WHILE dates for the launch of autonomous cars recede into the future and car manufacturers join forces to ramp up their already considerable investment in electric and autonomous vehicles, rail is quietly making important advances in the development of driverless and eventually autonomous trains.
As we report this month, Rio Tinto completed the transition to full automatic driverless operation of its entire heavy-haul rail system in the Pilbara region of Western Australia on June 14 making it the world’s first fully-automated mainline rail network.
While the AutoHaul project took longer and cost more than expected, it is a major step towards automating mainline railways. Full automation is already reaping benefits for Rio Tinto in terms of lower operating costs, shorter journey times, and greater reliability - vital for a company in a highly-competitive market. It is also improving the efficiency of scheduling, helping to eliminate bottlenecks on the network, and boosting productivity.
Hitachi STS, which developed AutoHaul in cooperation with Rio Tinto, is talking to North American railways about automating freight train operations. The challenge will be far greater in North America where multiple freight and passenger operators share the same tracks, trains are of different types and weights, and there are numerous junctions and yards as well as links to industrial plants.
Similar challenges apply in Europe. Nevertheless, progress is being made here as well. Following 18 months of research, French National Railways (SNCF) completed its first test run in July with a remotely-controlled locomotive- hauled autonomous train, as part of a project to develop driverless passenger and freight train prototypes by 2022.
Fully-automated metros have been around since 1981 when the first line opened in Kobe, Japan. According to the International Association of Public Transport (UITP), by the end of last year there were 64 fully-automated metro lines totalling 1026km in operation in 42 cities. However, the UITP forecasts a major acceleration in the development of automated metros with the total length expected to triple by 2023 to breach the 3000km mark. Fully-automatic metro lines are forecast to account for 48% of greenfield metro projects by 2022 in terms of kilometres compared with 10% in 2018.
“Every city that has built a new automatic metro line always does so again,” Mr Ramon Malla, chairman of the UITP’s Observatory of Automated Metros, and director of strategic projects with Barcelona Metropolitan Transport (TMB), told IRJ at the UITP’s Global Public Transport Summit in Stockholm in June. “It is difficult to see anyone building a new line that is not automatic.”
However, Malla says automation of existing lines is progressing slowly. “Converting a line to automatic operation is a real challenge, especially as we need to keep the service going. Cost is not the main issue; it is the complexity of managing the project and suppliers need to give operators more confidence.” While Barcelona’s newest lines (9 and 10) are fully automated, TMB is looking to convert Line 1, its oldest and busiest line, as the rolling stock will need to be replaced soon.
Malla listed the many benefits of automating a metro line, including greater capacity, availability, and operatonal flexibility; elimination of human error; reduced capital and operating costs and lower energy consumption. He also says the social element is very important. “We took enough time to negotiate with the unions to pave the way for automating the rest of the network. We are unique in not having any jobs with the title of train driver. Instead, we have flexible jobs which combine different tasks. Our staff now have more fulfilling jobs, they are more motivated, and we have less absenteeism than before. The happiness of our employees is important.”
Despite 38 years’ experience with automatic operation, improvements are still required, as Mr Paul-Edouard Basse, key account manager with Siemens, points out. He says failures on driverless lines are very rare but when they occur “they can be dramatic.” Basse says operators need to be able to restart operation quickly to avoid stranding trains in tunnels. He says cyber security is a must today, but was not a problem with early systems.
Artificial Intelligence (AI) is expected to improve the performance of automated metro lines by enabling predictive maintenance. Harnessing data retrieved from the CBTC and trains will help to reduce the criticality of incidents and assist operators to recover from incidents more quickly than today. As Malla says, a lot of time is lost in the control centre with visual and voice interactions during an incident. AI should help to reduce these interactions and allow traffic controllers to focus on the core problem.
Malla also points out that metros have been designing new lines and operating their networks the same way for decades. He says metros should take advantage of automation to introduce express services.
Some autonomous car technologies are also heading for rail. These include self-position detection, sensor fusion to allow environment recognition, multiple driving assistance systems, sensors to anticipate human behaviour, and platooning.
Such developments will help rail to drive automation forward and maintain its superiority in driverless operation over road transport. There is no going back.