The rise of autonomous control systems
Other sectors Maritime Oil and gas Power and renewables

Real autonomous control systems are only feasible with the availability of cheap sensors, the capacity to handle enormous amounts of data, and the processing capacity and methods to perform the necessary decision algorithms.

Interview with Asgeir Sørensen, Professor & Director AMOS, NTNU

Autonomous control means satisfactory performance under significant uncertainties in the environment and the ability to compensate for system failures without external intervention.1 This is different from automation, which is often defined as a process or procedure performed with minimal human assistance.2 The automotive industry has defined a scheme for levels of autonomy, from level 0 (no autonomy) to level 5 (full autonomy).3  Levels 1 and 2 denote increasing levels of the system assisting the operator, while levels 3 to 5 denote decreasing levels of monitoring required by operator. This level system is useful as it is important to be aware of the level of autonomy, design systems accordingly, ensure that safety is handled properly, and that user interfaces are working according to need.

Autonomy and automation go hand-in-hand, boosting one another and providing a fallback for one another. For example, think of a remotely-operated ferry, piloted from shore (this is automation as opposed to autonomy). If communication with shore is lost for some reason, then an autonomous system that will bring the ferry to a safe state is a needed backup.

Real autonomous control systems are only feasible with the availability of cheap sensors, the capacity to handle enormous amounts of data, and the processing capacity and methods to perform the necessary decision algorithms.

There are substantial risks that accompany “mixed” autonomous control systems. For example, autonomous elements of driving systems may be presented as safety assistance, but drivers that are aware of the presence of the safety assistance systems may change their driving, trusting more fully in those safety elements than they should. Without full autonomy, this is actually a danger; the overall risk picture for a driver on the road may increase. When full autonomy (level 5) is reached, this problem should be solved.

What lies ahead?

Though autonomous systems are still costly, their cost is coming down with the availability of cheaper sensors, computer storage, and processing power. A move from automation to advanced automation and autonomy is therefore expected within the next 10 years. Autonomy will not be achieved as a big leap, but in small steps through the 5 levels of autonomy.

Advanced automation systems will become widely available in the near future. These will be automated processes that include a much higher level of self-correction thanks to the addition of sensors and feedback loops. Fully autonomous systems will first emerge in niche areas, especially when safety is not an issue. Such systems are already becoming available in some areas, including mining applications and food delivery.

With respect to the wide application of full autonomy (level 5) for cars and ships, the situation is more unclear. While the number of car manufacturers performing tests with autonomous vehicle solutions is increasing by leaps and bounds, there are still major challenges to solve, especially in the field of safety and situational awareness in adverse environmental conditions. Wide public acceptance may also be challenging (In the early 1900s, even the introduction of the self-driving elevator took several decades.). Thus we expect that the introduction of autonomous cars and ships will take time, and come gradually.

Contributors

Main author: Siegfried Eisinger

Editor: Thomas Fries

  1. A. Panos et al. An Introduction to Autonomous Control Systems. IEEE Control Systems Magazine. 1991.
  2. M. Groover. Fundamentals of Modern Manufacturing: Materials, Processes, and Systems. Wiley. 2012.
  3. J3016. Taxonomy and Definitions for Terms Related to On-Road Motor Vehicle Automated Driving Systems. SAE. 2018.
Technology Outlook 2030 report cover
Download the Technology Outlook 2030 summary Click here