A Multi Fault Tolerant Control Architecture in Starter Generator System

An integrated starter generator is widely used in vehicles. In this study, the goal is to maintain an integrated starter generator system working accurately and respond instantaneously during sensors failure. A redundant signal is established by signal reconstruction and estimation method. To switch among multi-signal sources, a signal arbitration is done based on random forest algorithm. The accuracy of fault diagnosis is higher than 99.82%, and 14 kinds of signals failure can be identified.