This paper focuses on addressing the secure localization challenge faced by autonomous articulated vehicles when subjected to perception sensor attacks. Autonomous vehicles are generally equipped with multiple perception sensors with some for localization and some for obstacle sensing. Based on the redundancy provided by these sensors, a new attack detection and recovery approach is proposed to ensure secure pose estimation for autonomous articulated vehicles. The proposed approach encompasses three primary components. Firstly, redundant sensors are calibrated to make sure their pose observations are with respect to a common coordinate frame. Secondly, a cross-checking based monitoring algorithm is developed to capture anomalies and determine their origins. Lastly, a recovery strategy is designed to maintain continuous secure pose estimation, free from interference from the attack. The proposed approach has been validated on a real autonomous articulated vehicle. Four common types of attacks, namely deny-of-service (DoS), lateral-offset (LO), gradual-drift (GD), and replay, have been considered. Experimental results validate the efficacy of the proposed approach.
Secure Localization of Autonomous Articulated Vehicles: Attack Detection and Recovery
2023-09-24
2856174 byte
Conference paper
Electronic Resource
English
Autonomous underground tramming for center-articulated vehicles
British Library Online Contents | 2008
|Wiley | 2018
|Engineering Index Backfile | 1968
Simulation of Intelligent Convoy with Autonomous Articulated Commercial Vehicles
British Library Conference Proceedings | 2003
|Simulation of Intelligent Convoy with Autonomous Articulated Commercial Vehicles
SAE Technical Papers | 2003
|