Implementing Vibration-based Communication Channels for Secure Implantable Medical Devices
Abstract
Modern implantable medical devices (IMDs) often support wireless connectivity for post-deployment monitoring, diagnostics, and control. As IMDs have evolved into full-blown wirelessly-connected embedded systems, an unfortunate side effect has been the increased possibility of security attacks, with security vulnerabilities being reported in several commercial IMDs such as pacemakers, insulin delivery systems, etc. To securely communicate with an IMD, an encrypted wireless channel must first be established between the IMD and an external device. Recent work proposed a system (called SecureVibe) for sending a cryptographic key from a cell phone to an implanted medical device using a vibration-based communication channel. SecureVibe exploits the fact that many IMDs now contain accelerometers, and cell phones have vibration motors, meaning that vibration can be used as a secure side channel for communication between them. This thesis builds upon the SecureVibe architecture and presents a fully self-contained system for performing key negotiation between an IMD and a cell phone using a vibration-based communication channel. Cryptographic keys are encoded into vibration patterns using on-off keying and transmitted using the phone's vibration motor. The vibration signal is measured and decoded by the IMD. Since the vibrations are attenuated by the human tissue between the cell phone and the IMD, an efficient error-correction scheme that allows for ambiguous bits is used with the ambiguous bits being subsequently resolved by the cell phone. The decoded cryptographic key is then used to encrypt the wireless channel and secure all future radio transmissions. This thesis also presents a method for calibrating the parameters used for bit decoding in the IMD to account for differences in the physical properties of the human tissue. The effectiveness of the approach is demonstrated using synthetic human tissue, which is a more rigorous model than past work. For standard thicknesses of synthetic subcutaneous fat, the system can negotiate a 128-bit key in 20 seconds 90% of the time without requiring precise positioning of the phone relative to the IMD.
Degree
M.S.E.C.E.
Advisors
Raghunathan, Purdue University.
Subject Area
Computer Engineering
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.