A Design of Decision Engine for 5G Auto-driving Based on Ontology Slice

With the access to 5G network, auto-driving vehicles can upload their sensors' information to Mobile Edge Computing (MEC) and get driving decision results relaying on the powerful computing ability of MEC. This article presents an architecture of intelligent decision engine running on MEC for 5G auto-driving (IDEA) by adopting the concept of ontology slice. IDEA constructs an ontoligy slice for every vehicle including sensor information, vehicle status, rule bases, etc. and provides logical isolated resources to prevent interference between ontologies. This paper gives out the method of implementing and managing of the ontology slices and method of scheduling reasoner. Besides, to make better and safer decisions, this paper also proposes an algorithm, combining an analytic strategy algorithm (WADD) and a heuristic strategy algorithm (TTB), considering both vehicle's velocity and sensors' refresh rate, to deal with incomplete information before making decesions. With simulation, it is shown that IDEA can provide efficient and reliable auto-driving decision service with low processing time.