Classification of Mental Workload (MWL) using Support Vector Machines (SVM) and Convolutional Neural Networks (CNN)

In the current era of technological advancements and rising human-machine interaction, urged the vital importance of human factors and ergonomics in an industrial collaborative environment. These ergonomic needs have made it essential to analyze the industrial cognitive processes like mental workload (MWL), stress and vigilance in the ecological environment. Conventionally Electroencephalography (EEG) was used for assessment of brain electrical activity but recently functional Near-Infrared Spectroscopy (fNIRS) has immerged as a better substitute for acquiring brain signals with fewer protocols and enhanced spatial resolution. Over the period of time Machine learning algorithms (LDA, k-NN, ANN) are used to classify MWL and affiliated brain functions. Now the trend of employing Deep learning techniques is gaining popularity. In this study, we analyzed and classified MWL states using Machine learning (SVM) and Deep learning (CNN) algorithms. The classification accuracies achieved with Deep learning (CNN) outperformed the accuracies achieved with Machine learning algorithms. The best accuracies were achieved using CNN that are in the range of 80-87%. Finally, a comparison is drawn between Machine learning and Deep learning algorithms for better classification and discrimination of cognitive loads.