Cyber Physical Analysis of System Software Survivability by Stimulating Sensors on Drones
Abstract
We developed an acoustic injection testbed for MEMS gyroscopes and accelerometers; this testbed enables automated testing of the influence of compromised sensor values on drones, without the risk of physical damage to the drones. Using this testbed, we conducted rigorous experiments and discovered that sampling jitter is the essential factor influencing drone crashes during attacks. Notably, sampling jitter has not been discussed in previous studies. During our investigations, we discovered that sampling jitter produces noise-like signals. Based on this finding, we propose a novel prototype recovery system, UNROCKER, and demonstrated its capability through various experiments including real-world scenarios on physical sensors.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 29, 2022
- Accession Number
- AD1183688
Entities
People
- Dohyun Kim
- Dongkwan Kim
- Jaehoon Kim
- Yongdae Kim
Organizations
- KAIST