Deep Feature Representation Based Imitation Learning for Autonomous Helicopter Aerobatics

The drastic changes in flight parameters during aerobatics and the high instability of the system make the control of autonomous aerobatics unusually difficult. In this paper, we propose a deep feature representation based imitation learning method for autonomous aerobatics, which leverages expert demonstrations to efficiently learn the mapping of high-dimensional flight observations onto continuous actions (pitch, tail, and thrust). Different from the existing methods, our proposed method requires neither trajectory specification and alignment nor any assumptions and processing of system uncertainties, so it can greatly simplify the controller solution steps and reduce the computational burden. Particularly, our method uses the proposed deep feature representation network (DFR-network) to directly map the experts' demonstration trajectory to a deep representation space spanned by a set of learned subspaces which represent the motion patterns with the same statistical property among demonstration trajectories. Various aerobatic maneuvers can be encoded in the deep representation space through a simple combination of embedding features. Accordingly, the proposed method can perform arbitrary aerobatic maneuvers by observing a limit set of expert demonstrations. The effectiveness of the deep feature representation based imitation learning method is verified on the real-world flight data. Experiments show that compared with the existing methods, our proposed method has higher control accuracy, stronger robustness and anti-interference ability. Impact Statement-Helicopter aerobatics is a fascinating performing art but extremely challenging, which requires years of training for a human pilot to achieve reliable aerobatics such as tictocs and dodging. At present, very few methods have been reported to achieve autonomous helicopter aerobatics. Indeed, the most existing advanced method cannot perform aerobatics without expert demonstrations. Our proposed method can reliably complete the helicopter autonomous aerobatics including aerobatic maneuvers without expert demonstrations, which can be widely used in military and civil fields, including military reconnaissance and confrontation, performance, canyon rescue, and so on. It is worth indicating that our method learns embedding features that represent motion patterns, and thus is not limited to helicopter autonomous aerobatics, but can also be employed to address similar problems in other areas of robotics. © 2021 IEEE.