This special issue focuses on theory and application of diagnostics and prognostics (DnP) analytics and methods for condition monitoring, anomaly detection, and health management of automotive systems. With elevated system complexity in electrified powertrains, connected vehicles, and automated driving; new DnP analytics and technologies will support accurate and reliable estimation and predictions of state of health for these complex vehicular systems. The increased capability of the vehicle to exchange data with passengers, surrounding infrastructure, and other vehicles has been both an enabler and provide motivation for developing advanced vehicle health management (VHM) features. Robust VHM systems can mitigate degradation of vehicle components by shifting operating points and even enable reconfigured operation until repairs can be made. At the same time, the development cycle for these new and more complex systems needs to be shortened in order to reduce development cost.
Recent advances in systems and analytics tools such as machine learning and cloud-based computation can leverage the classic model based DnP algorithms to address the complexities, and performance requirements of future VHM systems. Extensive researches have been done to analyze and predict performance of machine learning algorithms that is required for reliable modeling, estimation, and detection algorithms. Re-configurable and stochastic reinforcement learning techniques have also been successfully utilized in recent research to mitigate degradation in autonomous systems. Still, much work is needed to ensure robust performance of the vehicle with any newly developed component and algorithms. This special issue provides critical information on these important matters.