SACHER：事後経験再生を組み込んだソフトアクター・クリティックによるUAVナビゲーションと制御

原文アブストラクト

In this paper, we propose SACHER (soft actor-critic (SAC) with hindsight experience replay (HER)), which constitutes a class of deep reinforcement learning (DRL) algorithms. SAC is known as an off-policy model-free DRL algorithm based on the maximum entropy framework, which outperforms earlier DRL algorithms in terms of exploration, robustness and learning performance. However, in SAC, maximizing the entropy-augmented objective may degrade the optimality of learning outcomes. HER is known as a sample-efficient replay method that enhances the performance of off-policy DRL algorithms by allowing the agent to learn from both failures and successes. We apply HER to SAC and propose SACHER to improve the learning performance of SAC. More precisely, SACHER achieves the desired optimal outcomes faster and more accurately than SAC, since HER improves the sample efficiency of SAC. We apply SACHER to the navigation and control problem of unmanned aerial vehicles (UAVs), where SACHER generates the optimal navigation path of the UAV under various obstacles in operation. Specifically, we show the effectiveness of SACHER in terms of the tracking error and cumulative reward in UAV operation by comparing them with those of state-of-the-art DRL algorithms, SAC and DDPG. Note that SACHER in UAV navigation and control problems can be applied to arbitrary models of UAVs.