Learning a deep neural net policy for end-to-end control of autonomous vehicles | IEEE Conference Publication | IEEE Xplore