Abstract
Cooperative driving using connectivity services has been a promising avenuefor autonomous vehicles, with the low latency and further reliability supportprovided by 5th Generation Mobile Network (5G). In this paper, we present anapplication for lane merge coordination based on a centralised system, forconnected cars. This application delivers trajectory recommendations to theconnected vehicles on the road. The application comprises of a TrafficOrchestrator as the main component. We apply machine learning and data analysisto predict whether a connected vehicle can successfully complete thecooperative manoeuvre of a lane merge. Furthermore, the acceleration andheading parameters that are necessary for the completion of a safe merge areelaborated. The results demonstrate the performance of several existingalgorithms and how their main parameters were selected to avoid over-fitting.