Seattle Transit Blog Argues Sound Transit Treats Automation as Unproven for Ballard Extension

Automated train control is not experimental technology: Boston was testing it in 1901, Vancouver’s SkyTrain has run driverless since 1985, and new lines in Montreal and Hawaii shipped with it from day one.

Yet, as Seattle Transit Blog argues, Sound Transit still treats automation as unproven — even as it weighs it for the Ballard extension.

That framing is the news here. The blog’s latest history of automated train control makes the case that Puget Sound’s light rail spine, fully grade separated in the north, could run driverless from Everett to downtown — with tighter headways, faster turnarounds, and less trackside hardware.

Seattle Transit Blog reported that Sound Transit evaluated automated light metro modes as far back as 2014, choosing conventional light rail to keep the Downtown Seattle Transit Tunnel and at-grade segments in play.

The gap between perception and deployment is striking. Airports like SeaTac and the JFK AirTrain have run automated systems for decades. So have whole metro networks.

“Some people still consider automated train control (ATC) as unproven even though many airports (SeaTac, JFK AirTrain…) and subways use this technology such as Vancouver Skytrain and new lines on Hawaii and in Montreal already use the technology.”

Why trains are easier than robotaxis

The engineering case is straightforward. A train on a dedicated right of way needs no lidar, no camera arrays, no neural perception stack. The “sensor” is the track itself: track circuits, axle counters, and a signaling layer that knows exactly where every train is. That’s why automated train control predates the transistor — the problem is deterministic, not probabilistic.

Platform screen doors close the remaining risk surface. Boston, Berlin, Paris, and Tokyo all proved the concept long before software made it elegant.

From Paris 1966 to Nürnberg’s mixed traffic

The milestones stack up quickly. Tokyo started limited automated operation in 1962. Paris ran a short automated line from 1952 to 1957, but lacked funds for upgrades until its first main line was converted in 1966. London’s Victoria line reached fully automated operation in 1967 — initially with a driver on board.

True driverless service arrived in Kobe, then Lille, Barcelona, and the airport shuttles. Nürnberg did something harder in 2008: mixed operation, running driverless trains alongside conventionally driven ones on the same tracks. That’s the interoperability problem Sound Transit would face, solved.

  • 1901 — Boston experiments with automated trains
  • 1962 — Tokyo begins limited automated operation
  • 1966 — Paris upgrades its first main line
  • 1967 — London’s Victoria line runs fully automated
  • 1985 — Vancouver SkyTrain opens as a designed-from-scratch automated light metro
  • 2008 — Nürnberg mixes driverless and driven trains

The designed-from-scratch systems — SkyTrain in 1985, London’s Docklands Light Railway in 1987, Copenhagen’s metro — show the compounding benefits. No driver cabs means shorter trains. Shorter trains mean smaller stations. Smaller stations mean higher frequency. It’s an architectural cascade, not a bolt-on.

The vendor stack behind driverless metros

Today’s market has consolidated around a few signaling platforms. Hitachi, which automated the Disney and Linimo monorails, grew out of Ansaldo’s ATC-first architecture. Incumbent suppliers Alstom (Urbalis Fluence) and Siemens (Trainguard MT) retrofitted ATC support onto their existing product lines — the classic platform-versus-bet-the-company dynamic, played out in signaling cabinets instead of cloud consoles.

Siemens is pushing further with its AutomatedTrain program, targeting regular heavy rail with onboard video detection of track obstacles. That’s the one place where the robotaxi comparison becomes real: once a train shares right of way with cars and pedestrians, perception matters again.

The economics, per the source material, are Alstom’s own claims: up to 20% less trackside equipment and up to 30% lower energy consumption, thanks to more precise acceleration and braking. Treat vendor figures with appropriate skepticism — but the physics of optimized traction profiles is sound.

What Everett-to-downtown automation would buy Seattle

The concrete prize for Link is headway. Automated control can push trains as close as 60 seconds apart, with faster end-of-line turnarounds that reduce the need for pocket tracks — expensive, space-hungry crossovers in a tunnel. For a region planning extensions toward Ballard, that changes the capacity math without laying a foot of new track.

The trade-off Seattle Transit Blog flags: Ballard may work better as a separate, automated line than as another interlined branch. Interlining maximizes one-seat rides; separation maximizes frequency and lets the automation architecture work as designed.

Sound Transit, formally the Central Puget Sound Regional Transit Authority, was created in 1993 by King, Pierce and Snohomish counties to build a regional rapid transit system. Voters approved the “Sound Move” plan in November 1996 after an unsuccessful proposal in 1995, and the agency’s first light rail line opened in Tacoma in August 2003, with Seattle service following in 2009.

Funded by three major ballot measures — most recently Sound Transit 3 in 2016 — the agency plans to expand the light rail network to 116 miles and 83 stations, with construction scheduled to continue until 2041.

Sound Transit’s own 2014 evaluation considered automated light metro and chose light rail for the spine. The northern Link line’s full grade separation reopens that question — and the blog’s closing ask is simple: that Sound Transit seriously consider automation where feasible.

Photo of author

Technology Desk

Technology Desk at Archyde.

Phil Parkinson Pleased Zak Vyner Earned First Kenya Cap