Heterogeneous multi-channel neighbor discovery formobile sensing applications: theoretical foundationand protocol design

Abstract

Neighbor discovery is of paramount importance in mobile sensing applications that rely heavily on data timely collected and shared among nearby users. Guaranteed discovery with bounded latency and supporting heterogenous duty cycles to provide fine-grained control of energy conservation levels are among the most crucial requirements in the design of efficient neighbor discovery protocols. While simultaneously satisfying these two requirements is non-trivial, the situation is exacerbated if the operating frequencies of mobile devices span multiple channels and discovery occurs only if nodes switch to the same channel. In this paper, we formulate this problem as heterogeneous multi-channel neighbor discovery problem and establish a theoretical framework of the problem, under which we derive the performance bound of any neighbor discovery protocol. Based on the theoretical results, we then develop Mc-Dis (Multi-channel Discovery), a novel multi-channel discovery protocol that (1) achieves guaranteed discovery with order-minimal worst-case discovery delay and (2) supports almost all duty cycles to provide fine-grained control of energy conservation levels.