WSNs event aggregation is a process of compounding some low-level events to a high-level event to removeduplicate information to be transmitted and thus save latency and energy. Current work on event aggregation reflects either latency restraint or aggregation job, but not both. Furthermore, current works only consider best aggregation for single high-level event, but many applications are composed of several high-level events. This paper studies the problem of aggregating multiple high-level events in WSNs with dissimilar latency restraints and aggregation functions. We propose relative matrix to describe aggregation function, which labels the resemblance among partial number of primitive events rather than the rising number of high-level events. Based on it, we suggest an event aggregation algorithm togetherseeing the two matters for single high-level event. This algorithm chains partial combination which is more general than completely aggregation. Through range the ideal base events, the work is stretched to multiple high-level events and consider the applied reliable constraint. The simulation results show that our algorithm outperforms currentmethods and saves substantial amount of energy (up to 15%in our system).

#