Cognitive Load Estimation of Drivers in Cars

Though the Gaze-controlled HUD addressed the eyes-off-road distraction, it is still challenging to detect the distraction caused due to cognitive load of driver. I investigated distraction detection of drivers by estimating their cognitive load while operating secondary tasks. My studies analyzed different physiological parameters involving pupil dilation, rate of fixation, saccadic intrusion, head movement, and EEG (Electroencephalogram). Initially, a series of studies were conducted in laboratory involving standard psychometric tests like N-back and Arithmatic tests followed by trials involving driving simulator. I also undertook studies involving professional drivers performing different secondary tasks while driving actual car to validate performance of the proposed system. I used machine learning methods to train on data of multiple drivers to learn a global threshold as it was difficult to find a single threshold for all the drivers to classify cognitive states. An intelligent system generating visual, auditory, and haptic alerts was developed and integrated to the proposed distraction detection system. I also developed a cognitive load monitoring dashboard for comparing HMIs in automotive with real time graphical feedback on ocular parameters and cognitive load.

User Studies:

I undertook a study to investigate differences in values of ocular metrics for standard psychometric tests in laboratory. I collected data from 21 participants (16 Male and 5 female) with an average age of 26 years from our university.

I hypothesised that the occular metrics

are robust to ambient light variations
can be used to distinguish different levels of cognitive load with re- spect to change in task difficulty of visual and auditory tasks.

I undertook another study to investigate differences in values of ocular metrics for a standard lane changing task and performing secondary tasks while driving in simulated automotive environment in laboratory. In this study, my motive was to find whether these metrics can classify cognitive state of participants while driving and performing secondary task.

I proposed a Neural Network model to increase accuracy of the system by 75%. The ML models classify driver’s cognitive states into two states viz., ‘No Task’ and ‘Task’. The working of our proposed ML based cognitive load monitoring system is illustrated in the figure below.

Demo Video:

Summary:

Ocular metrics were able to distinguish between differences in cognitive states corresponding to driving with and without undertaking secondary task
A feed-forward Neural Network model outperformed individual metrics and other Machine Learning models with 75% accuracy (classifying between ‘No Task’ and ‘Task’)
Rooms for improvement in the expansion of dataset and inclusion of wide range of age groups
To investigate further levels of ‘Task’ like operating music, maps, calls, Air Conditioner and perceiving a road hazard

Publications:

Prabhakar, G., Mukhopadhyay, A., Murthy, L., Modiksha, M., Sachin, D., & Biswas, P. (2020). Cognitive load estimation using ocular parameters in automotive. Transportation Engineering, 2, 100008, ISSN 2666-691X. DOI: 10.1016/j.treng.2020.100008
Hebbar, A., Bhattacharya, K., Prabhakar, G., and Biswas, P., Correlation Between Physiological and Performance-Based Metrics To Estimate Pilot’s Cognitive Workload, Frontiers in Psychology 2021, 12 (2021), 954, ISSN 1664-1078. DOI: 10.3389/fpsyg.2021.555446
Biswas, P., & Prabhakar, G. (2018). Detecting drivers’ cognitive load from saccadic intrusion. Transportation research part F: traffic psychology and behaviour, 54, 63-78, ISSN 1369-8478. DOI: 10.1016/j.trf.2018.01.017
Babu, M. D., JeevithaShree, D. V., Prabhakar, G., Saluja, K. P. S., Pashilkar, A., & Biswas, P. (2019). Estimating Pilots’ Cognitive Load From Ocular Parameters Through Simulation and In-Flight Studies. Journal of Eye Movement Research, 12(3), 3, ISSN 1995-8692. DOI: 10.16910/jemr.12.3.3
Prabhakar, G., Madhu, N. & Biswas, P. (2018). Comparing Pupil Dilation, Head Movement, and EEG for Distraction Detection of Drivers, Proceedings of the 32nd British Human Computer Interaction Conference 2018 (British HCI 18) DOI: 10.14236/ewic/HCI2018.69

Patent:

Biswas, P., Deshmukh, S., Prabhakar, G., Modiksha, M., Sharma, V., K. and Ramakrishnan A., A System for Man- Machine Interaction in Vehicles, Indian Patent Application No.: 201941009219, PCT International Application No. PCT/IB2020/050253

Google Sites

Report abuse