Driver Cognitive Workload Classification Using Physiological Responses

Abstract

Motor vehicle crashes (MVCs) are a leading cause of death for law enforcement officers (LEOs) in the U.S. LEOs and more specifically novice LEOs (nLEOs) are susceptible to high cognitive workload while driving which can lead to fatal MVCs. To help address this issue, machine learning algorithms (MLAs) can be used for predicting the workload of nLEOs. These MLAs can then be implemented into adaptive in-vehicle technology that will better be able to manage the cognitive workload of LEOs in police operations. A naturalistic ride-along study was conducted with 24 novice nLEOs. Participants performed their normal patrol operations while their physiological responses such as heart rate variation (HRV) and percentage change in pupil size (PCPS) were recorded. After data collection was completed, anMLA was developed and trained based on these data using subjective responses collected from participants and pre-established thresholds for the features extracted from the physiological signals as the ground truth. It was found that the developed MLA could predict cognitive workload with relatively high accuracy given that it was entirely reliant on physiological signals. Future studies should implement the developed MLA into adaptive in-vehicle technology for the prediction of cognitive workload in real-time. Having this technology adapt to the cognitive workload of drivers should reduce cognitive workload of nLEOs and improve road safety.