On Using Machine Learning for Vertical Handover Decision Making in a Vehicular Platoon

Abstract

Abstract Mobility management is required in the heterogeneous vehicular network because a variety of radio access technologies are integrated into these networks. Mobility handling importance rises when vertical handover is required in a vehicular platoon. This paper tackles the mobility problem for platoon vehicles in a heterogeneous vehicular network. More precisely, a vertical handover is initiated when the Quality of Service (QOS) level experiences degradation due to disturbances caused by moving vehicles. In this context, we propose a Vertical Handover (VHO) scheme achieved efficiently in two steps: the first one is the VHO decision based on various machine learn- ing techniques; the second step consists of a cooperative game based on network selection (CGNS) used to select the suitable technology. We investigated various machine learning techniques where our results indicate that the random forest algorithm is considered the most suitable approach for our scenario. Moreover, CGNS performance analysis demonstrates a reduced network overload and a decrease in useless handover requests.