Abstract
Distressed road surface conditions such as Potholes consitiute a major contributor to two-wheeler crashes and injuries in urban India; however, forward-facing hazard warning systems have rarely been empirically evaluated for single-track vehicles. This study investigates the behavioural impacts of a Driver Monitoring System (DMS)-based auditory pothole warning (50 m advance) on rider kinematics during naturalistic urban riding. Sixty-one riders (aged 18–46 years) completed a 12 km urban route on an instrumented two-wheeler, comprising of a 4 km baseline segment without alerts, followed by an 8 km segment with real-time auditory alerts preceding mapped potholes. High-resolution kinematic data collected using a Racelogic V-Box data (10 Hz) were analysed using unsupervised clustering to classify riding styles based on speed, acceleration, jerk, and yaw rate at multiple approach distances. Auditory warnings moderated 52.1% of initially aggressive riders, resulting in substantial reductions in speed and yaw rate while preserving robust jerk-based discrimination between riding styles across all distances. Conversely, 44.7% of initially cautious riders exhibited alert-induced behavioural escalation, characterized by aggressive steering responses, independent of hazard proximity. Markov and Hidden Markov Models identified jerk and yaw rate as dominant behavioural discriminants, revealing latent transition dynamics underlying riding style shifts. These findings indicate that non-adaptive auditory alerts can produce counterproductive behavioural shifts in cautious rider, highlighting potential safety trade-offs. Rider-style-adaptive warning systems that suppress cues for cautious profiles, while intensifying feedback for aggressive riders may therefore be necessary to achieve population-level risk reduction.

This publication has 24 references indexed in Scilit: