NAVIGATING AERIAL VEHICLES USING DEEP REINFORCEMENT LEARNING

摘要

The technology relates to navigating aerial vehicles using deep reinforcement learning techniques to generate flight policies. A simulator may produce simulations of a flight occurring in a region of an atmosphere, a replay buffer may store frames of the simulations, and a learning module having a deep reinforcement learning architecture may process a set of frames and output a neural network encoding a learned flight policy. A meta-learning system may include stacks of learning systems, a coordinator to provide an instruction to the learning systems, and an evaluation server to evaluate resulting rewards from learned flight policies. An operational system for controlling flight of an aerial vehicle may include a computing system to process an input vector representing a state of the aerial vehicle and output an action, an operation-ready policies server to store a trained neural network encoding a learned flight policy, and a controller.