Radio Frequency-Based Microcontroller Anomaly Detection
Technical Report,01 Sep 2014,24 Mar 2016
AIR FORCE INSTITUTE OF TECHNOLOGY WRIGHT-PATTERSON AFB OH WRIGHT-PATTERSON AFB United States
Pagination or Media Count:
The research presented here focuses on applying the Correlation-Based Anomaly Detection CBAD process to a Microcontroller Unit MCU as a means of detecting an anomalous behavior by verifying that the device is behaving normally. This research utilizes a previously developed method to collect Unintentional Radio Frequency RF EmissionsUREs Texas Instruments MSP430F5529 microcontrollers. Once captured, the time domain signals are processed and used as one of two inputs into the CBAD algorithm, which generates a statistic to be compared to a threshold. Like-device performance met the arbitrary benchmark of Equal Error Rate EER 10 . Cross-device performance met the benchmark when the reference signal originated from the same manufacturing lot, but did not meet the benchmark when the reference signal originated from a different manufacturing lot.
- Computer Systems Management and Standards