Computational Modeling of Human Comfort in Automated Vehicles from Maneuvering Behaviors

Abstract

Autonomous vehicle (AV) is a promising solution to many problems in the nowadays transportation system. Massive efforts have been dedicated to improving the technical performance of AVs to push its prevalence among transportation systems. However, insufficient confidence in human comfort in AVs is a major concern that harms its user acceptance. Therefore, it is important to investigate human comfort in AVs. Despite recent efforts, there still lacks a computation model to capture the relationship between human comfort and related AV factors, especially the AV maneuvering behavior factors. Therefore, this paper proposed a new computational comfort model to quantify human comfort from in-situ AV behavioral factors in human-autonomous-vehicle interaction (HAVI). Ex-perimental studies were conducted to collect human comfort data in AVs where a set of AV journeys were implemented. Human comfort data collected during the studies were used to validate the comfort modeling approach. The results illustrated the effectiveness of the proposed human comfort model in capturing the relationship between human comfort and AV maneuvering behaviors.