Decentralized Most Permissive Observer Architecture for Brake Leakage Diagnosis in Automotive Systems

One of the most important parts enabling the safety of contemporary cars is brake systems. Often recognized as one of the most prevalent vehicle-related factors causing fatal collisions are they. Usually, brake fluid leaks influencing the performance and responsiveness of the whole system cause the faults in this kind of component; hence, manual maintenance techniques are necessary to find the source of the problem. This work presents a unique Discrete Event System (DES) model for fault diagnosis using a distributed Most Permissive Observer (MPO) architecture able to detect brake fluid leakage faults. The developments were confirmed using an enhanced virtual environment based on a simulation model of the brake mechanism using Simscape and evaluating the event-driven fault diagnosis system with Stateflow. In order to evaluate test cases that combined edge values, a “Design of Experiment” method is applied employing a factorial design of 3k, where k is the number of levels of a factor in the factorial experiment (in this paper k is 4). Over a broad spectrum of simulated scenarios including fault locations, lower pressures, and single and multiple fault combinations, the effectiveness of the proposed DES model for fault diagnosis is shown. The proposed approach shows its efficiency under several operational conditions of the braking system since it regularly detects the injected fault conditions.