Secure Embedded Systems
As embedded systems take on larger roles in our lives, security is becoming a major concern—not only because the stored information is sensitive, but also because critical systems like vehicles and power plants have hefty consequences of failure. ECE researchers are tackling security of embedded systems on multiple levels.
Paul K. Ampadu
Thidapat (Tam) Chantem
Cameron D. Patterson
Ryan M. Gerdes
ECE researchers are designing a semi-automated, efficient, secure emergency response system that will provide guidance on the best action to take in an uncertain, potentially dangerous traffic situation. In some cases, the system will even take the appropriate action without human intervention. The system helps reduce the time needed for emergency vehicles to reach their destinations, while increasing the safety of non-emergency and emergency vehicles alike. We are also taking a real-time system approach to design efficient runtime algorithms to provide maneuver guidance to automated vehicles in both highway and urban settings.
Another project involves a custom microprocessor chip that can detect hardware tampering. This chip is designed for embedded, portable devices that a hacker can physically manipulate, such as a chip-enabled credit card or mobile phone. This chip will detect tampering and can be programmed to actively respond to it.
ECE researchers are deriving real-time task models that capture the dependencies between the physical environment state and timing parameters, and allow for tighter, less pessimistic timing guarantees. We have found that it is possible to make resource management decisions quickly without impacting accuracy, thus enabling the design of systems with size, weight, and power constraints.
There are many potential benefits of connected and autonomous vehicles, including increased fuel-efficiency. Using real-time information from connected and automated vehicles, we can holistically optimize vehicle operations at all layers, including routing, speed planning, vehicle dynamics control, and engine control. However, doing this requires optimization techniques that are orders of magnitude faster than existing approaches, and we are making significant progress towards this goal.
With the rise of the Internet of Things, the need is growing for secure devices in small packages—which can be more vulnerable to certain kinds of attacks. One project investigates common hardware attacks, such as side-channel and fault injection attacks, on reconfigurable platforms like field programmable gate arrays (FPGAs), Systems-on-Chip (SoCs) and microcontrollers. We are then designing low-cost countermeasures to secure sensitive information.
Our researchers are developing a platform for tamper-resistant embedded software, focusing on two forms of hacking that target the processor hardware. Side-channel analysis exploits the physical effects of computing as a means to extract internal, secret details. Fault injections probe the processor hardware for similar internal details. Our technique creates multiple virtual processors in a single physical platform. The resulting software will be resistant to both attacks. The platform will also demonstrate novel logic minimization algorithms.
Our infrastructure is relying more and more on autonomous systems and machine learning. Machines, however, don’t always react predictably to scenarios they’ve never encountered before. Before our systems can become fully autonomous, we need assurance that they won’t fail. ECE researchers are working on multiple ways to provide this assurance at the hardware level for applications such as unmanned vehicles.