A Study On The Approximation Algorithms Of Aggregation Query In Wireless Sensor Networks

In the time of Internet of things,wireless sensor network is one of the most important techniques.The energy,computation,communication and storage capabilities of node are very limited in sensor network,we are facing a big challenge to make efficient aggregation query.The aggregation query is one of the most important operations of users.In most cases,the user can accept the results of certain error.This paper studies approximate aggregate queries algorithm about SUM and Top-k,and put forward the approximate query algorithm of high efficiency and energy saving.Since the existing summation aggregation algorithms are almost duplicate-sensitive,an approximate algorithm Flajolet-Martin SUM(FM-S)of distinct summation query for Wireless Sensor Network(WSN)was proposed.In FM-S,each node in WSN combined the FM-S algorithm and the random number sample of binomial distribution to make hash conversion so as to fill a summation sequence of length L.Each node forwarded the generated sequence to the father node in routing tree.Then the root node received the summation sequence of whole network.Finally,according to the sequence of root node,the approximation summation value of distinct records in sensor networks could be obtained.The experimental results show that the distributed algorithm is low power consumption,high fault tolerance,robustness and scalability.This paper introduces the approximate algorithm of Top-k query based on sampling and weight in wireless sensor network.The algorithm divides the network into several disjoint clusters,respectively in the sink node and the nodes in cluster to take sampling process.Tax greater weight for reliable and important sensor node in the process of sampling.According to the sensor node sensing data has a time correlation,sampling threshold filtering in the cluster.Each cluster head node receives a Top-k candidate subset of the cluster,and then sent the subset to the sink node,finally,the sink node can receives a Top-k sample candidate that represent the whole network.The simulation experiments show that the algorithm only needs to send small data,the sample of smaller,and can satisfy arbitrary precision requirement.