Impact of Traffic Characteristic at Unsignalized Intersection in Mixed Traffic Condition using Logistic Regression Method (LRM) and Artificial Neural Network (ANN)

Abstract

Fajaruddin Mustakim, Azlan Abdul Aziz, Lim Heng Siong, Yaser Bakhuraisa and Noor Ziela Binti Abd Rahman

This research aims to investigate blackspot location at the chosen unsignalized intersection site (S) and the progression of right-turn motor vehicles (RMV) using the logistic regression method (LRM) and artificial neural network (ANN). In the initial stages, eleven unsignalized intersection site (S) were selected as the blackspot location and the study concentrating on urban road network. The network recognizes as heterogeneous road that all vehicle category utilizing same network. Consequently, the traffic conflict (TC) is analysed with a focus type of conflict, TC frequency at different time and sites. An LRM and ANN model were developed for right-turn motorists using datasets by combining three unsignalized intersection site (S2, S9 and S10). They are all vehicle (AV) model which consist (841 datasets), Passenger Car Model involved (357 dataset) and Motorcycle Model content (399 dataset). Using this approach, the identification of the variables that influence the decision-making process of right- turn motor vehicles (RMV) was done. Furthermore, the measurement of critical for vehicle category were implementing LRM approach. Among the sixteen variables examined in this statistical model, we found that vehicle gap, channelization (Chlzation), second vehicle passing RMV is motorcycle (SMc), second vehicle passing RMV is motorcycle (SCar), angular conflict (AGc), rear-end conflict (REc), traffic volume (TV) and RMV is rider were significant. This study purpose by implementing intelligent vehicle equipped with internet of thing (IoV), vehicle to vehicle communication (V2V) and advanced driver assistance system (ADAS) might solve partially traffic conflict and as well as traffic accident.

PDF