Date of Award:

5-2016

Document Type:

Thesis

Degree Name:

Master of Science (MS)

Department:

Electrical and Computer Engineering

Committee Chair(s)

Ryan Gerdes

Committee

Ryan Gerdes

Committee

Rajnikant Sharma

Committee

Tam Chantem

Abstract

In the past few years, automated cars have ceased to be part of science fiction, and have instead become a technology that has been implemented, with partially automated systems currently available to customers.

One benefit of automated vehicle technology is the consistent driving patterns due to automation, instead of the inconsistency of distractible humans. Passengers of automated vehicles will be exposed to much less danger than the passengers of human-driven vehicles.

These statements will only be true as automated vehicle systems are scrutinized by experts to find flaws in the system. Security enthusiasts have already hijacked control of an automated car remotely with a cell phone [1]. As security flaws are exposed and removed, the vehicle automation community will become safer.

This research investigates the design of an automated vehicle and potential methods to protect the security of such a system. Vehicle control and guidance algorithms are analyzed and presented.

In the event of sensor failure, which could be naturally caused or performed by a malicious entity, automated vehicles cease to operate correctly, either crashing or returning control to people. Minimizing the risks of sensor failure can be achieved in multiple ways. One potential solution is to use available sensors to detect additional information about parameters of interest, such as using two radio antennas to triangulate the origin of a communicated signal. A method to extend the capability of a standard monocular camera by determine inter-vehicle distance from an image. This is one example of a countermeasure to an attack on a vulnerable system.

The focus of this work was to create a testbed for evaluating attacks and countermeasures against automated vehicle systems. A set of automated vehicle was designed to provide a platform to evaluate the security and reliability of individual cars and cars as a group. The development of this system is presented in this work. Vehicle guidance requirements and algorithms are discussed. A method to determine distance using a single camera is proposed. Finally, the vehicle system was evaluated as a viable method to validate an attack on a highway system.

Checksum

8aa2dc846656a2e7c79e9bd40b5ba6c8

Share

COinS