https://www.mdu.se/

Modern embedded systems, particularly in the automotive domain, have seen significant advancements in functionality and complexity. This has driven demand for high data-rate sensors such as cameras, radars, and lidars, which generate vast amounts of data that require transmission with low, predictable latencies. However, traditional onboard communication protocols in the automotive domain, such as Controller Area Network (CAN), have limited support for these requirements. The IEEE Time-Sensitive Networking (TSN) standards have emerged as a solution, providing high-speed, low-latency communication that can be used as a backbone network connecting nodes and networks in the system. The challenge lies in fully utilizing TSN while maintaining compatibility with low-cost legacy CAN systems. This thesis aims to address the challenges of integrating CAN and TSN networks. We investigate various design techniques for the gateway that connects a CAN domain to a TSN domain, ultimately proposing the interface architecture for a CAN-TSN gateway. During our investigation, we identified the lack of timing analysis for the next generations of CAN, namely CAN Flexible Data-rate (FD) and CAN Extra Long (XL), and developed timing analysis for them. We further develop the analysis for the CAN-TSN gateway. As part of the thesis and working towards a scheduling method for TSN traffic sent by a CAN-TSN gateway, we extend a heuristic algorithm to schedule TSN traffic, increasing the schedulability of lower-priority traffic,  particularly in scenarios involving the use or non-use of frame preemption in TSN. Finally, we demonstrate the proof of concept by implementing the timing analysis in an industrial tool suite and through an industrial use case utilizing the gateway. This demonstrates the feasibility and potential benefits of integrating CAN and TSN networks.