PI: Prof. Hsuan-Jung Su
Co-PI: Prof. Hung-Yun Hsieh
Intel Champion: Dr. Minnie Ho & Co-Champion: Dr. Shu-ping Yeh
Investigate and provide physical layer (PHY) and medium access control layer (MAC) technologies to support future machine-to-machine (M2M) wireless networks, with the initial focus on improving the capability of current human-to-human (H2H) wireless networks in handling high density M2M applications and intelligently share spectrum with them.
The basic concepts are to improve frequency reuse, utilize idle spectra through sensing and opportunistic access, and improve areal spectral efficiency through interference mitigation and distributed coding which takes advantage of correlation between machines.
- Clustering: The machines form smaller clusters and transmit to “cluster heads”. This results in short range transmission, power saving and better frequency reuse. Cluster heads perform multi-hop transmission to the data collection center. Machines sense to avoid interfering or being interfered by H2H communications. The idea is comparable to the concept of traffic off-loading in the next-generation cellular networks to deploy smaller pico or femto cells, but clustering is more flexible, does not need network planning.
- Distributed Coding: Distributed source coding for correlated data through Wyner-Ziv coding. For multi-hop transmission, combined routing and spatial diversity. The goal is to provide practical yet close-to-theoretical-bound coding/decoding design (with finite code block-length), tailored for M2M traffic.
- Interference Mitigation: Apply inter-cluster interference mitigation to achieve frequency reuse one, with: soft frequency reuse, adaptive frequency reuse, coordinated transmission/scheduling, and interference alignment and cancellation. The machine-human interference is handled by sensing and opportunistic access.