Digital-analog hybrid unmanned cluster brain-like crowd-sourcing collaborative navigation method

The invention discloses a digital-analog hybrid unmanned cluster brain-like crowd-sourcing collaborative navigation method. The method comprises the following steps: step 1, establishing a motion model and a perception model of an unmanned agent cluster; 2, constructing a navigation algorithm based on a mathematical model method by using a dynamic window algorithm; 3, constructing a double-delay depth deterministic gradient network with time sequence relevance in combination with a long-short-term memory network; 4, obtaining the coordinates of the nearest obstacle of each agent of the unmanned cluster, and carrying out the density clustering of a coordinate set; step 5, calculating the barrier density rho obs according to the local information map, and calculating a dynamic fusion weight # imgabs0 # by combining the nearest distance between the intelligent agent Si and the barrier; step 6, performing hybrid fusion based on a mathematical model method and deep reinforcement learning on the decision speed of t