Abstract
The increasing use of motorised mobility scooters (MMSs) has raised significant safety concerns, particularly related to user behaviour during operation. Although various advanced driving assistance systems have been incorporated into MMSs to identify potential environmental hazards, few studies have investigated the impact of user behaviour in MMS driving. This study was the first to incorporate automated behaviour monitoring into the evaluation of MMS driving using an add-on driving behaviour monitoring system (ADBMS). The ADBMS was a platform for contactless measurement that used a pre-trained convolutional neural network to estimate posture and two inertial measurement units to record steering and throttle operations. Experiments were conducted to demonstrate the usability of the ADBMS and evaluate the coordination of head movements and steering manoeuvres when driving an MMS in the outdoor environment. Cross-correlation analysis revealed that head movement consistently preceded steering operation during the driving tasks, indicating the potential of the proposed system to quantify user behaviour related to attention toward the surrounding environment. The lag time between these two parameters may serve as a novel index of driving safety. These findings could support a comprehensive understanding of users' driving behaviours and provide valuable insights into developing behavioural interventions to promote safer MMS use.