| Underactuated mechanical systems refer to the mechanical systems with fewercontrol inputs than degree of freedom. Compared with the fullyactuated mechanicalsystems, underactuated mechanical systems have huge advantages in saving the space,reducing weight and energy consumption, because of the actuator missing. Acrobot,obtained from the cantilever motion of biomimetic robots, is a typical underactuatedsystem, which is one of the most popular benchmark for research on underactuatedsystem.In this paper, as the research target, we focus on the stabilization control ofAcrobot. We try to find a control algorithm, which can not only stabilize Acrobotfrom the downward position to the upward position in theory, but also easy toimplement. The main contents are as follows:Lagrange equations in analytical mechanics are used to build the precisemathematical model of Acrobot, approximately linearization at the equilibrium point,We build the simulation model in Matlab/Simulink, verifying the mathematicalmodel with “Necessary Condition Lawâ€.To achieve the stabilization control of Acrobot, we build the controllers withtwo different control strategies: control strategy based on single controller, controlstrategy based on switching of the swing-up controller and balance controller.Backstepping approach is adopted for the design of stabilizing controller in controlstrategy based on single controller. In control strategy based on switching thecontrollers, the swing-up controller based on Lyapunov theory and the balancecontroller based on LQR are designed. We verify the different strategies inMatlab/Simulink.As the complex mechanical structure and the parameters distribution of Acrobot,it’s difficult to get the actual parameters by theoretical calculation. In the paper, weuse UKF to obtain the identification parameters of the system.Ultimately, we do the semi-physical experiments in the experimental platformbased on dSPACE, achieving the control objectives based on switching controllers,verifying the control algorithm. |
