Expanding transmission capacity of can systems using dual communication channels with kalman prediction

The controller area network (CAN) protocol is widely used for in-vehicle network (IVN) systems, and many automotive companies also use the CAN in chassis network systems. However, the increasing number of electronic control units (ECUs) dictated by the need for more intelligent and fuel-efficient functions requires an IVN system with a greater transmission capacity and less network delay. Automotive companies have tried several approaches such as segmenting CAN systems and developing time-triggered protocols. This paper presents a practical method for increasing the transmission capacity and reducing the network delay in CAN systems using dual communication channels with a traffic-balancing algorithm based on Kalman prediction to forecast the traffic on each channel and allocate frames to the one that is most appropriate. An experimental testbed using commercial microcontrollers with two or more CAN protocol controllers was used to demonstrate the feasibility of the Kalman traffic-balancing algorithm. Experimental results show that the traffic-balancing CAN system with Kalman prediction reduced the transmission delay of all priority messages compared to that of a simple method, such as a channel-switching CAN, without sacrificing the performance for high-priority messages.