A two layers diagnosis system is proposed for a passenger car gearbox system. Conventional physical constraints based diagnosis is placed on the first layer. Model based diagnosis and specific symptom diagnosis are built on the second layer. Diagnosis rules for the first layer are described in a tabular form, then an algorithm is described with a symbolic coding tool. The residual method was selected for the model based diagnosis. The proposed strategy was implemented into time domain simulation to evaluate accuracy of the FDI system. A simulation study illustrates the proposed diagnosis process and shows effectiveness of the method.

View full abstract

The Controller Area Network (CAN) is a widely used in-vehicle network and its schedulability is particularly important. In a CAN system, implementing the priority queues is better than FIFO queues with respect to the schedulability, which is widely recognized by users. In this paper, we investigate whether and how much implementing the priority queues is better than FIFO queues when the offset assignment for CAN messages is conducted. Our results show that usually the priority queues are better, whereas the FIFO queues are better in some specific situations. Furthermore, we propose a new scheduling method for obtaining the advantages of both queues, which improves the schedulability effectively.

View full abstract