Research On Small-scale Unmanned Aerial Vehicle Flight Control Based On Robust Compensation

As the development of electronic elements and sensors, the flight controller can be made very small, light, and still with high performance. Therefore, different kinds of small-scale unmanned aerial systems(UAS) thrive rapidly in recent years, and become increasingly popular in a wide range of applications. Comparing to middle-scale and large-scale UAS, small-scale UAS usually is small, light weight, low cost, and possesses features like easy launch and retrieve, simple control, fast deploy, expendable, etc. Large portion of rotor-crafts are even more popular due to their ability of vertical takeoff, hovering, and vertical landing. Thus this type of UAS has great potential in both military and civil applications. However, because of their various structures and the complexity in low-speed small-scale aerodynamics, the small-scale unmanned aerial vehicles(UAV)are often described as high-order nonlinear systems which are under-actuated, strong coupling, and significantly disturbed by low altitude gusts. Therefore, it is very challenging and important both in academic and practical areas for researching on this topic. This thesis is focused on the theoretical study and experimental verification of autonomous flight control problem for small-scale UAVs. Based on the robust compensation concept, the thesis studies the improvement of linear controller for single vehicle, the nonlinear robust attitude control for single vehicle, the decentralized second-order consensus tracking control problem of multiple vehicles, and the decentralized formation flight control problem of multiple vehicles. The purposes are to improve the flight control performance of small UAVs and reduce the complexity of the controller. The main contents can be summarized as follows:The most popular and mature control methods for small-scale UAVs are the classic linear approaches because of their advantages as simple structure and easy implementation. However, the classic linear approaches also have limitations in performance, robustness, and resistance of disturbances, etc. Therefore, the thesis considers the robust compensation concept and nonlinear sliding mode observer technique, proposes a new robust output-feedback attitude control method for small-scale rotor-craft UAV via sliding mode observation technique, by adding the uncertainty and disturbance compensation term to the classic linear controller. The proposed method uses the simplified secondorder model of small rotor-craft UAV, designs a classic linear quadratic regulator and a output-injection sliding mode observer separately. The proposed sliding mode observer is able to estimate all states and the equivalent effect of uncertainty and disturbance on the controller output side, by only using the position measurements. The convergence of the estimated signals to their real values in finite time is proved using Lyapunov-based method. The controller developed using the proposed method is able to improve the control performance by using the observed compensation term, and is also able to achieve output-feedback control, eliminating the requirements for velocity sensors. The effectiveness of the proposed methods is verified through numerical simulations.Although adding nonlinear robust compensator to the linear controller is able to improve the control performance in a certain degree, it still has limited performance when the system states are far from the designed linearization point, as the design of linear controller is based on the linearized plant model. In order to further improve the performance and stability of output-feedback controller, the thesis proposes a robust attitude controller for small-scale rotor-craft UAV based on the novel robust integral of the sign of the error(RISE) method. The proposed method uses simplified second-order model of small rotor-craft UAVs which only includes limited nonlinear dynamics. A auxiliary filtering system is firstly designed, then the filtered error signals and the control law are developed based on the Lyapunov stability analysis results, and the expression of uncertainty and disturbance estimation term is also obtained by using the integral of the sign of filtered error signal. By using Lyapunov based analysis method, it is proved that the proposed controller is able to achieve semi-global asymptotic tracking of any bounded and thirdorder differentiable reference signal. The convergence of the estimated uncertainty and disturbance to the real values is also proved. In order to verify the effectiveness of the proposed method, it is implemented on a three-degree-of-freedom(3-DOF) experimental platform, and the experimental results are analyzed and compared.The above mentioned two proposed methods are used for the control of single UAV and useful results are obtained. As the cost of small-scale UAV decreases, the use of multiple UAVs corporately attracts more and more attentions because of the higher flexibility and redundancy. However, the present coordinate control methods for multiple UAVs have various limitations. Therefore, by using the results of proposed single UAV control methods and the recently developments on consensus problem research, the thesis further proposes a second-order consensus tracking method for multiple UAVs via output-feedback and robust compensation. The proposed method converts the decentralized consensus problem into the control of multiple local subsystems, by designing a new consensus controller. Then by using the results for single UAV control, the output- feedback nonlinear robust controller of each subsystem is designed. The second-order consensus tracking ability of the multi-UAV system is proved by using Lyapunov based method. In order to verify the effectiveness of the proposed method, an experimental platform is built using four of 3-DOF helicopters, where the proposed method has been implemented and tested. The experimental results are further analyzed and compared.One important application of solving the consensus problem is the formation flight control of multiple UAVs, as consensus problem is the theoretical abstract of one class of multi-agent coordinate control problem. Therefore, as the last part of the thesis, and based on the previous mentioned results on second-order consensus tracking problem,a decentralized formation control method for multiple UAVs via output-feedback and robust compensation has been proposed. The proposed method converts the decentralized formation flight control problem into a consensus tracking control problem by re-defining the error system of consensus tracking controller. In order to deal with the under-actuated problem of the UAV, a cascade controller structure is also proposed and integrated into the decentralized formation controller. Similarly, the output-feedback formation tracking ability of the proposed method is proved using Lyapunov based analysis method. The effectiveness of the proposed method is verified through numerical simulations, and useful conclusions are obtained.