to the inertial coordinate system can generally be expressed as:

（1）

(1)

In actual flight, the lift of the quadrotor is mainly generated by the rotation of the propeller. First, according to the gravity of the aircraft, the rotation speed of each propeller is obtained by inverse calculation, so that the aircraft can offset its own gravity through the four propellers, so as to obtain the basic lift. Then, on the basis of the basic thrust, the controller mainly performs incremental and decrement adjustments, that is, adjusts the speed of the four motors according to the deviation between the target point and the current position. As shown in Equation (1), the rotation information, , is related to U2, U3, and U4, respectively; similarly, the positional accelerations ẍ, ÿ, z̈ are related to U1, m, , . Practical problems may lead to difficulties in modeling and control: (1) different forms of wind gusts in the actual flight environment, i.e. sudden changes, high frequencies, cyclones, etc., affect the robustness of the controller; (2) caused by external factors such as air resistance Body vibrations can lead to model uncertainty. An effective method is to decouple the position and attitude parameters and develop an adaptive control algorithm based on the inner and outer loop subsystems. When tracking the target trajectory, the outer loop position controller will call the corresponding algorithm to control the angle of the quadrotor according to the input target trajectory and actual position information. The output regular velocity deviation is passed to the inner loop attitude controller. Then the inner loop adjusts the attitude of the aircraft according to this deviation, and transmits the thrust information to the final execution unit.

A. Outer loop position subsystem decoupling

By introducing two dummy variables 𝑈𝑥 and 𝑈𝑦, a decoupling method is proposed to solve the coupling relationship between the inner and outer loop control systems:

（2）

(2)

in,

The outer-loop subsystem model with internal and external disturbances applied is:

（3）

(3)

表示系统在x，y和z三个方向上的内部不确定性，

代表实际环境中的阵风、突变等因素导致的外部扰动。in,

represents the internal uncertainty of the system in the three directions of x, y and z,

It represents the external disturbance caused by factors such as gusts and sudden changes in the actual environment.

B. Inner loop attitude subsystem decoupling

According to the model expression in Eq. (1), the equation of the attitude subsystem of the inner ring can be expressed as

（4）

(4)

。in,

.

2) Construction of the outer loop position controller

As mentioned in (3), by introducing three dummy variables Ux, Uy and Uz, the translational motion equation can be simplified as:

（5）

(5)

Although model uncertainties and disturbances are unpredictable in actual flight and cannot be acquired as prior knowledge, the effects of these factors can be mitigated by introducing control terms. Here, a robust sliding mode controller is constructed to achieve accurate and stable motion in three dimensions, as shown in Figure 3.

Take the x-direction as an example to illustrate the design process of the controller. The controller ensures that the system is asymptotically bounded by designing dummy variables, so that the control output can not only suppress interference, but also achieve asymptotic tracking of the reference signal. Defining two state variables x ₁ , x ₂ , equation (5) can be rewritten in the following state space form:

（6）

(6)

Among them, 𝑈𝑥 = 𝑈1(𝑐𝑜𝑠𝜙𝑠𝑖𝑛𝜃𝑐𝑜𝑠𝜓 + 𝑠𝑖𝑛𝜙𝑠𝑖𝑛𝜓𝑥) ; 𝑥1 , 𝑥2 are used to describe the speed and acceleration of the system in the x direction, respectively. Inspired by the backstepping algorithm, a dummy variable 𝛽 is defined here, and a first-order tracking error is introduced to ensure the robustness of the tracking.

（7）

(7)

Define two error variables 𝑒 _1𝑥 ,𝑒 _2𝑥 , and set

（8）

(8)

，根据方程（8）得：Introduce sliding surface

, according to equation (8):

（9）

(9)

Use the exponential approximation law to improve the robustness of the position control system:

（10）

(10)

The final control amount is as follows:

；

;

，求导得：Define a Lyapunov function

, find the derivation:

（11）

(11)

使，滑动面s→0，位置控制子系统是渐近稳定的。when satisfied

Let, the sliding surface s→0, the position control subsystem is asymptotically stable.

Further control quantities in the y and z directions can be calculated:

（12）

(12)

3) Construction of the inner loop attitude controller

For the attitude sub-control system, an online adaptive fuzzy sliding mode controller is proposed. As shown in Fig. 4, this method uses fuzzy model and SMC to overcome uncertainty and external disturbance, respectively. Furthermore, to alleviate the chattering problem existing in traditional sliding mode control, we introduce an improved STA that applies a chattering switching function to the higher-order derivatives of sliding mode variables by introducing integral and exponential terms. On this basis, an online adaptive strategy is designed, which integrates the fuzzy model and the robust sliding mode controller into a unified framework, and realizes the real-time identification of the control system parameters.

A. Fuzzy Approximation System

Since it is difficult to obtain an exact f(t, q, q) equation, according to the universal approximation theorem, there must be a fuzzy system f̂(t, q ,q̇) highly approximate to f(t, q, q̇). Therefore, a two-dimensional fuzzy controller can be constructed to approximate f(t, q, q̇). In order to ensure the approximation accuracy, it is defined here as the approximation error ε.

It can be seen from formula (4) that there are three angle variables in the attitude control system, which should be considered comprehensively when designing the fuzzy system. Here, we use the pitch angle ϕ as an example. The detailed fuzzy system is shown in Figure 5.

Assuming that A1 and A2 represent the fuzzy result sets of φ1 and φ2, respectively, the fuzzy rules are established as follows:

对应于隶属函数的最大值。此外，借助于乘积推理机，推理结论可以表述为

。因此，基于中心平均解模糊器获得系统输出为：In the fuzzy system, a single-value fuzzer is used to calculate the rule result, the function value

corresponds to the maximum value of the membership function. In addition, with the help of the product inference engine, the inference conclusion can be expressed as

. Therefore, the system output obtained based on the center-averaged defuzzifier is:

（13）

(13)

Define the parameters 𝜔 _𝜙 and introduce a new weight 𝜁(𝜙), then the fuzzy output can be further transformed into:

（14）

(14)

Then, the weight 𝜁(𝜙) of the member l ₁ l ₂ (l ₁ = 1,2, ⋯ , 𝑚; l ₂ =1,2, ⋯ , 𝑛) can be expressed as:

（15）

(15)

，可得：Assumption

,Available:

（16）

(16)

where ε is the approximation error.

有：definition

Have:

（17）

(17)

有：for

Have:

（18）

(18)

（19）The dynamic attitude equation can be rewritten as

(19)

B. Adaptive Synovial Controller Based on Fuzzy Compensation

Using the former fuzzy approximation system, the model uncertainty of the system can be compensated. In order to obtain satisfactory tracking accuracy and robust performance against external disturbances, an adaptive robust controller is further proposed. In this controller, an improved STA is introduced as a robust term to alleviate the chattering problem of traditional sliding mode controllers. The robust approximation law changes with the real-time feedback state and the fuzzy approximation term. The structure of the robust sliding mode algorithm is shown in Figure 6.

且有

。因此，s_q可进一步表示为：Introduce sliding surface

and have

. Therefore, s _q can be further expressed as:

（20）

(20)

It is well known that when the sliding modal function s _q → 0, the height error e and its derivative exponentially converge to zero. Therefore

（21）

(twenty one)

Here, an improved STA algorithm is proposed as the robust term of SMC, which inherits the characteristics of traditional linear and nonlinear STA in interference suppression to overcome the chattering of the second-order sliding mode and improve the robustness of the system to external interference:

（22）

(twenty two)

The final control amount is as follows:

（23）

(twenty three)

，求导得：Define a Lyapunov function

, find the derivation:

（24）

(twenty four)

的第一部分满足

；此外，当满足

成立。因此，这里只需计算𝑠_𝑞收敛到0的边界条件即可。最终推导得到的边界条件如下：Obviously,

The first part satisfies

; in addition, when satisfying

established. Therefore, it is only necessary to calculate the boundary conditions where 𝑠 _𝑞 converges to 0. The final derived boundary conditions are as follows:

（25）

(25)

Finally, by integrating these sub-control systems into a unified closed-loop system, the precise and stable control of the quadrotor UAV under uncertain internal parameters and external disturbances can be realized.

The online self-adaptive control method for the quadrotor UAV of the present invention relies on the actual mechanism model, takes into account the internal uncertainty and external disturbance of the model, and has good robustness. According to the nonlinear coupling characteristics of the system, the system is decoupled into inner and outer loop subsystems, and the characteristics and disturbance characteristics of each subsystem are fully considered to design the controller independently, which effectively ensures the stability of the control process. In the design of the inner loop controller, according to the idea of feedback compensation control, the fuzzy modeling method is used for the adaptive compensation of the internal uncertainty of the model to effectively restore the real system dynamics; The model controller adopts an improved super-twist algorithm, which can not only effectively suppress the chattering of the control system on the sliding mode surface, but also improve the robustness of the entire closed-loop control system.

The simulation and comparison experiments of the present invention are carried out based on three commonly used control algorithms. This is done to demonstrate its specific implementation in detail and to verify the effectiveness of the proposed method. The parameters of the quadrotor used in the simulation are shown in Table 1.

Table 1 Four-rotor related

1. Attitude tracking simulation

To illustrate the robustness and convergence performance of the proposed fuzzy adaptive sliding mode controller, we take the attitude subsystem as an example to conduct a comparative simulation of its attitude tracking effect and input stability.

(1) Cascade PID controller (PID)

(2) Improved sliding mode controller (M-SOSM)

(3) Fuzzy sliding mode controller

(4) Fuzzy adaptive sliding mode controller (AdapFuzzy M-SOSM)

Among these three kinds of controllers, the cascade PID controller contains the double closed-loop PID control strategy of angular velocity and angular velocity. The sliding mode control of the M-SOSM controller and the fuzzy M-SOSM controller is optimized by an improved superwarping algorithm (STA), which is the same as the robust term of the fuzzy adaptive M-SOSM controller. This is also to demonstrate anti-jamming robustness and online approximation performance. The model uncertainty selection for the attitude subsystem is:

（26）

(26)

As shown in equation (27), the multi-frequency function u(t) and the sinusoidal signal τ(t) are used as the frequency conversion reference signal and the high-frequency dynamic external disturbance, respectively.

（27）

(27)

The parameters used in the comparative simulation are shown in Table 2:

Table 2 Control parameters used in the attitude simulation process

The fuzzy membership function is explained as follows:

（28）

(28)

时，三个姿态角的跟踪效果和跟踪误差如图7-图9所示。从图中可以看出，采用改进的超扭曲算法可以有效地缓解高频干扰。此外，与M-SOSM和模糊M-SOSM相比，该控制器的跟踪误差减小很快，在0.5s内收敛到零。这意味着自适应律和模糊系统的引入大大提高了跟踪精度。通过与PID串级控制器和二阶滑模控制器的比较，说明了所设计的自适应控制策略只需在时间上（近0.6s）进行极小的调整，就可以达到令人满意的跟踪效果。另外，其他算法的跟踪误差分别在和[-0.7,0.4]、[-0.3,0.3]和[0,0.3]之间。对比仿真的均方根误差（RMSE）在表3中进行了定量总结。给出了这些结果，并与串级PID、M-SOSM和模糊M-SOSM进行了比较，所设计的控制器在姿态跟踪稳定性方面具有更快的收敛速度和稳定性。When selecting the membership weight Wq, it is found that the tracking effect decreases with the increase of the initial value before the tracking process is stable. But it has no effect on tracking convergence time. This means that when choosing the initial value of Wq, try to choose a small initial value. Here, when the membership weight value is

, the tracking effects and tracking errors of the three attitude angles are shown in Figures 7-9. It can be seen from the figure that the use of the improved superdistortion algorithm can effectively alleviate the high frequency interference. Furthermore, compared with M-SOSM and fuzzy M-SOSM, the tracking error of this controller decreases rapidly, converging to zero within 0.5s. This means that the introduction of adaptive laws and fuzzy systems greatly improves the tracking accuracy. By comparing with the PID cascade controller and the second-order sliding mode controller, it shows that the designed adaptive control strategy can achieve satisfactory tracking with only a very small adjustment in time (near 0.6s). Effect. In addition, the tracking errors of other algorithms are between [-0.7, 0.4], [-0.3, 0.3] and [0, 0.3], respectively. The root mean square error (RMSE) of the comparative simulations is quantitatively summarized in Table 3. These results are given and compared with Cascade PID, M-SOSM and Fuzzy M-SOSM, the designed controller has faster convergence speed and stability in attitude tracking stability.

Table 3 Quantitative comparison of tracking performance

2. Trajectory tracking simulation

Through trajectory tracking simulation, the effectiveness and mobility of the control strategy are verified. The object trajectory used is a cylindrical spiral, as shown in Eq. (29),

（29）

(29)

The simulation time is 20s, and the initial position coordinates and attitude of the quadrotor are set as [x, y, z]=[0 0 0]; in addition, [ϕ, θ, ψ]=[0 0 0] certainty. Suppose the uncertain part of the model is as follows:

（30）

(30)

Two forms of gusts are proposed in the simulations: high-frequency random disturbances (high-frequency sinusoidal signals) and multi-frequency random disturbances (square-wave signals). The interference used in the tracking simulation consists of the following two parts:

（31）

(31)

Related research shows that the parameter cx does not affect the stability of the control system, but has a direct relationship with the convergence time. If the parameter cx is too large, the convergence speed is too fast and strong chattering will occur; if the parameter a is too small, the convergence time will be long. Therefore, in this simulation, we used the empirical value of cx=5. The fuzzy basis vectors for the altitude and attitude controllers are w=[0.1*one(75,1)] and w=[zeros(25,1)]. In addition, since practical factors such as measurement noise and navigation error will affect the control performance to a certain extent, additional tracking simulations are performed in a random noise environment to verify the effectiveness of the proposed control strategy. In this simulation, relevant parameters such as simulation time, initial position coordinates, target trajectory, model uncertainty, and external disturbance are analyzed, and the results are shown in Figure 10-11. The random noise in Figure 10 reflects simulated measurement noise and navigation errors in a real environment. Referring to the treatment of measurement error in some literatures, when factors such as measurement technology, people and instruments are ignored, the value and sign of measurement error change randomly in an unpredictable manner and obey a normal distribution. Therefore, in the simulation, we choose white Gaussian noise as the measurement noise, which follows a normal distribution with 0 mean and 0.05 variance. Noise generated by random number generator, sampling period is 0.001s. Figure 11 shows the results of trajectory tracking. The results show that despite the presence of Gaussian random noise, the control system still has stable tracking performance and can quickly converge to track the cylindrical helical trajectory. According to the tracking error, the positional response of the quadrotor UAV to random noise and external disturbance shows a fluctuating trend, and the amplitude is very small (bounded at ±0.1m), which proves the suppression effect of the proposed controller on random noise. Table 4 shows the confidence metrics in the x, y, and z directions, with a 0.95 confidence level for random noise.

Table 4 Confidence level (95%) under random measurement noise

To sum up, in view of the accurate trajectory tracking problem of quadrotor UAV under the uncertainty of time-varying model and external interference, on the basis of modeling the kinematics and dynamic mechanism of quadrotor UAV, the quadrotor unmanned aerial vehicle has no problem. The inner and outer ring models of the human-machine are decoupled, which effectively solves the coupling problem between the position parameters and attitude parameters of the quadrotor UAV. Firstly, for the model characteristics of the outer loop position subsystem, a robust controller based on the sliding mode control algorithm is developed; by introducing the integration of the first-order tracking error into the controller, the robust tracking of the system to external disturbances is effectively improved performance. Secondly, for the attitude subsystem of the inner loop, an adaptive fuzzy sliding mode controller based on the improved STA algorithm is designed by combining the fuzzy algorithm with the sliding mode control. real-time identification and adjustment. Finally, these sub-control systems are integrated into a unified closed-loop system. When controlling the position subsystem, the chattering caused by the control switching of the attitude subsystem is regarded as the external disturbance, and the integral of the first-order tracking error is introduced to improve the robustness to the external disturbance. During attitude control, the chattering caused by the control switching of the position subsystem is regarded as an external disturbance, and fuzzy and super-distortion algorithms are introduced to ensure the robustness of the system to the disturbance, so as to effectively suppress the chattering caused by the control switching.

The above are the preferred embodiments of the present invention. It should be pointed out that for those skilled in the art, without departing from the principles of the present invention, several improvements and modifications can be made. It should be regarded as the protection scope of the present invention.

Claims (4)

1. An online adaptive control method for a quad-rotor drone, comprising:

analyzing the nonlinear coupling characteristic of the model based on a mechanism model of the quad-rotor unmanned aerial vehicle, and introducing an intermediate variable to process the coupling problem of the model on two loops of position and attitude to obtain an outer loop position subsystem and an inner loop attitude subsystem;

for the outer ring position subsystem, designing a robust controller based on a sliding mode control algorithm according to a first-order tracking error, and introducing an exponential approximation law into the controller to serve as a robust item to realize the suppression of the outer ring control system on external disturbance;

designing a fuzzy compensation strategy aiming at model uncertainty for the inner ring attitude subsystem, and establishing a self-adaptive fuzzy sliding mode controller based on fuzzy compensation for real-time identification and online adjustment of model parameters;

and the stability of the model of the quad-rotor unmanned aerial vehicle is proved according to the analysis of the outer ring position subsystem and the inner ring attitude subsystem.

2. An online adaptive control method for a quad-rotor unmanned aerial vehicle according to claim 1, wherein the step 1 specifically comprises:

the dynamic equation of a quadrotor relative to an inertial coordinate system can be generally expressed as follows according to newton mechanics and newton-lagrange equation:

（1）

decoupling the outer ring position subsystem by introducing two virtual variable sums;

（2）

wherein,

the model of the outer ring subsystem with the applied internal and external disturbances is:

（3）

wherein,

representing the internal uncertainty of the system in the three x, y and z directions,

representing gusts in the actual environment, external disturbances caused by sudden change factors;

decoupling the inner ring attitude subsystem;

according to the model expression in equation (1), the equation for the inner ring attitude subsystem can be expressed as

（4）

Wherein,

。

3. an online adaptive control method for a quad-rotor drone according to claim 1, characterized in that said step 2 comprises in particular:

by introducing three virtual variables Ux, Uy and Uz, the translational motion equation can be simplified as:

（5）

two state variables x are defined₁，x₂Equation (5) can be rewritten as the following state space form:

（6）

wherein,𝑈𝑥 = 𝑈1(𝑐𝑜𝑠𝜙𝑠𝑖𝑛𝜃𝑐𝑜𝑠𝜓 + 𝑠𝑖𝑛𝜙𝑠𝑖𝑛𝜓) ，𝑥₁ , 𝑥₂respectively representing the speed and the acceleration in the x direction and used for describing the motion state of the system in the x direction𝑥Direction;

defining a virtual variable, and introducing a first-order tracking error,

（7）

defining two error variables𝑒_1𝑥, 𝑒_2𝑥And is provided with

（8）

Introducing slip form surfaces

From equation (8), we obtain:

（9）

the robustness of the position control system is improved by using an exponential approximation law:

（10）

the control quantities are finally obtained as follows:

defining Lyapunov functions

And obtaining a derivative:

（11）

when it is satisfied with

So that, sliding surface s → 0, the position control subsystem is asymptotically stable;

the control quantities in the y and z directions can further be calculated:

（12）。

4. an online adaptive control method for a quad-rotor drone according to claim 1, wherein said step 3 comprises:

the method comprises the steps of solving modeling deviation caused by an uncertain part in a mechanism model by using a fuzzy compensation model, establishing a self-adaptive fuzzy sliding mode controller based on fuzzy compensation for real-time identification and online adjustment of model parameters, inhibiting buffeting of a control system on a sliding mode surface by adopting an improved super-distortion algorithm, and realizing real-time identification of parameters of a closed-loop control system by integrating the fuzzy model and a robust sliding mode controller into a unified frame.

Images