摘要

For real-time applications in mobile low-duty-cycle (DC) wireless sensor networks (WSNs), discovering neighbours needs to be achieved in both short average latency and high fraction of discoveries (FDs) with low energy consumption. A novel asynchronous probabilistic method is proposed, in which a node randomly chooses a prime, e.g. p, from a pre-configured prime set and wakes up only at the first slot during every p continuous slot which is for a hyper-cycle length, i.e. a fixed number continuous slots, determined by the prime sets, and then selects another prime for the next hyper cycle. For low-DCs, especially <2%, one can flexibly configure different prime sets for the given DC, and the more the primes in the sets, the lower the average discovery latency (ADL) that is achieved. Simulations show that the proposed method can not only achieve a shorter ADL than SearchLight, but also obtain a shorter and smaller tail on FDs than Birthday.