It is often assumed that automation makes the operator’s job easier, safer or less critical. Instead, it redefines what qualifies as a "safe" machine and shifts the operator’s role from active control to passive supervision. This change introduces new, often underestimated safety challenges.
Marea de Koning puts it straightly: "It's clear that autonomous machines can do a lot, but as soon as humans get involved, things get messed up. The real danger isn’t that machines will outthink us, but that we’ll overestimate their intelligence – ceding control before they've earned it, and stepping back long before we understand the risks."
In her research, De Koning examines the challenges that arise in meeting safety compliance as automation levels increase. She also analyses the implications of the forthcoming EU Machinery Regulation 2023/1230, which adapts essential health and safety requirements for the EU machinery market in response to automation. Machines classified as ‘high-risk’ – those capable of executing behaviours without direct human input – are expected to perform safety functions independently and maintain a safe state at all times.
To address these requirements, De Koning proposes that such machines be equipped with a safety reasoning module. In contrast to conventional safety mechanisms that trigger fixed, immediate responses to hazards, a reasoning module enables machines to evaluate the situation and coordinate safety measures more flexibly.
”Safety functions may still execute the responses, but the goal of this concept of a reasoning layer introduces the critical step of deciding what’s needed and when,” she says.
She also investigates the feasibility of embedding such reasoning into off-road mobile machinery. Her research reveals that – at least for now – human oversight remains a necessary element in ensuring system safety.
“However, my findings also emphasize a major limitation that remains unsolved: humans struggle to effectively monitor complex automated systems for long periods. This can lead to reduced alertness, slower reactions, and even psychological fatigue, future work should focus more on this direction,” she adds.
In conclusion, Marea De Koning argues that when full removal of the human operator is not viable, achieving adequate risk reduction requires a careful balancing act between hardware, software and human factors.
Originally from the Netherlands, Marea De Koning now lives in Tampere, Finland, and works as a Functional Safety Expert at HULD Oy.
Public Defence on Friday 13 June
The doctoral dissertation of MSc (Tech) Marea de Koning, titled Safety in Highly Automated Off-Road Mobile Machinery, will be publicly examined at the Faculty of Engineering and Natural Sciences at Tampere University. The public defence takes place on Friday 13 June at 12 o’clock at Hervanta campus, Tietotalo building, in the Auditorium TB109 (Korkeakoulunkatu 1, Tampere).
The Opponent is Associate Professor Fredrik Asplund from KTH Royal Institute of Technology, Sweden and Co-opponent is Dr Jeremy Hubbard from System Safety & Sustainability Ltd, UK. The Custos is Associate Professor Tatiana Minav from Tampere University.