Cryptographic Strength and Machine Learning Security for Low Complexity IoT Sensors

Abstract

Hacking of every-day sensors is now a common problem with a disproportionate impact over life-critical assets. It is precisely because they are low-cost and low-complexity that these sensors make such easy targets that this problem will grow exponentially as sensors become more indispensable to modern life. This paper examines the impact of sensor hacking and offers a two-phase approach to security: a low-cost authentication scheme and a novel Machine Learning approach using a Malware Predictive Interpreter.