Title of talk: Data-driven modelling of cognitive workload using Virtual Reality and Eye Tracking Technology
Speaker: Dr. Souvik Das, Post-Doctoral Research Fellow, Purdue University
Date & Time: Wednesday, 10th July 2024, 10:00 am - 11:00 am
Venue: Seminar room IE 211, Second Floor, IEOR Building.
Abstract: Cognitive workload is associated with specific task which demands cognitive efforts for its completion. Maintaining a proper level of cognitive workload in a dynamic environment presents real challenges to different occupations. Although it is well established that both overloading and underloading can lead to severe consequences, a feasible technology has not been implemented to measure and monitor the operator's cognitive workload in a dynamic environment. This is partially due to a lack of consensus among the researchers on the construct of cognitive workload and its measurement. Eye tracking technology and methodology provide a promising avenue for clarifying how cognitive workload is conceptualized and measured. Compared to other measurement methods, the advantage of eye-tracking technology is that it can be implemented with minimal to no interference in real work environments. This research aims to develop a data-driven modelling approach for assessing cognitive workload using eye tracking metrics, which can help stakeholders make better decisions by gaining a deeper understanding of cognitive process and the factors that impact the level of cognitive effort.
The study addresses five research objectives: (i) development of a methodology to assess cognitive workload based on eye movement analysis, (ii) development of perceived cognitive workload metrics based on NASA-TLX with uncertainty and fuzziness modeling, (iii) development of information theory enabled eye metrics to assess the cognitive processing during accident/incident scenarios, (iv) development of cognitive workload prediction model considering the individual differences and time aggregated effect, (v) development of a methodology to assess operator training effectiveness considering both cognition and performance. To achieve the objectives mentioned above, a combination of design of experiments, multivariate statistical modeling, fuzzy set theory, and machine learning & explainable AI concepts are employed. These methodologies are utilized to ensure that the research is conducted in a comprehensive and thorough manner, and to provide accurate and reliable results.
From practical implications point of view, the research has designed a comprehensive methodology to assess the mental workload of the operators for an industrial operation. The research has adopted the multi-measure approach where all such measures like subjective, objective, and physiological measures are used together to understand the cognitive processing of the operators. The multi-measure approach assimilates different aspects of the cognitive functioning of human which provides a deeper understanding and helps in bringing reliable decisions. Further, the research has demonstrated the methodology which incorporates VR to design the research experiment and train the operators. This opens a new paradigm to the industry practitioners as it helps them to train their workers in a hazard free environment and make them more situationally aware of the abnormal scenarios. Furthermore, the research has devised a methodology to compute the training effectiveness considering both the task performance and cognitive task demand. Earlier, the training effectiveness is measured based on task performance only which may sometimes lead to wrong decisions. However, the inclusion of cognitive task demand may lead to a better decision making. In summary, the current study helps industry to identify the source of error and to improve performances in the workplace based on operators’ mental processing capability.
Bio-note: Dr. Souvik Das is a Postdoctoral Research Fellow in the School of Engineering Technology at Purdue University, USA with a focus in safety by design through Analytics. His research interests include Safety Engineering and Analytics, Risk Assessment, Virtual and Augmented Reality, Eye Movements Analysis, Human Factors and Ergonomics, Artificial Intelligence and Machine Learning, and Fuzzy Set Theory. Previously, Dr. Das worked as a Principal Research Scientist at the Centre for Excellence in Safety Engineering and Analytics (CoE-SEA), IIT Kharagpur, India. He earned his Ph.D. in Industrial and Systems Engineering from IIT Kharagpur, India, his master’s degree (M.Tech) in Industrial Engineering and Management from IIT Kharagpur, India, and his bachelor’s degree (B.Tech) in Electrical Engineering from Regional Computer Centre Institute of Information Technology, Govt of West Bengal, India.