Improved Automotive CAN Protocol Based on Payload Reduction and Selective Bit Stuffing

Abstract

In this research paper, the implementation strategies of automotive controller area network protocol are investigated and the short messaging scheme with selective bit stuffing method to improve the effective utilization of its bandwidth has been proposed. There would be a sharp decrease in the performance of traditional CAN protocol because of considerable increase in the number of ECUs (Electronic Control Units) and infotainment gadgets connected in the vehicle architecture. The demand for safety, emission, diagnostics and comfort norms has steeply increased the number of messages in the 250 Kbps CAN network as the computational power of ECUs has gone up. To overcome this problem, the short CAN method has been proposed and the work is benchmarked with SAEJ1939 Heavy commercial vehicle CAN standard. The Matlab Simulink based short CAN has been modeled and the performance of the proposed system has evaluated using virtual instrument cluster. Experimental results have shown that compared to the traditional CAN, the proposed method has reduced the worst case response time of CAN extended frame from 160 μ sec to 144 μ sec. Selective bit stuffing technique has reduced the impact of bit stuffing over the payload and improved the utilization factor for the CAN bus without affecting the CAN message ID properties. The proposed algorithm has been modeled and simulated using CAN Matlab model Simulink and it has been verified using virtual CAN tool and real time CAN bus hardware.