As traffic levels in civil aviation continue to increase, the possibility of keeping up with capacity in Air Traffic Management (ATM) by simply applying “more of the same” is faltering. The antidote, automation, is frequently cited as yet another source of accidents. The Human Factors Research group at KTH, the Royal Institute of Technology in Stockholm, is one of the many actors that try to combat the drawbacks of badly implemented automation.
Whereas earlier increases in air traffic could be mitigated by redesigning the ATM sectors and increasing the number of air traffic controllers, this is no longer the case. Consequently, further work must be undertaken in the area of automation in order to cope with increases in traffic intensity.
From the viewpoint of a human operator, automation can decrease the continuously developing understanding of system behavior, or Situation Awareness. This in turn can create situations where automation together with other system components, e.g. humans, performs in a counter-productive manner.
The Human Factors Engineering group at KTH has performed research in the area of aviation during the past decades and currently works on modeling large, highly automated systems in order to develop indicators for safety assessment. The main rationale is to be able to maintain the trend of increasing system integration while diminishing safety risks and their associated costs, ultimately progressing sustainable aviation.
Whereas small decoupled systems are fairly easy to model, large socio-technical systems are very hard since the number of tightly coupled interactions between components in the system is large as well as being non-linear in their nature. This characteristic not only makes the systems hard to analyze but the consequence of non-nominal operations tends to snowball and cause large costs.
The main driver of creating larger integrated systems is basically to gain economic advantage, but if larger systems create sufficiently large consequences, this argument becomes obsolete. Currently, we do not possess sufficiently good tools and methodologies to assess large systems in near real-time in order to prevent large break-downs. Attempts have been done to create incident reporting systems in order to elicit data, some of them very good, but still there is a haphazard application of Human Factors in many domains.
As a response to this, the European Commission funded the HILAS project (Human Integration into the Lifecycle of Aviation Systems) in which KTH was a partner. The project ventured on a system-wide integration in the aviation sector, from flight-deck technology through operations to maintenance. One of the major achievements in the HILAS project was the inter-company sharing of potentially competitive information to obtain industry-wide benefits.
Do developers have the appropriate tools to understand the operational context of automated systems?
In order to overcome at least some of the most cost inefficient causes, Europe has since the 1990’s taken steps toward a more harmonized ATM structure. The latest such step is the SESAR program, which is often compared to the NextGen initiative in the U.S. Responding to emerging trends of higher automation and complexity, SESAR supports the Higher Automation Levels In ATM (www.hala-sesar.net) and Complex World (complexworld.innaxis.org) research networks, of which KTH is taking an active part.
Does criticism of automation imply that we should avoid automation altogether? No, of course we shouldn’t. But we should retain an awareness of the potential consequences in order to make informed decisions. Automation may to some extent remove humans from the “sharp end” of operations, but the implementation of automation tends to emphasize human intervention in the development phase. That redistribution would arguably make a rationale for increased research regarding the effects of decision making in early stages of development. Do developers have appropriate tools to understand the operational context of the automated system, especially in event-driven operations where some scenarios can only be assessed post-hoc?
In the evolving highly automated ATM system comprising a broader diversity of aircraft, broader diversity of equipage in avionics and communications, broader diversity of agents (human or autonomous), the main question posed by the KTH Human Factors Research group is how to model the impact of automation in order to increase the safety and cost efficiency of aviation.
Fredrik Barchéus, MSc Ph.D, is a member of the Human Factors Research group of KTH, the Royal Institute of Technology in Stockholm. He can be reached at firstname.lastname@example.org