Radio-Frequency Interference for Fault Injection and Sensor Manipulation
Abstract
Cyber-Physical Systems (CPS) couple sensors and actuators with automation logic to control and monitor physical processes. In this project we consider those CPS that must perceive objects of interest and react to their presence; e.g., an automated vehicle equipped with a vision system must be able to detect traffic obstacles and maneuver the vehicle accordingly, while surveillance systems will often employ cameras in an effort to detect intruders and issue an alarm. Our goal is to enable the subvertion of the sensing and computational components of systems so as to prevent the identification and tracking (surveillance) of mission personnel and allow battlefield commanders to extract sensitive information to thwart the systemsÕ proper operation. We propose to investigate two attack vectors that utilize the same attack mechanismÑviz., specially crafted radio-frequency interference (RFI)Ñto allow a battlefield commander to subvert perception-based CPS, and cryptographic hardware engines, in general, by: 1) manipulating the analog sensing chain to alter or obscure the sensed signal, and 2) introducing erroneous inputs or state transitions into an encryption module, resulting in biased output, to recover sensitive information (e.g., encryption keys). Our approach is end-to-end, in that we seek to control both the input to the system (sensor attacks) and its digital output (e.g., breaking of authenticity and confidentiality protocols), to increase the likelihood that the system can be subverted and/or controlled.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Oct 07, 2021
- Source ID
- W911NF2110320
Entities
People
- Ryan Gerdes
Organizations
- Army Contracting Command
- United States Army
- Virginia Tech