May 28, 2024

Automatic Train Operation (ATO): Driving the Future of Rail

Automation is revolutionizing operations across various industries by enhancing safety, efficiency, and capacity - and the rail industry is no exception. This article serves as your guide to understanding how automation is applied in rail through Automatic Train Operation (ATO). We’ll explore its mechanics, advantages, and transformative potential for train travel, as well as highlight how OTIV is advancing automation in rail.

What is Automatic Train Operation (ATO)?

Automatic train operation (ATO) is a system that enables trains to be operated and driven automatically. It consist of multiple different subsystems which control different parts of the train. ATO works together with Automatic Train Protection (ATP), which is a safety system that checks speed limits, signalling input and can emergency brake, and Automatic Train Supervision (ATS), which allows for remote supervision of the train. These three systems together are called Automatic Train Control (ATC)

Automatic Train Operation is meant to increase safety, improve operational parameters such as efficiency, capacity, energy usage and provides an answer to increasing labour shortages within the rail industry. The degree by which a train is automated, is determined by its Grade of Automation (GoA).

Grades of Automation (GoA)

Implementing automation in any system typically involves a step-wise approach, as various  technologies, systems and even laws need to be created before achieving full autonomy. In industrial automation, machinery, robotics and cars, this progression is categorized by levels of automation. Similarly, in rail transportation, this categorization is known as Grades of Automation.

The Grades of Automation (GoA) start from GoA 0 or Line of Sight Operation, to GoA 4, which is Unattended Operation or Full Autonomous Driving. The Grade of Automation is determined by a list of basic functions which need to be automated before that system can be awarded a certain Grade.

💡 Difference freight versus passenger transport

There is an important distinction to be made when it comes to automating freight trains and passenger trains. When transporting freight you don’t have to open and close doors and be mindful of passengers. This is why in freight there is no distinction between Grade of Automation 3 and 4. Instead for freight we make a distinction between GoA2 with the driver on the train, and GoA2+, when the train is remotely controlled.

Grade of Automation overview by OTIV
Grade of Automation overview by OTIV

Grade of Automation 0: Line of sight

When trains are operated by drivers who use line of sight, and which don’t interfere with rail systems, we call it Grade of Automation 0. Everything has to be done manually, zero automation is applied. This means drivers have to manually check lights, switches, possible obstructions, oncoming traffic. They have rules about communicating what they see through walkie-talkie or some other form of communication which connects them to a dispatch service. This type of operations is most often  found on private property such as shunting yards, industrial sites and/or depots. The reason for this is because access is limited and everyone who does have access, has been trained or has to go through safety certification to know how to deal with these trains. Accidents can still happen ofcourse, that is why we developed our OTIV.ONE camera system for shunting, so we can provide more visibility to locomotive drivers and increase safety on industrial sites.

A lot of trams are also driven on line of sight and have to rely on manually reading signalling systems as a replacement to ATP. That is why we have developed our OTIV.TWO Advanced Driver Assistance Systems (ADAS) to intelligently detect obstructions and provide various warnings to drivers or intervene in braking when necessary.

Grade of Automation 1: Non-automatic

The first level of automation means the train is still operated manually, where a driver is responsible for every task in managing the trains operations. The only exception to this is the fact that on mainline the train uses Automatic Train Protection (ATP). This system monitors the speed of the train and has access to information about speed limits and signalling. This way it ensures the speed of the train doesn’t exceed local limits and can automatically stop at certain signals. It can even enable emergency braking when deemed necessary. As the name states, this is an emergency system to protect the train and its content as much as possible. To this day, a lot of tram and mainline rail routes are still GoA1.

Grade of Automation 2: Semi-automatic

Grade of Automation 2 (GoA2) is where active automation starts coming into effect. In this grade the system will be responsible for the following processes:

  • Accelerating and stopping the train
  • Supervising speed and braking when exceeding speed limits = ATP
  • Determining its position on the track
  • Transmitting data from the vehicle to a wayside monitoring system and requesting authorization for the route

Although this system is already automating quite some tasks, a driver is still present to perform several vital tasks as well as take over when the system should fail:

  • Stopping the train if there are obstructions or if a person has fallen on the track
  • Be mindful of crews working alongside tracks and adjust speed accordingly
  • Open and close train doors
  • Authorize departure from the station

Different sensors are now used to monitor everything in front of the train and near trackside. This is often done with a combination of camera’s, lidars and/or radars. Most of the metro lines in Europe are at least using GoA2. A complete overview can be found here.

In freight we talk about GoA2+ when a train is capable of automating the functions mentioned above, but is remotely operated. Seeing as no interactions with passengers or stops are required, the main functionality is getting a train ready for operation and driving it from point A to B. However advanced perception is required to be able to detect problems which otherwise a driver would be able to detect. OTIV.THREE, our remote supervision and control center, allows operators to remotely control trains and trams.

Grade of Automation 3: Driverless

In Grade of Automation Level 3, the train is in control between platforms. There’s no driver, but there are attendants on the train. As mentioned before this only applies to passenger trains.

  • All core operations of GoA 2 are covered
  • System monitors all platform elements
  • System stops the train if there are obstructions which are partially or fully on track
  • System protects crews at a track level

Attendants on board are assigned to take care of the following tasks:

  • Open and close doors
  • Authorize departure from the station

Examples of GoA3 in Europe are the Budapest Metro Line 4 or the Istanbul Metro UU line.  A complete overview of GoA3 and 4 can be found here.

Grade of Automation 4: Unattended or fully autonomous

In Grade of Automation 4, the train or tram is in full control of all systems on track as well as in the station. This is what we would describe as full autonomy.

  • All operations of GoA 3 are covered
  • System opens and closes doors
  • System determines if it is safe to depart from a station
  • Systems performs self-tests
  • System detects emergency situations
  • System can take the train in and out of the depot

Paris metro Line 4 is the latest track to support GoA4
Paris metro line 4 is the latest track to support GoA4

Levels of complexity to become autonomous

As you can see with the examples listed above, there is quite some difference in applying automation towards different rail vehicles. This has to do with the environment they operate in:

  • Closed environment: This is where most metros and some mainline trains and trams operate. No one can enter these tracks because of safety doors and they are closed off from all other forms of mobility. This means there are very little variables to account for, which makes enabling autonomous driving a lot easier.
  • Open environment: This is where most trams and mainline trains operate. Cars, pedestrians, motorcyclists and every type of mobility option has (partial) access to the same roads the tracks go on for urban transport. In mainline rail there are level crossings and signalling to take into account, as well as other variables such as environment (a tree falls on the tracks) or the weather (rain changes the braking distance).

As you can see, open environments require more advanced systems to go towards full autonomy. It needs better perception from its sensors, more intelligent decision making because of the wide range of situations it has to deal with and it needs to be able to communicate with a lot of different systems like ETCS or ETRMS, other trains and trams, control centers and more.

The benefits of automation in rail

Technological advancements such as automation (or more recently machine learning and A.I.) are often accompanied with skepticism. Any time we relinquish control of systems which can have an impact on our safety, we need to be careful in implementation. However we can’t understate the benefits these technologies can bring as well as the hazards human operation still carry. For one technology is not susceptible to human errors such as fatigue or other health issues impacting performance. Let’s dive into the primary advantages automation can bring to rail.

Increase safety

Basic levels of automation in rail start with increasing safety and implementing emergency protection systems. When you are a driver, no matter how skilled you are, you only have two eyes to work with. Human eyes are amazing at detecting obstacles and dangers, but they have some limitations.

  • Humans have a pretty wide total vision field of view, around 210° with both eyes combined. However of that range, 150° is peripheral vision. We will be able to see things there, but it is much harder for us to discern what it is and how it is moving. On top of that, suppose we focus on an object to one side whilst driving, we become oblivious to what’s happening on the other side. Sensors have the advantage that they are looking at everything within their Field of View equally. This makes the first step in automation, when we combine a human driver with technological support, so powerful. You have better detection capabilities, but retain the high resolution focus and critical decision-making of a qualified driver.
  • A human is driving from inside the vehicle, which inherently creates blind spots. The larger the vehicle, the harder it becomes to have a comprehensive view of everything that happens around the vehicle. Everyone has been told as a child to keep their distance from buses, trams and trucks for this exact reason. Sensors on the other hand can be placed on the outside of the vehicle and in places to specifically account for these blind spots.
  • The human eye doesn’t work as well in different environmental conditions. When it’s raining heavily or snowing a lot, visibility becomes limited. If there is a lot of light or it hits you at a certain angle, it can impair you for a short time. In contrast we can choose sensors and adapt the calibration of our sensors to better detect during certain conditions.

When we developed OTIV.ONE, our camera system for shunting, it was exactly to solve some of the issues above. During pushing of convoys, the drivers can’t see in front of the tram and they have to rely on visual cues of their colleagues. Locomotives are also very large, steel cages, which makes it very hard to see below a certain height or close to the convoy.

OTIV.TWO, our ADAS for urban rail, provides a solution to the same challenges in complex urban environments. We provide better detection of the environment, combined with smart algorithms to warn a driver of incoming collisions or even help emergency brake. Here the combination of the complex decision-making capability of a competent driver, with the capabilities of our system, really shine in elevating safety. The driver can override the system when it might be overly careful and the system can warn the driver of the few things they might miss.

More capacity

When running rail vehicles on a network, only a certain amount of them can be on a given track at a time. Suppose something goes wrong, enough distance needs to be between two rail vehicles for it to come to an emergency stop without hitting the next rail vehicle, which is what Automatic Train Protection is for. If you want to run more rail vehicles, they need to be able to talk to eachother so you need to reserve less space between each one. The first grades of automation will reduce the distance between rail vehicles by a certain amount. However real maximalization of capacity will come when the systems are able to talk to eachother and can dynamically adjust the distance between them.

Optimize for energy usage, passenger experience and wear-and-tear

Automation can improve operational parameters because the route becomes a mathemathical problem to be optimized. When a systems knows all the different parameters of the track, we can calculate the optimal time and intensity with which it needs to accelerate and/ or decelerate to reach the next stop. This means we can optimize energy usage, as well as reduce wear-and-tear on the engine and brakes. Smoothing the curve of acceleration and deceleration also improves the passenger experience.

In conclusion

Automatic Train Operation (ATO) is changing the rail industry and OTIV is a part of pushing this movement forward. By offering different solutions for every Grade of Automation, we help improve safety, capacity and operational parameters for in urban, mainline and industrial rail. Are you interested in raising the Grade of Automation in your rail operations with Automatic Train Operation? Drop us a message and let's talk about how OTIV can help you realize your goals.