Autonomous decision making for a driver-less car

Autonomous driving has been a hot topic with companies like Google, Uber, and Tesla because of the complexity of the problem, seemingly endless applications, and capital gain. The technology's brain child is DARPA's autonomous urban challenge from over a decade ago. Few companies have had some success in applying algorithms to commercial cars. These algorithms range from classical control approaches to Deep Learning. In this paper, we will use Deep Learning techniques and the Tensor flow framework with the goal of navigating a driverless car through an urban environment. The novelty in this system is the use of Deep Learning vs. traditional methods of real-time autonomous operation as well as the application of the Tensorflow framework itself. This paper provides an implementation of AlexNet's Deep Learning model for identifying driving indicators, how to implement them in a real system, and any unforeseen drawbacks to these techniques and how these are minimized and overcome.