Hybrid ARQ for DC-DC Converter Noise in Controller Area Networks

Controller area networks (CANs) are primarily used for communications among electronic devices in an automobile. The noise resistance mechanism defined by the CAN standard accommodates frame/bits errors caused by relatively small noise. However, when it comes to the modern electronic vehicles (EVs) and hybrid vehicles (HVs), the noise level caused by high voltage switching in DC-DC converters could be very large. This significantly reduces the performance of the CAN protocol. To tackle this issue, we first perform noise injection experiments to characterize noise patterns generated by a DC-DC converter. The experiments show that the noise level caused by voltage switching varies from time to time. Therefore, in this paper, we propose an application layer-based hybrid ARQ protocol, which consists of the automatic repeat request (ARQ), forward error correction (FEC), and HALT modes. Depending on the noise level, the proposed protocol selects an appropriate mode. The first prototype implemented on a real CAN node show that our protocol can adaptively switch its mode according to the noise level.