LiDAR sensors play a central role in the development of autonomous vehicles. LiDAR is a sensor that uses laser light to obtain 3D information over a wide range of areas. It can precisely detect the surrounding environment of a vehicle and measure the distance to objects with high precision, greatly improving the safety of autonomous driving. However, new security challenges have arisen, such as exploiting vulnerabilities in LiDAR and injecting false data with attack lasers. This issue requires a thorough analysis of LiDAR vulnerabilities and finding effective countermeasures against potential threats.

　For the first time, the research group conducted a large-scale measurement study on nine types of LiDAR sensors, old and new, and evaluated sensor vulnerabilities and safety. As a result, the attacker irradiates the sensor with a high-frequency laser (named "HFR (high-frequency laser removal) attack"), which disrupts the measurement of the target LiDAR like radio wave jamming and erases a wide range of objects on the sensor. It has been shown that attacks such as debugging or injecting falsified data are possible.

　In the future, the research group will develop technologies to improve the resistance of LiDAR sensors to malicious laser attacks, new algorithms to prevent the injection of false data, and improve safety by combining different types of sensors (radar, cameras, etc.). Focus on developing defenses to counter vulnerabilities. Ultimately, the company aims to strengthen the security of self-driving vehicles and thereby contribute to the safety and security of society as a whole.

Paper information:[NDSS Symposium 2024] LiDAR Spoofing Meets the New-Gen: Capability Improvements, Broken Assumptions, and New Attack Strategies