7. Conclusions
In this paper we have considered a scenario where a large number of WBSNs meets at one point and competes for channel resources (time and frequency). In particular, we have assessed the relative impact of a range of static and adaptive schemes for frequency allocation on two new major performance metrics, the satisfaction rate and the carrying capacity. We have found that for most schemes by far the most influential parameters are related to the MAC backoff process. By comparing the considered frequencyadaptive schemes against an idealized scheme allocating WBSNs in both frequency and time, we have furthermore found that on the one hand frequency adaptation can provide substantial gains, but on the other hand further gains can potentially be reached by adding the capability to properly separate WBSNs operating on the same channel in time (“phase adjustment”). There is substantial potential for future work. As a next step, we plan to design phase adjustment schemes which operate without any centralized entity and are integrated with frequency adaptation. Secondly, we intend to assess the considered schemes in more general scenarios with dynamic WBSN populations, the presence of hidden-terminal situations (i.e. where not all WBSNs are located on the same spot) and heterogeneous loads. We are also currently working on a measurement campaign to gain experimental insight into the performance of the schemes discussed in this paper. Another interesting direction is to extend this work towards a combination of internal and external interference, e.g. coming from WiFi or Bluetooth.
