Coordinated Control Method Of A Class Of Underactuated Two-wheeled Pendulum Suspension Vehicle Via A Double Closed-loop Strategy

Two-wheeled pendulum suspension (WPS) vehicle, different from the traditional two-wheeled inverted pendulum (WIP) self-balancing vehicle, is a free stable pendulum vehicle. Because of its simple structure, low energy consumption, flexible obstacle avoidance and the advantage of stable and easy control, WPS vehicle is more suitable for working under unstructured environment, therefore it has broad application prospects, such as space exploration, rescue and get more and more attention in recent years.The dynamics of WPS vehicle possesses the characteristics of nonlinearity, strong coupling, underactuation, nonholonomic constraints etc. In order to facilitate the research system of motion control and trajectory tracking, it is necessary to describe the system dynamics. First, we apply the Euler-Lagrange energy method to establish the mutual coupling of four coordinates nonholonomic dynamics model; To eliminate the influence of the nonholonomic constraints, we get the three-dimensional dynamic model by reduced order processing.Secondly, we propose double closed-loop control strategy for motion control and trajectory tracking. In the inner loop, we design the controller using the sliding mode variable structure strategy to realize speed tracking; In the external loop, we design the posture error controller to achieve arbitrary reference trajectory tracking under the inertial coordinate system.Again, aiming at the underactuated characteristic of dynamic model, an adaptive hierarchical sliding mode control strategy is put forward to solve speed tracking and at the same time to suppress the vibration of the suspension body. The adaptive rule makes the control system to get rid of the dependence on the accurate values for the physical parameters.Finally, considering the influence of friction, using continuous parametric friction model, we design a tracking controller with friction compensation to eliminate the adverse effects of friction, at the same time, use RBF network to identify the uncertain parametric friction model online. The simulation results verify the feasibility and effectiveness of the proposed control strategy.