Unmanned aerial vehicle path dynamic planning method for minimizing information age

The invention discloses an unmanned aerial vehicle path dynamic planning method for minimizing information age, and the method comprises the following steps: firstly describing an unmanned aerial vehicle dynamic trajectory planning problem as a partially observable Markov decision process with a non-uniform time step size, secondly, designing a deep recursion reinforcement learning algorithm to find a strategy of minimizing an age expected value of weighted average information, processing challenges from non-uniform time steps by using an improved discount mechanism, and introducing an action elimination mechanism to solve a coupling problem between effective actions and observation; the invention designs a trajectory planning algorithm based on deep recursive reinforcement learning, so that weighted average information age expectation at a base station is minimized, the unmanned aerial vehicle learns the dynamic change of the environment and makes an intelligent decision based on the dynamic change, and thro