Secure FMCW LiDAR Systems with Frequency Encryption

Abstract

Robust and secure ranging is among the most vital capabilities demanded by future autonomous vehicles and robotics for precise navigation and avoiding collisions. Light detection and ranging (LiDAR) is a promising 3D imaging technology for this aim. However, the security vulnerabilities of LiDAR systems can impose critical threats to human safety and security, similar to other types of sensors. While LiDARs are becoming standard technologies for self-driving cars, their security aspects have not yet been studied well so far. In this paper, we will first summarize various security attack scenarios against different LiDAR types. We focus on beam-steering and frequency modulated continuous wave (FMCW) LiDAR systems as they have been considered the most secure LiDAR systems proposed so far. We will show that an attacker can reverse engineer the victim's LiDAR system and build a spoofing system using commercially available electro-optical components. To do so, we will develop an electro-optical co-simulation framework in MATLAB Simulink and use that to study the feasibility of the spoofing attack in today's FMCW LiDAR systems. Finally, we propose the frequency encryption technique as a countermeasure to mitigate the possibility of spoofing FMCW beam-steering LiDAR systems. The proposed approach can ensure the security of future FMCW LiDAR systems without compromising functionality or accuracy.