Investigating User Acceptance of Autonomous Vehicles in Developing Cities Using Machine Learning: Lessons from Alexandria, Egypt

Sherif Shokry, Ahmed Mahmoud Darwish, Hazem Mohamed Darwish, Omar Elsnossy Ibrahim, Maged Zagow, Marwa Elbany, Usama Elrawy Shahdah

The willingness to adopt Autonomous Vehicles (AVs) represents a crucial advancement from the sustainable mobility perspective. This is progressively continuing in the developed countries. A comparable shift is expected in developing nations; however, empirical studies remain limited, especially in areas where AVs have not yet been deployed. This study investigates the willingness to adopt AVs in a developing city where AVs have not been deployed yet. A comprehensive travel behavior questionnaire was conducted among local commuters in Alexandria, Egypt, to identify the influential variables affecting AV choice. The well-known machine learning classifier, Extreme Gradient Boosting (XGB), was employed to develop a forecasting model, which indicated a notable accuracy. The results indicated that trip cost was the most influential feature. On the other hand, there is a considerable level of mode captivity, since most travelers prefer to remain with their current mode, regardless of the effects of other variables. A significant share of travelers expressed concerns about shifting to AVs due to safety worries associated with the travel behavior of other transportation modes’ commuters. The analysis provides nuanced perspectives on the variables promoting modal shift toward the AVs, supporting future policies for smart urban mobility.