CARMEN project adds new dimension to Capito’s PhD research

female in green shirt and glasses with long black hair, smiling

Linda Capito is an electronics engineer turned safety researcher. She has her master’s in electrical engineering from The Ohio State University and is currently working on her PhD.

Capito has always been interested in safety and autonomous systems, specifically autonomous vehicles. She saw growth in autonomous vehicles and knew it would impact a lot of people in the near future. Her research involves operational safety and adversarial testing for autonomous vehicles, which means she generates specific situations where the vehicle may be in danger and tests how the vehicle reacts.

She is set to graduate with her PhD this year, thanks in part to her involvement with the Center for Automated Vehicles Research with Multimodal Assured Navigation (CARMEN) project. This project aims to find concrete solutions for transportation cybersecurity risks that may be associated with autonomous vehicles.

Her advisor, Professor Keith Redmill, is one of the Principal Investigators of CARMEN. “Since I am Professor Redmill's student, and I had previous experience in the area of designing adversarial tests for vehicles, it felt like a good opportunity to extend my work,” she says.

CARMEN incorporated a new dimension into her research: devising challenges for automated systems considering attacks on positioning, navigation and timing (PNT) sensors. Capito had previously developed algorithms to control Principal Other Vehicles (POVs) in order to create adversarial testing scenarios for Subject Vehicles (SV) at the operational level.

Subsequently, she introduced PNT sensors into the equation, thereby adding another layer of complexity to the problem. Attacking the readings from the sensors or the actuators in the vehicles can potentially activate detection safeguards within the vehicles. In essence, if the sensor readings differ significantly or exhibit excessive variation between consecutive time intervals, the detection algorithm may raise an alert, thereby exposing the attack. Therefore, her current objective is to determine how the vehicles under investigation can deviate from the intended path or navigate toward an incorrect destination without triggering the detection systems.

Capito has always been interested in the safety of autonomous vehicles; she saw the growth of autonomous vehicles and knew that they would impact a lot of people in the upcoming years. She says, “I wanted to know how to make these vehicles safer by analyzing the risks associated with them and then doing something about it.” The CARMEN project, in addition to Linda’s work, will tackle some of these pressing issues.

Written by Cassie Forsha, CAR writing intern