openETCS: Rail Systems Engineering Inspired by PolarSys
The vision is compelling: You enter a train in Portugal, which takes you through all of Europe, the only thing indicating a border crossing is your cell phone, telling you about the roaming fees in the country you just entered. And this vision is already a reality in some parts of Europe. But behind the scenes, the industry is still struggling with incompatibilities of the various implementations of ETCS, the European Train Control System. While it exists, it is prohibitively expensive and does not scale well yet.
The itea 2 project "openETCS" is changing this by a shift that is considered radical in the rail industry: To lower costs, the software component of the on-board unit (the hardware that is installed on the train) will be modeled using formal methods, following the "open proofs" approach. The idea is that vendors can reuse these artifacts, thereby lowering costs and improving quality at the same time.
Open Proofs goes beyond open source: Not just the source code, but all artifacts of the systems developement process are open, including documentation, models and tools. Besides the known benefits of open source (inspection by all, sharing of effort), this makes long-term support less risky, as no vendor-lock-in takes place.
This talk will report on the progress of the openETCS project and present its results, in particular the openETCS toolchain, which is inspired by PolarSys (formerly Topcased), and which reuses some PolarSys components, in particular Papyrus. After outlining the challenges that are particular to the rail industry, we will provide a demo of the tool (which can be downloaded at openetcs.org/tool). We will explain why we still have to rely on some closed-source components (Scade in particular), how we integrate such components in an open Eclipse-based environment, and how these components will be eventually replaced by open source equivalents