Aditya Mogli

Autonomous Vehicle Safety

Started May 23, 2022

Abstract or project description

Research Question: How do {Tesla’s, Waymo’s} autonomous systems work and how safe are they compared to human driven cars?

Abstract: Autonomous vehicles are vehicles which leverage the artificial intelligence systems of the future to automate the act of driving. In recent years, many automotive companies have started to incorporate AI systems in their cars to reduce driving workload with an emphasis of increasing safety. In recent years however, automotive companies have done a poor job explaining how their systems work, especially when it comes to creating an understanding with their consumers. If AI systems are really the future, the people who are going to be using these systems and even those who are going to encounter them on the road should understand how they work. In my study, I looked at the safety of driverless cars including severity of injuries, rate of car accidents, as well as other safety metrics. I found that autonomous vehicles have helped to reduce the severity of car crashes in comparison to those driven by humans, however, driverless cars still have a long way to go before they can help reduce the rate of car accidents that occur on streets. Additionally, we sought to explore how autonomous systems are developed and in what ways their safety can be improved. Using AWS Deep Racer, we were able to create and train a CNN to classify and localize objects. This system was then tested using AWS SageMaker to evaluate and compare the safety of our model to the likes of a human driven vehicle. As AI and technology as a whole begin to improve, the safety of autonomous vehicles will improve as a result.