Application of reinforcement learning based on neural network to dynamic obstacle avoidance | IEEE Conference Publication | IEEE Xplore