CAREER: Protocols for Low-Power Wide-Area Networks in White Spaces

Iowa State University, Ames IA

Investigators

Abstract

This project will integrate research and education in the design and implementation of a Low-Power Wide-Area Network (LPWAN) in the TV White Spaces. LPWANs enable low-power devices to transmit over long distances for various Internet-of-Things (IoT) applications. Today, LPWANs have a number of major limitations making their adoption challenging. First, they rely on wired infrastructure for integrating multiple networks to cover very large areas (e.g., smart city). Lack of proper infrastructure hinders their applicability to rural/remote wide-area applications such as agricultural (e.g., smart farming) and industrial IoT (e.g., oil-field monitoring). Second, rapid growth of LPWANs in the limited spectrum raises the challenge of coexistence. Third, current (non-cellular) LPWANs are not designed to effectively support mobile nodes, e.g., tractors and drones in smart farming. Finally, LPWANs are not yet designed to support real-time communication hindering their adoption for many applications (e.g., process control). This project will address these challenges by developing theoretical foundations and systems for SNOW (Sensor Network Over White Spaces), an LPWAN that exploits unused TV spectrum, called white spaces. The results are extendable to many other LPWANs. Due to abundant white spaces, this project will enable connectivity for many rural applications. It will integrate education through course development, student research, and outreach to minority/underrepresented and K-12 students. This project will design and implement an LPWAN architecture and complete protocol stack based on SNOW to support scalable integration, coexistence, mobility, and real-time communication as follows. (1) It proposes a scalable seamless integration of multiple SNOWs by minimizing latency in the integrated network. Considering tradeoffs between quality and execution time, the solution approach includes both global optimization and fast heuristics. (2) It proposes a novel approach based on Reinforcement Learning to handle coexistence with many independent networks. This is done by developing an efficient Q-learning framework that is practical at low-power nodes. This will be the first Q-learning approach for LPWAN and for handling coexistence in a low-power network. (3) It proposes lightweight cross-layer approaches to enable mobility of SNOW nodes by handling Doppler effect as well as the effects of geospatial variation of white spaces. (4) It proposes a real-time communication framework for SNOW which will be the first result on real-time scheduling for LPWAN. (5) It will implement the proposed protocols on TI CC1310 and also on universal software radio peripheral devices. The protocols will be evaluated through experiments in two different radio environments -- an urban test-bed and an agricultural field piloting smart farming. 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.