Using instrumented probe bicycles to develop bicycle safety and comfort prediction models

Common deterrents to cycling in North America are the real and/or perceived concerns on the safety, comfort, and practicality of choosing cycling over other modes of transportation, concerns that may be addressed by improved cycling facilities. The challenge lies in effectively quantifying the desirability of cycling facilities to assess return on investment for bicycle infrastructure decisions. In this paper an ordinal logit regression model is proposed as a potential Bicycle Comfort and Safety Prediction Model (BCSPM) to quantitatively predict a cyclistrs perceived safety and comfort. These BCSPMs were developed by conducting experiments utilizing an Instrumented Probe Bicycle (IPB). The IPB used in this study was developed using research from around the world. Many sensors were used, including: a 3DM-GX3 inertial sensor collecting time-stamped, position, velocity, and roll/yaw/pitch angles; and, a Microsoft Kinect sensor (still being operationalized) to record time-stamped eye/head positions, facial expressions, pulse, and ambient noise levels. Data for the BCSPM was collected from IPB sensors (numeric), field assessments (subjective numeric and categorical), and IPB rider questionnaires (categorical, Likert scales of comfort/safety). This paper outlines the potential applications of the BCSPM, the early modeling results of the study, the challenges faced, potential improvements that can be made to the IPB, and the next steps in this research.