Skip to content

This repository contains the implementation of a DQN agent with a camera and collision sensor for navigating a map autonomously.

Notifications You must be signed in to change notification settings

AAnirudh07/DQN-CARLA

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

17 Commits
 
 
 
 
 
 

Repository files navigation

DQN-CARLA

This repository contains the implemetations for the following:

  1. Internal state-based car control.
  2. Self driving in CARLA's Town02 map.

The reinforcement learning method used for self driving was the Deep Q-Network. The Xception model was used as the DQN neural network.

The car has a camera attached to the hood. The frames captured by the camera are fed into the neural network for training in batches. A collision sensor is also attached to the vehicle. It is used in tandem with the episode duration to determine the reward.

The highest observed reward for the was around -130 at episode 100. The lowest observed reward was -340 at episode 1.

About

This repository contains the implementation of a DQN agent with a camera and collision sensor for navigating a map autonomously.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages