EdgeDrive: Supporting Advanced Driver Assistance Systems using Mobile Edge Clouds Networks

In this paper, we present EdgeDrive, a networked edge cloud services framework which can support low-latency applications during mobility taking into account needs of the driver, nature of the required service and key network features. We implement head-mounted device (HMD) based Augmented Reality (AR) ADAS applications such as navigation, weather notification and annotation based assistance to drive the evaluation. These services are then coupled with the Mobile Edge Clouds (MECs) wherein the container based service migration is enabled based upon migration cost and required Quality of Experience (QoE) to support mobility. An emulator based evaluation is carried out on the ORBIT testbed using realistic San Francisco taxicab traces running over nine edge cloud nodes and AR HMD being used by drivers. The experiments show that the EdgeDrive can support low-latency ADAS applications with an average system latency less than 100 ms for the applications under consideration.