Multi-policy Soft Actor-Critic Reinforcement Learning for Autonomous Racing | IEEE Conference Publication | IEEE Xplore