Propagation and system modeling: New propagation models are needed for ultra-dense networks. Driven by legacy cellular systems, most models currently available are tower-to-ground with base stations typically at 30 m and above. Device-to-device links, however, will be ground-to-ground: picocells on lightpoles and indoor femtocells placed on tabletops or mounted on walls or ceilings will be < 10 m above ground. The problem is that few ground-to-ground channel models to date exist in the literature. In addition, most channel models assume independent fading between links. In ultra-dense networks in particular, this assumption does not hold because there will be significant correlation between links, necessitating the study of joint links. NIST also has expertise in this area, but much more work needs to be done for sufficient characterization. In addition, the introduction of carrier aggregation to provide additional bandwidth will require developing ground-to-ground and indoor channel models for new frequency bands. In addition to channel propagation models, detailed system models are needed to accurately characterize and develop signal processing algorithms at the physical (PHY) layer, as well as Medium Access Control (MAC) layer and routing protocols. A big issue that needs to be dealt with is interference mitigation:
- Coordination algorithms for synchronizing cells, managing power, and selecting carriers needs to be examined in terms of their inaccuracy (timing offsets, etc.), and how these impact performance.
- Determining the limits of densification (i.e. how much capacity reuse is possible before interference or channel effects or overhead causes saturation) is an important issue that needs to be addressed . This will require development of additional models and simulation tools to answer.
Mobility management and self-organization: Mobility management for hyper-dense networks is an important problem that needs to be addressed. A typical mobile device will be within coverage of multiple small cells. Thus, determining which standard and spectrum to utilize and which base station to associate becomes a truly complex task. While high-mobility nodes can be safely restricted to macrocells, leaving the low-mobility cells to exclusively use the small cells, there is still an issue involving rapid handovers and ping-pongs for UEs moving through very small coverage areas. Possible solutions to be examined include:
- The use of so-called virtual cells, in which a collection of small cells are treated as a single large cell.
- Coordinating communication between an ensemble of base stations and the mobile device is required, in addition to determining how much traffic to transmit from each base station.
- Using mobility state information to modify the mix of traffic from multiple base stations in response to the mobile's changing position.
- Developing methods for measuring and disseminating channel information in order to successfully implement massive cooperative multi-point transmission schemes.
Also, small cells in ultra-dense deployments that are not serving UEs may go into sleep mode to conserve power, resulting in a network populated by base stations that turn on and off in response to UE migration. Additionally, future networks may include small cells that are deployed by customers without direct setup by the service provider. This means that base stations in ultra-dense deployments must be able to autonomously perform self-organizing functions such as discovery, synchronization, authentication, and power control, and selection of appropriate common carriers in networks that use carrier aggregation.