Speaker:Chen-Yeou Yu

A Situation Enabled Framework for 5G Vehicular Edge Computing --- Optimizations on Energy Consumption, Navigation and Service Response

Vehicular edge computing (VEC) is a promising technology to deliver quality of services (QoS) for accessing public cloud from vehicles. In-vehicle computers, especially for battery electric vehicles (BEVs) and hybrid electric vehicles (HEVs) in 5G vehicular edge computing environments, are typically energy and resource constrained for their limited electricity. Meanwhile, due to the high cost in building sustainable infrastructure to support vehicular computing networks, edge servers are often proposed to be deployed and they should operate with high energy-efficiency for better acceptability and applicability in the marketplace.

In this study, we propose a novel situation-enabled 5G vehicular edge computing framework consisting of the following components. The first component aims to enhance the energy-efficiency in vehicle computer systems. It learns the driving situations of a vehicle owner and adaptively turns off unnecessary applications launched in the wrong situation.

As the second component, the vehicular edge servers measure their current workload and dynamically offload their workloads to available servers if they are in the busy status. Idle servers can redirect client requests to available servers to minimize the number of busy and idle ones. In this way, we can maximize the utility of available servers and minimize the energy consumption on the infrastructure layer.

The third component is a path exploration and determination algorithm, which uses deep reinforcement learning (DRL) to find a traveling route for a vehicle to deal with road emergencies. The algorithm is designed to minimize the traveling time and distance, thus ensuring the QoS in terms of response time.

Implementations using long short-term memory (LSTM) on top of the situation model and simulation of the vehicular edge offloading algorithm have been conducted to evaluate the energy efficiency for the first and the second components, respectively. For the third part, we plan to simulate the path exploration for an ambulance. In order to evaluate the performance of the proposed path finding algorithm in the 5G vehicular edge computing environment.

Committee: Dr. Carl K. Chang(Major Professor), Dr. Johnny Wong, Dr. Ying Cai, Dr. Wensheng Zhang, Dr. Simanta Mitra

Join on WebEx: https://iastate.webex.com/iastate/j.php?MTID=m81d55131ebc7b894035518ca6f1380bf