Heterogeneous Platooning Vehicle with Robust Sensor Fault Detection and Estimation

The research on ACES, which stands for Autonomous, Connected, Electric, and Shared, brings a significant impact on safe driving, fuel efficiency, traffic smoothness, and society's quality life, especially for the aging population mobility. One of the ACES services is the platooning vehicle, where a group of vehicles moves together with the cooperative adaptive cruise control (CACC) and other safety features. Sensors and vehicle-to-vehicle (V2V) communication are a crucial part of succeeding in the implementation of platooning vehicles. In this work, we consider radar and camera sensors as the primary concern. Those sensors provide relative distance and velocity between vehicles where is the key to maintain a safe distance between vehicles during platooning. In practice, both sensors are subject to several critical issues, such as faults, malfunction, and sensor degradation. Besides, in the transmission and processing of sensor data, errors may occur due to delay, noise, and cyber-attacks. These issues may affect the safety and stability of an ACES system. For a platoon operating under CACC, the string stability may be addressed. Therefore, this paper aims to solve sensor fault issues with the proposed strategy for fault detection and estimation in the presence of noise, uncertainty, and perturbation. The proposed strategy integrates parameter estimation and the state estimation approach. The modified, extended recursive least square (MERLS) with exponential forgetting factor and Kalman filter is considered to deal with the system have noise. The model is obtained by MERLS and used by the Kalman filter to detects the sensor's fault based on the residual. When the sensor fault detected, a robust estimation from the Kalman filter is employed as fault-tolerant control that guarantees impressive performance after sensor fault occurs, subject to safety and string stability.