Key points are not available for this paper at this time.
In this paper, we propose a novel DQN-based global path planning method which enables a mobile robot to efficiently obtain its optimal path in a dense environment. The method can be broken into three steps. Firstly, we need to design and train a DQN to approximate the state of the mobile robot - the action value function. Then, we determine the Q-value which corresponds to each possible action such as turning right and going forward. And the input to the DQN is the original RGB image representing the environmental structure. Finally, the current optimal action is selected by the action selection strategy, so that the mobile robot can reach the target point while avoiding obstacles. Multiple simulations in different environments show that the proposed method is effective and more efficient than traditional path planning methods.
Zhou et al. (Wed,) studied this question.