Research Initiation Award: Secure Spectrum Sharing for Intelligent Internet-of-Things (IoT) Wireless Communication Networks

点击翻译按钮获取中文摘要

Research Initiation Awards provide support for junior and mid-career faculty at Historically Black Colleges and Universities who are building new research programs or redirecting and rebuilding existing research programs. It is expected that the award helps to further the faculty member's research capability and effectiveness, improves research and teaching at the home institution, and involves undergraduate students in research experiences. The award to Morgan State University supports research on developing systems that defend Intelligent Internet of Things (IoT) devices – a project that supports one of “The 10 Big Ideas for Future NSF Investments,” i.e. “Work at the Human-Technology Frontier." The goal of this project is to secure the IoT users' communication by designing and developing schemes to share the spectrum with licensed users while circumventing the attackers, thereby maximizing the network capacity. First, vulnerabilities of IoT wireless communication networks will be investigated by designing optimal attacking strategies based on adversarial multi-armed bandits, and randomized time-varying feedback graphs. In the defense design paradigm, an intelligent IoT device will apply novel adversarial multi-armed bandit methods for effective frequency channel selection and data transmission while minimizing channel switching delay. In addition, both objectives of channel and power allocation will be integrated into a unified framework to optimize the IoT user's power consumption. Characterizing the wireless channels' inherent conditions with a time-varying stochastic process and considering both the attacker and the IoT user to be learning-based, the two agents form an online repeated stochastic game. It is intended to identify any possible equilibrium between these two agents using stochastic optimization methods. Furthermore, using a data-driven methodology based on statistical factor analysis approaches, the IoT device's robustness will be characterized as a function of its intelligence factors. Then, the proposed defense frameworks will be extended to centralized and decentralized multi-user IoT networks where combinatorial online learning approaches and various collision resolutions techniques will be utilized in the centralized and decentralized settings, respectively. Eventually, the collective intelligence of multiple IoT users will be quantitatively evaluated by extracting the group intelligence and examining its impact on the IoT users’ resiliency and network capacity in adversarial settings. Successful implementation of the proposed optimal spectrum access policy based on multidisciplinary techniques will provide secure spectrum sharing mechanisms for the IoT users with practical considerations which enables them to circumvent the adversary and maximize the network capacity.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.