Control and prediction of traffic crashes in the residential streets in Iraq using the expert system

• Main Author: Mohammed, Ali Ahmed
• Format: Thesis
• Language: English; English; English
• Published: 2022
• Subjects: T Technology (General); T173.2-174.5 Technological change
• Online Access: http://eprints.uthm.edu.my/8388/1/24p%20ALI%20AHMED%20MOHAMMED.pdf; http://eprints.uthm.edu.my/8388/2/ALI%20AHMED%20MOHAMMED%20COPYRIGHT%20DECLARATION.pdf; http://eprints.uthm.edu.my/8388/3/ALI%20AHMED%20MOHAMMED%20WATERMARK.pdf; http://eprints.uthm.edu.my/8388/
• Institution: Universiti Tun Hussein Onn Malaysia

Residential streets suffer from various traffic safety problems related to traffic accidents, especially in low and middle-income countries. This research's objective was to develop a novel system for controlling and predicting traffic accidents in the residential streets using the expert system (CPTCRSI-ES). Knowledge of the CPTCRSI-ES was collected from domain experts and published sources. The system comprised nine modules: the module to identify types of car accidents–causes-preventive actions effects, module to identify safety problems and related solutions, module to find the speed calculation, module to find the cut-through traffic calculation, module to find the Equivalent Property Damage Only (EPDO), module to find residential density calculation, module to rank traffic safety parameters, module to control the traffic accidents, and module to find the traffic accidents prediction. Verification, validation, and evaluation process (V, V and E) was conducted. Results were statistically analyzed using Number Cruncher Statistical Systems (NCSS) version 26. Verification was done by asking three groups of experts involving 20 professional engineering computers and domain experts. The arithmetic mean for evaluators' responses was higher than 4.2 out of 5, indicating a strong agreement. The Cronbach's alpha was 0.960, and internal consistency reliability (ICR) showed excellent reliability. Results for verification demonstrated the satisfaction of the experts with the proposed system. In the validation process, the experts were requested to propose appropriate strategies and solutions to address the safety problems and reduce traffic accidents. The arithmetic mean for matching the experts' answers and the outputs of CPTCRSI-ES was higher than 4, and the Cronbach's alpha was 0.917 and ICR, indicating excellent reliability. Finally, evaluation of the system by its end-users showed that the overall assessment rating was more than 4, Cronbach's alpha was 0.932, and ICR showed excellent reliability. Results for V, V, and E demonstrated that the system had met its primary objectives. The proposed system adopted in this research can help traffic safety authorities control and predict traffic accidents.