Mobility Prediction for Efficient Resources Management in Vehicular Cloud Computing

Abstract

Vehicular Cloud Computing (VCC) has becomea significant research area recently, due to its potentialadvantages and applications, especially in the field ofIntelligent Transportation Systems (ITS). However, thehigh mobility of vehicular environment poses crucial challengesto resources allocation and management in VCC, which makesits implementation more complex than conventional clouds. Many works have been introduced to address various issuesand aspects of VCC, including resources management andVirtual Machine Migration in vehicular clouds. However, usingmobility prediction in VCC has not been studied previously. Inthis paper, we introduce a novel solution to reduce the effect ofresources mobility on the performance of vehicular cloud, usingan efficient resources management scheme based on vehiclesmobility prediction. This approach enables the vehicular cloudto take pre-planned procedures, based on the output of anArtificial Neural Network (ANN) mobility prediction model. The aim is to reduce the negative impact of sudden changes invehicles locations on vehicular cloud performance. A simulationscenario is introduced to compare between the performanceof our resources management scheme and other resourcesmanagement approaches introduced in the literature. Thesimulation environment is based on Nagel-Shreckenberg cellularautomata (CA) discrete model for traffic simulation. Simulationresults show that our proposed approach has leveraged theperformance of vehicular cloud effectively without overusingavailable vehicular cloud resources.