A Practical Tutorial on Systems Integration 

Director of Systems Engineering
South Coast British Columbia Transportation Authority (TransLink)
New Westminster, BC

Integration is often the source of consternation as it is the final hurdle that projects must overcome before reaching completion. But what is it, and how can projects successfully undertake integration in large transportation projects? On February 17, APTA hosted a webinar where a panel of industry experts discussed the importance of integration. It covered the roles and responsibilities in the project lifecycle, interface management, applicable case studies from a transit agency, and practical steps teams can take to deliver their next project.

For the purposes of this article, integration is simply connecting two pieces of technology that have different functions and enabling them to function as a single unit. This is a common concept in the modern world, be it connecting your mobile phone to your vehicle via Bluetooth or connecting your laptop to your printer. However, in transportation projects, there is more complexity, as different technologies may not necessarily be designed to function with each other, and the technology can be dated. Imagine trying to record your Xbox games onto video cassettes (Google this if you don’t know). The same challenges apply to fare collection systems, security and safety technologies, and telecommunications equipment, but the stakes are certainly higher.

Successful integration starts early in the project lifecycle and requires clear identification of roles and responsibilities. Understanding the various components of the project and how they need to interact with each other is a critical first step, followed by identifying a responsible party to ensure all the interactions (interfaces) are properly managed and charged with resolving interface issues as they arise. This is best started already at the requirements definition phase, as it may be time-consuming and/or costly to address interface incompatibilities if they are not identified until final design or implementation.

Once the interfaces are identified, it is then important to define the processes in which the integration tasks will be executed. For example, who is responsible for each side of the interface, how will teams collaborate, what tests need to be performed, and what are the acceptance criteria? It is easy to assume that interfaces are identical to ones completed in the past, so it is important to validate these assumptions and to identify where there may be “hidden” interfaces within the scope of work. It is also important to tailor the amount of rigor required to address an interface—going to the store to buy an adaptor for your phone is less complex than asking competing vendors to fix a firmware bug.

It is also critical to examine the interfaces at an appropriate level of detail. Mapping the interfaces, using block diagrams, and taking a systematic approach are helpful strategies to avoid any scope gaps. Remember that the size of the component does not dictate the number of interfaces. A single programmable logic controller (PLC) can have several complex interfaces with signaling, safety, and telecommunications equipment.

Once interfaces are implemented, they must be adequately tested to verify that they meet design requirements and to validate that they meet user needs. Some common issues include unexpected challenges in brownfield conditions and different performance characteristics in production as compared to the lab. This is why it is critical to identify a party responsible for resolving interface issues early in the project. If no such party is defined, the owner, typically the transit agency, becomes the de facto party responsible.

The unfortunate truth is that integration issues can manifest in very visible and public failures, often well-documented in the media. Too often, we hear about “bugs” in the signaling software, train doors not aligning with platform doors, and passenger information displays showing erroneous information.

To avoid these situations, projects need to start integration early and empower a single, competent party to manage interfaces and resolve issues. Clear communication between various parties, including external projects, is key, as is respecting project control gates. Establishing robust interface registers, and testing interfaces rigorously can assist with technical development. Lastly, bringing interfaces into revenue service incrementally rather than bringing everything online in a single instance—the “big bang”—helps with troubleshooting as it is easier to identify which part of the interface is causing issues.

Remember, integration is not necessarily difficult, but it won’t happen on its own. For more details, watch the webinar on APTAU. Happy integrating!