Search Results

While machine learning in autonomous vehicles development has increased significantly in the past few years, the use of reinforcement learning (RL) methods has only recently been applied. Convolutional Neural Networks (CNNs) became common for their powerful object detection and identification and even provided end-to-end control of an autonomous vehicle. However, one of the requirements of a CNN is a large amount of labeled data to inform and train the neural network. While data is becoming more accessible, these networks are still sensitive to the format and collection environment which makes the use of others’ data more difficult. In contrast, RL develops solutions in a simulation environment through trial and error without labeled data. Our research expands upon previous research in RL and Proximal Policy Optimization (PPO) and the application of these algorithms to 1/18th scale cars by expanding the application of this control strategy to a full-sized passenger vehicle.

Autonomous vehicle development has benefited from sanctioned competitions dating back to the original 2004 DARPA Grand Challenge. Since these competitions, fully autonomous vehicles have become much closer to significant real-world use with the majority of research focused on reliability, safety and cost reduction. Our research details the recent challenges experienced at the 2017 Self Racing Cars event where a team of international Udacity students worked together over a 6 week period, from team selection to race day. The team’s goal was to provide real-time vehicle control of steering, braking, and throttle through an end-to-end deep neural network. Multiple architectures were tested and used including convolutional neural networks (CNN) and recurrent neural networks (RNN). We began our work by modifying a Udacity driving simulator to collect data and develop training models which we implemented and trained on a laptop GPU.