DEEP LEARNING BASED MOTION CONTROL OF A GROUP OF AUTONOMOUS VEHICLES

The present invention is related to a controller (1) for an agent (Ai) of a group of agents (An), in particular for a group of autonomous or semi-autonomous vehicles, and to a computer program implementing such a controller (1). The invention is further related to a temporal deep network (2) for such a controller (1) and to a method, a computer program and an apparatus for training the temporal deep network (2). The controller (1) comprises a temporal deep network (2) designed to calculate a desired trajectory for the agent (Ai), a nonlinear model predictive controller (3) designed to calculate commands for the agent (Ai) based on the desired trajectory and desired trajectories of the other agents (An) of the group of agents (An), and an augmented memory (4) designed to integrate historic system states of the group of agents (An) for the temporal deep network (2).