Public transport is going autonomous

 IEC E-Tech

Public transport systems are gradually becoming autonomous, despite the many challenges, including for standards developers.

article here

Cities across the world are struggling with increased traffic, “resulting in billions of dollars in lost time, wasted fuel, and excess carbon emissions”, according to this provider of data and insight into worldwide transport systems. One of the solutions could well be autonomous public transport systems, which city planners can organize to reduce traffic with little requirement for added personnel.

Driverless urban rail links such as those in Kobe, Japan; Lille, France; and London Docklands have been operational since the 1980s, with dozens of metro networks in cities from Doha to Lahore and Santiago following suit. In recent years, autonomous trams and driverless bus services have been enabled thanks to advances in the ways computers process data from cameras, sensors and scanners in response to live conditions.

Driverless transport will soon involve air as well as surface and underground systems. Advanced air mobility (AAM) vehicles such as air taxis are not yet mainstream, but developments are happening quickly. For example, there are plans for air taxi services between Heathrow and central London.

The pros and cons of autonomous public transport

There are four major constraints which impact the viability of autonomous public transport systems. They include the amount of investment required, regulatory hurdles, infrastructure readiness and cultural acceptance.

“These constraints vary, depending on the economy in question,” explains Ripin Kalra, a smart cities specialist inside the IEC and Senior Fellow (Participatory Urban Resilience & Environmental Assessment and Climate Policy) at the University of Westminster. “In labour-intensive economies, job creation may be prioritized. In others, particularly those facing labour shortages, autonomous transport is a way of addressing workforce gaps, in addition to being an answer to traffic jams.”

Human fatigue is also a major safety risk. Urban transport drivers face a lot of stress and abuse. “Many report harassment from passengers,” says Kalra. “Autonomous systems can operate longer, more consistently and without fatigue-related errors. Fleet utilization increases significantly when you remove human constraints like working hours, illness or personal commitments.”

There’s also a shortage of trained drivers in some countries. Germany is forecast to have a shortfall of 80 000 bus drivers by 2030. Trains, trams or buses often operate up to 15 hours a day, seven days a week, but drivers cannot. “Labour is often the most expensive component of any system and removing or reducing labour costs can improve the economic viability of public transport systems, which rely on public funding to operate,” Kalra adds.

Infrastructure is the biggest challenge

That said, building the infrastructure required for such systems remains the most expensive and complex barrier. Automated transport depends on data centres, high-speed connectivity (such as 5G), artificial intelligence (AI) processing and constant connectivity – all of which consume energy and water. Even scaling electric vehicles (EVs) has strained power grids in many countries. Automated transport adds further demands.

“Scaling up remains the major challenge,” says Kalra. “Many cities still face a deficit in basic public transport infrastructure. Some are wealthy and able to invest; others struggle to fund basic services. Birmingham, for instance, has faced severe financial constraints and must prioritize essentials like waste collection and education before considering major innovation projects, such as automated transport.”

While Glasgow, home to the world’s third-oldest metro network, is rolling out driverless underground trains this year, London, which opened the world’s first subway in 1863, has abandoned similar plans, citing cost. The London city mayor said the cost to upgrade just three sections of London’s vast Tube network would be GBP 20 billion. Opposition from the drivers‘ trade union, over fear of job losses, was another concern.

Standards can help drive things forward, though, by making sure the tech is interoperable and works safely – reducing costs in the medium term. The IEC and ISO joint committee for information technology, JTC 1, established a subcommittee which publishes many of the foundational standards for the use of AI. SC 42 has published ISO/IEC TR 24030, a technical report which looks at many AI use cases, including transport in smart cities. It notably addresses how AI can be used for enhancing traffic management efficiency and infraction detection accuracy and for traffic signal optimization based on multi-source data fusion.

Public acceptance of automated systems is cultural and varies by country. In wealthier nations, regulatory scrutiny, liability concerns and public expectations around safety are much higher and that can slow deployment. “In contrast, some developing nations are more open to experimentation and ‘leapfrogging’ to the latest technology, partly as a statement of modernization,” Kalra describes.

Perceptions of automated transport can shift quickly. In San Francisco, for example, autonomous vehicles are now seen by the majority of passengers as significantly safer than human drivers sharing the road.

Trials with trams and trains

A section of Oslo is in the middle of a year-long test of a tram fitted with computer vision. Supported by EU-Rail, the tram is operated remotely using technologies developed under the EU-funded research initiative FP2-R2DATO. The EU Rail initiative has already helped launch Europe’s first passenger-carrying driverless train operating on a 24 km stretch of open railway in the Czech Republic.

Both developments are significant in shifting the rollout of autonomous train and tram systems away from insulated sections of track, such as airport shuttle services, onto networks with greater variables of potential incidents, such as level crossings.

The EU lags behind some other parts of the world when it comes to autonomous trains, notably China. In the country, where the Wuhu Rail Transit monorail has been operating at Grade of Automation 4 for five years, the state owned rolling stock manufacturer CRRC has debuted a claimed world first driverless train operating at speeds of 200 km/h. (For more on the various grades of automation based on the Society of Automative Engineers (SAE), read Capturing data for autonomous vehicles | IEC e-tech).

Standards can help

Depots, where trains are stored, maintained and refuelled, are relatively straightforward to automate. Branch lines, which cover about 30% of total track mileage in Europe, have lower maximum speeds, more homogeneous traffic (than main lines) and simpler regulatory requirements, making automation more manageable.

“There is still a ‘not invented here’ mindset in parts of the industry, but there is also growing recognition that modern technology can significantly improve how things are done,” explains Alex Haag, CEO and Co-Founder of an autonomous train systems developer based in Strasbourg and Munich. There are parallels with the automotive industry, which was widely seen as an old, closed industry where startups couldn’t innovate, before perception changed with the launch of Tesla. Rail is at a similar inflection point.

“Regulation doesn’t prevent innovation, but it does require that things be done carefully and systematically,” says Haag. “At some point, innovation has to move from theory into the real world, which means dealing with hardware, safety certification and regulatory processes.”

That is also where standards come in. IEC TC 9 is the IEC technical committee set up to develop standards pertaining to electric equipment for railways. Different levels of automation are specified in IEC 62290-1. This TC 9 standard establishes fundamental concepts for urban transport management and control systems. Grade of Automation 4, for instance, applies to trains that run automatically at all times, including door closing, obstacle and emergency situation detection. TC 9 also publishes IEC 62267, which specifies the safety requirements for automated urban guided transport. “Many cities simply don’t know where to begin,” says Kalra. “Standards can provide clarity and reduce risk.”

Autonomous buses are on track

The first self-driving buses are now departing to and from Rotterdam The Hague Airport, shuttling passengers on public roads, albeit a short distance, to the Meijersplein metro station. It is one of a number of Society of Automotive Engineers (SAE) Level 4 proof of concepts now live. Others include a 4 km route between two Gothenburg stations; a 5,3 km route connecting a university campus in Michigan; a 2,5 km route with nine stops in Arbon, Switzerland; and a 7 km stretch of a district in Hannover with 13 bus stops, 10 traffic lights and “multiple complex manoeuvres”, including roundabouts, side and parallel parking, and pedestrian crossings.

Research suggests there are more autonomous bus trials on city streets than autonomous train trials. Flexibility and cost are the key reasons. “Once you invest in rail tracks and infrastructure, you’re locked into that route,” says Kalra. “Buses, by contrast, can be redeployed to match demand. If a bus fails, the system continues operating; if a train fails, it can disrupt the entire line.”

Buses also have lower capital costs and shorter build times compared to rail or light rail systems. With technology in this space evolving rapidly, it is easier to upgrade or replace a bus fleet than fixed rail infrastructure. “Rail works extremely well in high-demand corridors, but buses offer adaptability, resilience and lower investment risk,” Kalra says.

Problems of integration and interoperability

“Interoperability is crucial,” says Kalra. “If systems communicate with one another, waiting times can be reduced and capacity adjusted dynamically. Ideally, transport networks should respond intelligently to passenger flows.”

However, full integration increases cyber security risks. “The more connected systems are, the more vulnerable they become to hacking or malicious interference,” he says. “So integration must be balanced with strong security standards.”

There is, however, no overarching framework for autonomous transportation as a complete system. “Today, multimodal transport systems in cities are fragmented,” says Kishor Narang, Vice Chair of the IEC Systems Committee for Smart Cities. “There are many initiatives and good intentions, but very few cities have a truly harmonized, end-to-end transportation system. If that foundation isn’t in place, integrating autonomous vehicles becomes an even bigger challenge.”

The vision may be compelling, and the technology is progressing, but cities are still waiting for individual technologies and systems to mature before attempting large-scale integration into multimodal transport networks.

“Cities are systems of systems,” says Narang, who is also Principal Design Strategist and Architect at an India-based consultancy. “Technology alone can’t solve these challenges, and neither can standards or policy on their own. What’s required is the integration of multiple technologies, transport modes and ecosystems into a coherent whole – and that’s a very complex task.”