Hybrid Network Architecture for Inter-ECU Communication

Abstract

With the growth of the number of automotive electronics and the development of autonomous driving technology, the demand for high-bandwidth of in-vehicle networks increases significantly. Traditional technologies such as CAN and FlexRay are facing the challenge of bandwidth limitations. For example, cameras loaded on the side and back of the vehicle are used to detect objects in the blind spots, which needs a lot of bandwidth to transmit images and videos to the ECU. Traditional wired vehicular networks cannot afford such a large amount of data transmission. This is reasonable because it is difficult to predict the bandwidth demand of today’s automotive network at the time when traditional technologies are developed. To improve network communication performance, current vehicle network architectures use a combination of several different network protocols to meet the growing bandwidth requirements. Also, for communication reliability, many network protocols working together form a redundancy mechanism. An obvious example is that wireless networks can act as channel backups to carry on important data transmissions when critical transmission failures happen on wired networks. Beginning with this point, the thesis studies a hybrid network communication within vehicle ECUs. This work is basically divided into four parts. First, it introduces the background and related works. Second, a prototype of hybrid network architecture is demonstrated. It is developed for in-vehicle communications using several developed network protocols, including Ethernet, dual-band Wi-Fi, CAN and BLE. Four experimental results of single-interface benchmarking tests are carried out. Next, a preliminary hybrid network algorithm is carried out, which aims to improve the whole system performance in terms of data transmission reliability and bandwidth utilization. Redundancy transmissions are the basis of this algorithm. Finally, it presents several fault detection experiments which are used for developing detection rules. Based on these rules and the algorithm which is mentioned above, a preliminary recovery mechanism is developed and verified by using testbed, to simulate how the system performs when critical transmission failure happens.