An Efficient Steering Angle

Abstract

From the last few years, the number of self-driving cars has increased significantly. Udacity has released a dataset that includes, among other things, a compilation of images capturing the steering angle while driving. Aim of  Udacity challenge was to predict steering angle using only the images given. We investigate two separate models that  are using deep learning techniques such as Transfer Learning, 3D CNN, LSTM, and ResNet to perform high-quality steering angle prediction based on images. Both of our models would have put in the top ten of all entries if the Udacity challenge had still been running.