Flexible switching control of aircraft skin inspection robot with double frames

An aircraft skin inspection robot with double frames which crawls on the aircraft skin by switching between the two frames is used to inspection the damage of the aircraft skin. In this paper, a model of the adsorption system of the aircraft skin inspection robot is established. To reduce the time of the switching process for the two frameworks, a flexible switching control strategy is proposed. Taking into account the complexity of the model and uncertainty of the operating environment, the reinforcement learning is adopted to operate the adsorption system based on the interaction with the environment. Moreover, in order to improve the training efficiency, we add a prior knowledge. Simulation results show effectiveness of the control strategy.