An enhanced computational integrated decision model for prime decision-making in driving

• Main Author: Rabi, Mustapha
• Format: Thesis
• Language: English; English; English
• Published: 2019
• Subjects: T Technology (General)
• Online Access: https://etd.uum.edu.my/9024/1/s94764_01.pdf; https://etd.uum.edu.my/9024/2/s94764_02.pdf; https://etd.uum.edu.my/9024/3/s94764_references.docx; https://etd.uum.edu.my/9024/

Recent development of technology has led to the invention of driver assistance systems that support driving and prevent accidents. These systems employ Recognition-Primed Decision (RPD) model that use driver prior experience to make prime decision during emergencies. However, the existing RPD model does not include necessary training factors. Although, there is existing integrated RPD-SA model known as Integrated Decision-making Model (IDM) that includes training factors from Situation Awareness (SA) model, the training factors were not detailed (IDM has only six training factors). Hence, the model could not provide reasoning capability. Therefore, this study enhanced the IDM by proposing Computational-Rabi’s Driver Training (C-RDT) model that improves the RPD component with 18 additional training factors obtained from cognitive theories. The designed model is realized by identifying factors for prime decision-making in driving domain, designing the conceptual model of the RDT and formalizing it using differential equation. The model is verified through simulation, mathematical and automated analyses and then validated by human experiment. Verification result shows positive equilibrium conditions of the model (stability) and confirms the structural and theoretical correctness of the model. Furthermore, the validation result shows that the inclusion of the 18 training factors in the RPD training component of the IDM can improve driver’s prime decision-making. This study demonstrated the ability of the enhanced C-RDT model to backtrack and provide reasoning on the undertaking decisions. Hence, the model can also serve as a guideline for software developers in developing driving assistance systems.