Skip to main content
eScholarship
Open Access Publications from the University of California

Modeling the Effect of Driver’s Eye Gaze Pattern Under Workload: GaussianMixture Approach

Creative Commons 'BY' version 4.0 license
Abstract

This paper puts forward a Gaussian Mixture Model (GMM) foreye gaze behavior under workload and applies it to the analy-sis of gaze distributions in an automotive context. Specifically,it extends earlier work on Information Constrained Control(ICC) (Hecht, Bar-Hillel, Telpaz, Tsimhoni, & Tishby, 2019)(Hecht, Telpaz, Kamhi, Bar-Hillel, & Tisbhy, 2019) (Hecht etal., 2015) (Hecht, Telpaz, Kamhi, Bar-Hillel, & Tishby, 2018)by generating an ICC GMM derivative. We suggest a mea-sure for workload estimation based on the Kullback Leiblerdivergence (Dkl ) between tested eye gaze distributions and areference workload-free distribution. This derivative assumesdiagonal Gaussians that are distant from each other. Underthese assumptions, we achieve an analytical measure that hassignificantly fewer parameters than discrete grid-like distribu-tions (Hecht, Bar-Hillel, et al., 2019). Testing our measureon eye gazing data collected during real world driving experi-ments in a highway environment confirms the effectiveness ofthis approach.

Main Content
For improved accessibility of PDF content, download the file to your device.
Current View