edu.cn). FLEXIBLE FUZZY PRIORITY SCHEDULING OF THE CAN BUS

With the increasing complexity of distributed real-time systems, the need for improved CAN bus performance is continually increasing. Nor-mally, a scheduling scheme with static-priority has low network schedula-bility/utilization; while using dynamic priority will improve the QoS of the network at the cost of a narrow service range or a high overhead, compared with the fixed priority scheduling schemes. Actually, because of the fluc-tuation of network traffic, these priority policies may not guarantee flexibil-ity for different kinds of messages. Based on the broadcast nature of the CAN bus, a closed-loop fuzzy scheduling scheme is proposed in this paper. Compared with the dynamic priority schemes, this fuzzy scheduling scheme uses fewer bits to encode fewer priority levels, which widens the service range of the network without increasing overhead. Based on game theory, a fuzzy parameter updating algorithm for the fuzzy scheme is developed to improve the adaptation of the scheme, which guarantees the required QoS of the network even with traffic fluctuation. Simulation results well demon-strate the abilities of the fuzzy scheme to guarantee high schedulability for real-time messages, as well as the fairness and the same QoS for non-real-time messages in networks