Water-air amphibious unmanned vehicle path planning method based on reinforcement learning

The invention discloses a water-air amphibious unmanned vehicle path planning method based on reinforcement learning. The method comprises the following steps: S1, selecting an area S where an amphibious unmanned vehicle executes a path planning task, and extracting data of the corresponding area S in an electronic chart according to the area S to perform three-dimensional environment modeling; s2, constructing a Markov decision process (MDP) for path planning of the amphibious unmanned vehicle; and S3, giving a starting point and a target point, and completing global path planning according to different working scenes of the amphibious unmanned vehicle based on a depth Q network (DQN) algorithm according to the MDP of path planning of the amphibious unmanned vehicle. Compared with an existing environment modeling method for path planning of the amphibious unmanned vehicle, the planning range is widened to dozens of kilometers, the motion characteristics of the amphibious unmanned vehicle are effectively cons