Clustering-based Data Collection Using Constructive Interference In Wireless Sensor Network

With the trend of Internet of Things(Io T),wireless sensor network(WSN),has been focused by a lot of researchers in recent years,as one of the crucial technologies of Io T.Many giant technology companies have endeavored in this field.On the other side,applications like 5G,artificial intelligence,industrial internet have promoted the rapid development of data center.Increasing number of data centers brings the serious problem of energy wasting.To tackle this problem,wired sensor network is traditionally used to monitor the operation status of data center.Yet,wired sensor network is complicated,costly and inconvenient to maintain.WSN provides a novel strategy for this problem.However,efficient data collection has been a tough task for this large-scale wireless sensor network.This paper combines concurrent transmission with data collection,and achieves efficient data collection based on clustered data gathering algorithm.The main research results are given as follows:Concurrent interference based flooding data collection protocol has been proposed on the basis of thorough understanding of concurrent transmission and network flooding.The error caused by time offset in one period is decreased to satisfy the slot assignment of the proposed data collection algorithm based on concurrent transmission.The prototype is evaluated on COOJA simulation and testbed as well.Simulation and experiment results show that radio duty cycle of concurrent transmission based flooding data collection is better than that of CTP.Moreover,the energy consumption of sensor nodes is uniform.The advantage is more obvious with smaller IPI.The stability has been verified with 3 hours' operation under the condition of = 2.And the proposed achieves reliable data collection with the 100% packet reception rate.Synchronous transmission has scalability problems,which limited the development of data collection protocol based on concurrent transmission.This thesis proposes to tackle the problem from the viewpoint of clustering.The use of concurrent transmission in clustering-based data collection has not been explored in literature.Herein,a criterion for clustering principle is designed from viewpoint of load balance and connectivity balance to ensure the of rationality for clustering.And the data frames are fine-grained designed according to the requirements of clustering-based flooding data collection algorithm.Evaluation is conducted with COOJA simulation,and testbed experiment.From the viewpoint of theory analysis,we derive the ratio of radio duty cycle and packet load overhead between clustering flooding protocol and non-clustering flooding protocol.The metrics of goodput,packet reception rate,and radio duty cycle are verified by COOJA simulation and testbed experiment.Simulation and experiment show that goodput of three-cluster protocol is 2.78 times of non-cluster protocol,which is basically in accordance with theoretical analysis.Experiment results on a testbed consisting of 30 Telos B nodes show that packet receive rate of three-cluster protocol is almost 100%,and the radio duty cycle is 0.36 of non-cluster protocol.COOJA simulations and testbed experiments verify the high efficiency and scalability of clustering-based flooding data collection.