An offline reinforcement learning approach for path following of an unmanned surface vehicle

HIGHLIGHTS

• What: The main contributions of this paper are as follows: First, the authors propose a novel modelfree, offline learning method, a soft actor-critic with diversified Q-ensemble (SAC-N) steering controller, for continuous control in USV path following. The aim of SAC-N is to maximize both the actor`s entropy and the expected reward. In Figure 2, the authors demonstrate the application of SAC-N for learning steering control directly from offline datasets. The experiments show that the SAC-N steering controller is robust and efficient for path-following tasks in underactuated USVs.
•  

  

  If you want to have access to all the content you need to log in!

  Thanks :)

  Username or Email

  Password

  Remember me

  Lost your password?

  If you don't have an account, you can create one here.