The public's demand for data has been growing at an enormous rate and growth is showing no signs of diminishing. In particular, users are increasingly demanding data from wireless devices such as smart phones, tablets, and laptops. To accommodate this demand, the number of wireless access points (APs) being deployed in high-demand areas has increased substantially and this has resulted in significant interference between nearby wireless communications, which reduces performance and can result in unmet user demands. This research is studying the development of advanced coding, signal processing, and scheduling techniques based on multiple-input multiple-output (MIMO) communication to significantly reduce interference between competing transmissions, thereby substantially increasing the overall user demand that can be accommodated. The project will benefit society by opening up new types of high-demand applications, such as augmented and virtual realities, to wireless users and will maintain performance for these and other applications in the face of ever-increasing data demands. The project will also enhance education through unified courses that integrate advanced signal processing developments and wireless networking instead of the traditional approach, which treats these as separate disciplines.
This research is investigating new nonlinear processing techniques for multicell MIMO within a framework of AP coordination for single clusters of 2-6 APs, which covers home networks and small enterprises. To address the demands of large enterprise wireless networks, the project is also researching a hierarchical approach, which employs innovative techniques to deal with the problem of edge-node interference between clusters to scale performance across multiple AP clusters. The project will develop algorithms, techniques, and protocol extensions in the area of performance scaling for multi-AP wireless networks. Among the contributions are: (1) a novel non-linear precoding approach for maximizing downlink performance in an AP cluster, (2) a new algorithm for selecting groups of users for concurrent communication in a cluster and an algorithm, based on a classic algorithm from VLSI circuit partitioning, for iteratively improving the user groupings, (3) a slotted protocol that segments mobile and stationary users and allocates mobile slots on demand, (4) an approach to providing consistent access point orientation across AP clusters so as to eliminate interference between clients at cluster boundaries, (5) a loose coordination protocol executed between clusters that will allow neighboring clusters to handle interference between them without the costly overhead of exchanging full channel state information across the network, and (6) new physical-layer-aware techniques for client-coordinated access to multiple providers networks.
