Use this URL to cite or link to this record in EThOS: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.785252
Title: Prediction of drivers' performance in highly automated vehicles
Author: Alrefaie, Mohamed
ISNI:       0000 0004 7970 7948
Awarding Body: Loughborough University
Current Institution: Loughborough University
Date of Award: 2019
Availability of Full Text:
Access from EThOS:
Access from Institution:
Abstract:
Purpose: The aim of this research was to assess the predictability of driver's response to critical hazards during the transition from automated to manual driving in highly automated vehicles using their physiological data. Method: A driving simulator experiment was conducted to collect drivers' physiological data before, during and after the transition from automated to manual driving. A total of 33 participants between 20 and 30 years old were recruited. Participants went through a driving scenario under the influence of different non-driving related tasks. The repeated measures approach was used to assess the effect of repeatability on the driver's physiological data. Statistical and machine learning methods were used to assess the predictability of drivers' response quality based on their physiological data collected before responding to a critical hazard. Findings: - The results showed that the observed physiological data that was gathered before the transition formed strong indicators of the drivers' ability to respond successfully to a potential hazard after the transition. In addition, physiological behaviour was influenced by driver's secondary tasks engagement and correlated with the driver's subjective measures to the difficulty of the task. The study proposes new quality measures to assess the driver's response to critical hazards in highly automated driving. Machine learning results showed that response time is predictable using regression methods. In addition, the classification methods were able to classify drivers into low, medium and high-risk groups based on their quality measures values. Research Implications: Proposed models help increase the safety of automated driving systems by providing insights into the drivers' ability to respond to future critical hazards. More research is required to find the influence of age, drivers' experience of the automated vehicles and traffic density on the stability of the proposed models. Originality: The main contribution to knowledge of this study is the feasibility of predicting drivers' ability to respond to critical hazards using the physiological behavioural data collected before the transition from automated to manual driving. With the findings, automation systems could change the transition time based on the driver's physiological state to allow for the safest transition possible. In addition, it provides an insight into driver's readiness and therefore, allows the automated system to adopt the correct driving strategy and plan to enhance drivers experience and make the transition phase safer for everyone.
Supervisor: Not available Sponsor: Engineering and Physical Sciences Research Council
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.785252  DOI:
Keywords: Business and Management not elsewhere classified ; autonomous vehicles ; machine learning ; Human Factors ; physiological behaviour ; highly automated vehicles ; highly automated driving
Share: