All Software
Python

Advanced Planning RL

Implementation of advanced planning for autonomous vehicles using Reinforcement Learning and Deep Inverse RL.

Advanced Planning RL
View on GitHub Report Issue

Advanced Planning using RL

This project implements advanced planning algorithms for autonomous vehicles, combining Reinforcement Learning (RL) and Deep Inverse Reinforcement Learning (IRL). It focuses on learning driving behaviors and planning paths that mimic expert demonstrations.

Key Features

  • Deep Inverse RL: Recovers reward functions from expert driving data.
  • Reinforcement Learning: Trains agents to optimize the learned or specified reward functions.
  • Autonomous Driving: Tailored for highway and urban driving scenarios.
  • Behavior Learning: Capable of learning complex maneuvers like lane changing and overtaking.

Applications

  • Autonomous vehicle behavior planning
  • Learning from demonstration (LfD)
  • Simulation-based training for self-driving cars

Technical Details

  • Language: Python
  • Libraries: TensorFlow / PyTorch
  • Algorithms: Maximum Entropy IRL, Deep Q-Learning

Visit the GitHub repository for the implementation details.