Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Bayesian network approach to building an affective module for a driver behavioural model

Dorota Młynarczyk, Gabriel Calvo, Francisco Palmi-Perales, Carmen Armero, Virgilio Gómez-Rubio, Ana de la Torre-García, Ricardo Bayona Salvador

TL;DR

This work addresses estimating driver affective states, specifically mental load $Y_{ML}$ and active fatigue $Y_{AF}$, under uncertainty by leveraging Bayesian networks within the BERTHA driver behavioural framework. The approach builds a DAG-structured probabilistic model that links affective states to physiological signals ($Y_{MHR}$, $Y_{SDD}$, $Y_{RLH}$, $Y_{MNB}$) and demographic covariates, with $Y_{ML}$ and $Y_{AF}$ modeled via Bayesian logistic regression and the continuous nodes via Bayesian regression; inference is performed by MCMC using JAGS to obtain the joint posterior predictive distribution. The study demonstrates the ability to estimate conditional probabilities of mental states given physiology and demographics, revealing that age may have a protective effect while gender shows limited impact, with a concrete example yielding $P(Y_{ML}=1) \approx 0.72$ for a specific case. The findings highlight potential real-time safety applications, such as adaptive interventions, and point to the need for larger datasets and additional markers to enhance predictive accuracy and generalizability in dynamic driving environments.

Abstract

This paper focuses on the affective component of a driver behavioural model (DBM). This component specifically models some drivers' mental states such as mental load and active fatigue, which may affect driving performance. We have used Bayesian networks (BNs) to explore the dependencies between various relevant random variables and assess the probability that a driver is in a particular mental state based on their physiological and demographic conditions. Through this approach, our goal is to improve our understanding of driver behaviour in dynamic environments, with potential applications in traffic safety and autonomous vehicle technologies.

Bayesian network approach to building an affective module for a driver behavioural model

TL;DR

This work addresses estimating driver affective states, specifically mental load and active fatigue , under uncertainty by leveraging Bayesian networks within the BERTHA driver behavioural framework. The approach builds a DAG-structured probabilistic model that links affective states to physiological signals (, , , ) and demographic covariates, with and modeled via Bayesian logistic regression and the continuous nodes via Bayesian regression; inference is performed by MCMC using JAGS to obtain the joint posterior predictive distribution. The study demonstrates the ability to estimate conditional probabilities of mental states given physiology and demographics, revealing that age may have a protective effect while gender shows limited impact, with a concrete example yielding for a specific case. The findings highlight potential real-time safety applications, such as adaptive interventions, and point to the need for larger datasets and additional markers to enhance predictive accuracy and generalizability in dynamic driving environments.

Abstract

This paper focuses on the affective component of a driver behavioural model (DBM). This component specifically models some drivers' mental states such as mental load and active fatigue, which may affect driving performance. We have used Bayesian networks (BNs) to explore the dependencies between various relevant random variables and assess the probability that a driver is in a particular mental state based on their physiological and demographic conditions. Through this approach, our goal is to improve our understanding of driver behaviour in dynamic environments, with potential applications in traffic safety and autonomous vehicle technologies.
Paper Structure (7 sections, 5 equations, 3 figures)

This paper contains 7 sections, 5 equations, 3 figures.

Table of Contents

  1. Introduction
  2. Methods
  3. Bayesian networks and Bayesian inference
  4. Data structure and variables
  5. A Bayesian network for assessing active fatigue and mental load in drivers
  6. Results
  7. Discussion

Figures (3)

  • Figure 1: Basic direct acyclic graph example.
  • Figure 2: Bayesian Network for examining active fatigue and mental load. For clearer visualization, the nodes have been identified exclusively by the acronyms of their subindices.
  • Figure 3: Average posterior probability of each joint state of active fatigue (AF) and mental workload (ML) as a function of driver gender and age, with BMI held constant at 22, shown across $Y_{SRT}$ and $Y_{MNB}$ values.