Effectiveness of smart LED strips at mid-block crosswalks under distracted driving conditions

Abstract

We investigated the effectiveness of an LED-based smart mid-block crosswalk system in mitigating the detrimental effects of driver engagement in non-driving-related tasks (NDRTs) with behavioural, performance, and subjective measurements. We designed a 2 (Crosswalk: smart vs conventional) by 2 (Task complexity: low vs. high NDRT) within-subjects experiment. Thirty-six drivers drove along four urban scenarios in a static driving simulator. We collected data on driving behaviour (speed, reaction distance), and safety (minimum time-to-collision [MTTC]), as well as subjective driver ratings on the perceived task load and their trust in the technology used, and performance levels achieved while performing the NDRTs.

Behavioural and performance observations showed that the smart mid-block crosswalk resulted in greater reaction distances and MTTC values when drivers interacted with pedestrians, thus indicating improved safety. Remarkably, the results also revealed that increased NDRT complexity does not negatively affect the smart crosswalk effectiveness in terms of driver-pedestrian collision prevention (i.e., MTTC does not decrease significantly). However, the NDRT complexity influenced driving performance in terms of speed and reaction distance at brake pedal pressure, with drivers exhibiting lower speeds and lower reaction distances with higher task loads. Moreover, the subjective ratings and performance levels while performing a NDRT reflected the experimental manipulation, with drivers perceiving higher task loads and performing worse in the higher NDRT complexity condition. Overall, the smart mid-block crosswalk led to a safer driver-pedestrian interaction compared to conventional crosswalks and achieved a good acceptance level both of which augur well for the widespread future installation of this technology.