Simulated Self-Driving using Reinforcement Learning

I collaborated with a top-notch team at Georgia Tech to create a method for a simulated car to learn how to effectively drive itself on a race track. Our team used a reinforcement learning technique called Relative Entropy Policy Search (REPS) to compute a function that would map road conditions to car control inputs. Over many iterations, this function produced car controls that resulted in faster and faster lap times. This resulted in an autonomous car that learned how to drive better over time on a simulated racetrack! We used The Open Source Race Car Simulator (TORCS) in our work.  See the code at and the .pdfs of the poster and paper below:






Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s