yuhuang-ca / brain-inspired-deep-imitation-learning-for-autonomous-driving-systems Goto Github PK
View Code? Open in Web Editor NEWThis project forked from intenzo21/brain-inspired-deep-imitation-learning-for-autonomous-driving-systems
Autonomous driving vehicles have drawn a great deal of interests from both academia (e.g. Oxford, MIT) and industry (e.g. Google, Tesla). However, we find that it is very difficult to directly achieve fully autonomous driving (SAE Level 5) due to generalised knowledge. To deal with the problem, deep imitation learning is a promising solution which learns knowledge from the demonstration of human. In this project, we worked on how to use deep imitation learning to achieve vehicle dynamic control (e.g. steering angle, speed). We used a dataset and simulator provided by Udacity (https://github.com/udacity/self-driving-car-sim) and the real-world comma.ai dataset.