Mikołaj Sokołowski Jul 23, 2024 - Interior sensing
Unveiling the Aspects of Cognitive Workload in Modern Automation
Technological advancements have simplified many aspects of our lives, but in certain areas, technology is becoming increasingly demanding. This creates an even greater need to study human-machine interaction. The concept of “Automation Irony” can help us understand these challenges, particularly the cognitive load we face when monitoring automated or autonomous systems. If you are interested in further exploration of its impact on our daily lives, your potential contributions to the study of cognitive workload are invaluable. Let us delve into the significance of this research.
Imagine two situations: driving your car on a calm Sunday afternoon along a familiar route to work – your mind operates almost automatically without needing to focus intensely. You might even start to feel bored due to the lack of mental stimulation and high level of familiarity, experiencing what’s known as mental underload. This state may be characterized by decreased alertness and slower reaction times, potentially increasing the risk of accidents.
Now, think about driving during Friday rush hour in a city you do not know, trying to find the right highway exit while using GPS and reacting to changing traffic conditions. You have to process and respond to many different pieces of information at once, which takes a lot more mental effort. In simple terms, your mind can be pretty overwhelmed. This state can be seen as an example of mental overload, which can lead to errors in judgment, missed information, and higher stress levels, all of which can compromise safety. Looking at these two scenarios, it is useful to ask a few questions:
How busy is the driver in each situation?
How complicated are the tasks the driver has to do?
Can the driver handle extra tasks on top of the current ones?
Can the driver respond to unexpected events?
Theoretically, you can answer these questions by measuring the Mental Workload (MWL) of the “system”—considering the person involved and the interactions in which they are situated. Before we get to that, it is good to know what MWL is and what we need to measure it.
MWL is considered a mental construct – it is a hidden variable that reflects the load a task puts on a person (Cain, 2007). This means we can’t see it directly; we can only infer it using selected indicators. Choosing these indicators is up to a theory or a framework, which acts like a research lens applied to the phenomenon so we can start studying it (Cain, 2007).
Despite early interest in MWL starting in the 70s and 80s, scientists still do not agree on its different levels, such as definition or sources (Huey and Wickens, 1993, p. 54). Fortunately, this does not mean they gave up trying to systematize this construct.
A big milestone was a NATO workshop in 1977 that brought together experts from various fields with the aim of addressing the following key points (see Moray, 1979):
(1) Constructing a strong definition of workload everyone could agree on.
(2) Creating a well-informed practical theory or framework for workload estimation.
(3) Identifying a relationship between measurement techniques and theoretical concepts.
These goals were met, resulting in the groundbreaking book Mental Workload: Its Theory and Measurement, edited by Neville Moray (1979).
This book led to even more productivity in the research community, paradoxically creating even more definitions, theories, and research frameworks than initially expected (for a comprehensive review, see Salvendy, 2012).
Now, let’s examine some of the most influential theories about MWL, which I personally also find very interesting.
Multiple Resource Theory (Wickens, 2002)
This theory helps us understand why some task combinations are harder to do than others. According to this theory, our brain has different “resources” for processing information. Wickens identified four main types of resources:
Stages of processing: Receiving information and thinking vs. reacting (e.g., listening and thinking vs. speaking).
Senses: Vision vs. hearing (e.g., reading vs. listening).
Types of information: Pictures vs. text (e.g., looking at a map vs. reading instructions).
Types of vision: Focusing on details vs. the big picture (e.g., reading a book vs. observing the surroundings).
The main point of this theory is that it is harder to do tasks simultaneously if they use the same resources (e.g., looking at the road while maneuvering and browsing social media using a smartphone—both of these activities involve visuo-manual modalities and can be categorized as the same “type of information”). In practice, this theory can be used to design safer interfaces or cockpits, minimizing the presence of tasks that would interfere with each other.
Contextual Action Theory (CAT; Stanton et al., 1995, as cited in Stanton & Young 1997)
This framework aims to understand how people manage tasks, balancing the demands placed on them with their available resources.
Imagine driving a car. The demands include monitoring traffic, controlling speed, and steering, while your resources are your skills, experience, and attention.
CAT explains that mental workload, or the mental effort needed to do a task, can change depending on this balance. If the demands exceed your resources, you feel overloaded, which may lead to mistakes and stress. If the demands are too low, you might feel underloaded, which, on the other hand, may lead to boredom and reduced attention, affecting your reaction speed to unexpected events.
By analyzing these interactions, CAT aims to optimize workload, making sure it is neither too high nor too low, reaching the “sweet spot.” This balance is crucial for maintaining performance and safety, especially in tasks requiring constant attention, like driving.
A short summary
The phenomenon of cognitive workload and its consequences is fascinating and has important practical implications for industries, interface development, and more. That is why it is being studied using different approaches, including behavioral, physiological, and task performance methods.
In our next blog post, we will explore the strengths and weaknesses of these three different means of MWL measurement, discussing how they enhance our understanding of cognitive load and what challenges remain. Join us as we delve into practical tools that help us measure cognitive workload effectively.
Author:
Mikołaj Sokołowski, Junior Researcher
🔴 Ready for more automotive insights? Download our handbook on Driver Stress Detection.
🔴 Interested in full-driver state detection services, including complex testing for drowsiness, distraction, stress, and driving under the influence? E-mail us at: humanfactors@robotec.ai
🔴 Curious about our latest projects? Follow us on LinkedIn to keep up with our news.
References
Cain, B. (2007). A review of the mental workload literature. DTIC Document.
Huey, B. M., & Wickens, C. D. (1993). Workload transition: Implications for individual and team performance.
Moray, N. (Ed.). (1979). Mental workload, theory and measurement. Plenum.
Salvendy, G. (Ed.). (2012). Handbook of human factors and ergonomics. John Wiley & Sons.
Wickens, C. D. (2002). Multiple resources and performance prediction. Theoretical Issues in Ergonomics Science, 3(2), 159-177.
Young, M. S., & Stanton, N. A. (1997). Automotive automation: Investigating the impact on drivers’ mental workload.