Deep Reinforcement Learning and Randomized Blending for Control under Novel Disturbances. Issue 2 (2020)

• Record Type: Journal Article
• Authors: Sohège, Yves; Provan, Gregory; Quiñones-Grueiro, Marcos; Biswas, Gautam
• Is Part Of: IFAC-PapersOnLine. Volume 53:Issue 2(2020)
• Page Start: 8175
• Page End: 8180
• Publication Date: 2020
• Subjects: Design of fault tolerant/reliable systems -- Fault accommodation -- Reconfiguration strategies -- Methods based on neural networks and/or fuzzy logic for FDI; Automatic control -- Periodicals; 629.805
• DOI: 10.1016/j.ifacol.2020.12.2313
• Languages: English
• ISSNs: 2405-8963

Abstract

Abstract: Enabling autonomous vehicles to maneuver in novel scenarios is a key unsolved problem. A well-known approach, Weighted Multiple Model Adaptive Control (WMMAC), uses a set of pre-tuned controllers and combines their control actions using a weight vector. Although WMMAC offers an improvement to traditional switched control in terms of smooth control oscillations, it depends on accurate fault isolation and cannot deal with unknown disturbances. A recent approach avoids state estimation by randomly assigning the controller weighting vector; however, this approach uses a uniform distribution for control-weight sampling, which is sub-optimal compared to state-estimation methods. In this article, we propose a framework that uses deep reinforcement learning (DRL) to learn weighted control distributions that optimize the performance of the randomized approach for both known and unknown disturbances. We show that RL-based randomized blending dominates pure randomized blending, a switched FDI-based architecture and pre-tuned controllers on a quadcopter trajectory optimisation task in which we penalise deviations in both position and attitude.