The CAN protocol has served the automotive and related industries well for over twenty-five (25) years now; with the original CAN protocol officially released in 1986 followed by the release of CAN 2.0 in 1991. Since then many variants and improvements in CAN combined with the proliferation of automotive onboard microprocessor based sensors and controllers have resulted in CAN establishing itself as the dominant network architecture for automotive onboard communication in layers one (1) and two (2). Going forward however, the almost exponential growth of automotive onboard computing and the associated devices necessary for supporting said growth will unfortunately necessitate an equivalent growth in the already crowded wired physical infrastructure unless a suitable wireless alternative can be provided. While a wireless implementation of CAN has been produced, it has never obtained real traction within the automotive world. Other alternative methodologies for providing wireless connectivity have been much more pervasive and accepted, but none of them provide anything more to CAN interfaces than a CAN-to-Wireless Bridge; with Wi-Fi, Blue Tooth, and GSM being the primary wireless network architectures bridging to CAN. What is proposed within this paper is more than simply a wireless extension of CAN in that it does more than extend CAN into the wireless domain (as was the case with CANRF). As pure wireless CAN with no accommodations for heavy utilization would only exacerbate CAN's primary deficiency of starving out lower priority messages; since there would be no way to isolate devices in sub-networks as could be done with a wired infrastructure. Rather, the proposal within this paper would attack this deficiency head-on by modifying the newly defined wireless network protocol and architecture, DQWA (Distributed Queuing Wireless Arbiter) to not only extend CAN into the wireless domain, but also addresses CAN's more prominent shortcomings. DQWA is a solution that provides both security and reliability within a wireless framework, while maintaining CAN's distributed network communication methodology and implicit avoidance of single points of failure within the network.

Presenter
Graham Campbell, Arynga
Jon Barton Shields, Arynga