Interval-based Solutions for Reliable and Safe Navigation of Intelligent Autonomous Vehicles

The transportation systems reliability is addressed in this work. A comprehensive comparison between the probabilistic and the interval-based uncertainty handling approaches for autonomous navigation has been detailed. Based on this comparative study, a set-membership safety verification technique that monitors the correlation between variables has been proposed to achieve an optimal uncertainty assessment. Further, a Principle Component Analysis (PCA) diagnosis process has been extended to handle interval-data. Finally, a strong link between the proposed automotive diagnosis and risk management has been constructed to ensure a high robustness to uncertainty. The proposed interval-based solutions have been integrated on an Adaptive Cruise Control (ACC) system. Simulation results prove the proposed diagnosis and risk management efficiency in handling uncertainties and faults.