Navigation of Autonomous Vehicles using Reinforcement Learning with Generalized Advantage Estimation
Edwar Jacinto, Fernando J. Martínez, Fredy Martinez
Abstract
This study proposes a reinforcement learning ap-proach using Generalized Advantage Estimation (GAE) for autonomous vehicle navigation in complex environments. The method is based on the actor-critic framework, where the actor network predicts actions and the critic network estimates state values. GAE is used to compute the advantage of each action, which is then used to update the actor and critic networks. The approach was evaluated in a simulation of an autonomous vehicle navigating through challenging environments and it was found to effectively learn and improve navigation performance over time. The results suggest GAE as a promising direction for further research in autonomous vehicle navigation in complex environments.