Traffic accidents are often attributed to reduced attentional capacity, as vehicle operations involve complex tasks that depend on the driver's mental workload (MWL). Cognitive strain in MWL affects focus on other vehicles and pedestrians, which can result in serious accidents. Therefore, a method for effective cognitive-strain evaluation is required.
In this study, the driver cognitive strain was determined using the a posteriori probability method. The posteriori probability of the driver strain state is computed using multidimensional biological signals mapped into a probabilistic space based on a Gaussian mixture model. Ultimately, the driver’s cognitive strain was estimated using a regression function based on the posteriori probability. In the experiments, tasks tracked the preceding vehicle, and N-back evaluation helped elucidate the cognitive strain. Other experiments have shown an increase in the evaluation value during the satnav operation. Accordingly, the proposed method is considered suitable for real-time driver cognitive strain evaluation.