Data and Model-Based Approaches in Fault Detection and Identification for Connected Vehicles

In recent years, significant progress has been made in the application of data-driven, learning-based approaches to fault detection in distributed networks. These methods are optimized for quickly detecting and identifying faulty instruments, whether originating from within a single vehicle or from a network of connected vehicles. This paper provides a preliminary review of typical Fault Detection and Identification (FDI) techniques, with a focus on platoons of vehicles arranged in a rectilinear formation using a leader-follower architecture. Specifically, this paper discusses the advantages and disadvantages of data-driven versus model-based methods for addressing the FDI problem. In particular, the main characteristics of a novel immunity-based bio-inspired data-driven technique are highlighted, and numerical simulations of a multi-vehicle system under normal and faulty conditions are presented to support the discussion.