TrackSafe

Wayside worker protection system, deployed to carriers including NJ TRANSIT and MARTA. Tech lead at Bombardier Technology Solutions. IET Innovation Award and Altran UK Foundation Award, 2013.

role Technical Lead
dates 2013 — 2023
stack rfid · wireless · embedded · middleware
status deployed

What it is

A wayside worker protection system for passenger rail. TrackSafe uses RFID and wireless networking to make the presence of track workers visible to approaching trains and to dispatch, and to make the presence and movement of trains visible to workers on the ground. It turns a procedural safety regime — radio calls, paper protection forms, verbal confirmations — into an active technology layer that enforces the same rules without relying on any single person getting it right every time.

I was the technical lead on TrackSafe. The system shipped to passenger-carrier clients including NJ TRANSIT and MARTA. The Federal Transit Administration sponsored Phase II research and demonstration (Report 0194) on the system’s wayside worker protection capabilities.

Why it exists

Track workers occupy a position where a small number of small mistakes can be fatal — dispatcher error, radio misunderstanding, a train arriving on a track that was believed to be clear. The job of the system is to make those mistakes impossible at the technology layer rather than relying on procedural compliance alone. The technology and the procedure are the same discipline: they both describe what must be true before the train moves. The system’s contribution is to verify it rather than report it.

The award context

In 2013 the Institution of Engineering and Technology (IET) named TrackSafe one of fifteen Innovation Award winners globally, and the Altran UK Foundation recognized it with its own award the same year. International recognition matters less than the change in posture it reflected — that wayside worker safety was being treated as an engineering problem with a defensible solution, not a training problem with unavoidable residual risk.

What it taught

The discipline that carries into my current work is this: design the enforcement layer first, and design it deterministically. Trains do not move until the system agrees they can. That same posture is what the regex cascade in StrongAfter is doing, at a very different scale — hard overrides that run before any probabilistic reasoning, because the cost of the probabilistic layer being wrong is too high to absorb.

Awards

Sources

← All work