Contactless Side Channels on Mobile Wireless Charging: Exploration and Mitigation
DescriptionThis proposal considers the distributed optimization problems under imperfect communication. The purpose of distributed optimization is to solving an optimization problem through decentralized methods without central coordination. The global objective function is the sum of all agents' local objective functions. The agent is able to communicate information with its neighboring agents and only be accessible to its local functions. All agents cooperatively minimize the global objective function through information communication. Distributed optimization has been widely applied in many research areas, such as sensor networks, distributed control, distributed resource locations, and decentralized estimation. The distributed mirror descent algorithm (MDA) has become a popular tool in distributed optimization in recent years. Instead of using Euclidean distance in the traditional sub-gradient descent algorithm, a non-Euclidean based on Bregman divergence as a distance measure function is utilized in MDA to solve the distributed optimization problems. Further, it has been shown that MDA is an efficient tool to achieve a fast convergence rate based on the Bregman divergence approach and deal with optimization problems over large-scale distributed networked systems. The information transmission between different agents is the key to solve distributed optimization problems. However, the communication is not perfect in real life due to various kinds of limitations, such as delay, limited data-rate, and noises. In recent years, different kinds of methods have been proposed for distributed optimization problems with many imperfect communication schemes. The existing approaches cannot be directly applied to MDA. Thus, it is a significant direction to investigate distributed MDA to deal with these practical issues. This project aims to develop different types of methods for distributed MDA over communication network with limited data-rate or communication delay. Based on the traditional concept, some new algorithms will be developed and its performance will be tested and verified under imperfect communication channels. The theoretical results generated in this proposal, if well accomplished, will lay a solid foundation for the development of distributed optimization with imperfect communication channels. The theoretical results generated in this proposal, if well accomplished, will lay a solid foundation for the development of distributed optimization with imperfect communication channels.
|Effective start/end date||1/10/23 → …|