Compressive Sensing Technology Used On Wireless Sensor Networks

With the popular application of the Internet of things, wireless sensornetworks nowadays have been used in more and more applications. Animportant application of wireless sensor networks is monitoringenvironmental signals and each sensor node monitors the environment onthe sampling time, and passes the sensed data to the sink node. Sink nodecollects data and make the subsequent data processing.Because of the distinctiveness of monitoring applications, the networkneeds to collect a large quantity of data, and transmit the data to sink node.The process of transmitting the sensed data to sink node requires lots ofenergy. The power supply for sensor nodes in the wireless sensor networksare usually battery. Since sensor nodes are usually placed in areas whichare not easy for people to reach, the battery for the sensor node can hardlybe replaced after it is power off, so the energy issue is a crucial problem inwireless sensor networks. In order to control cost in practical applications,the sensor nodes constructing the sensor networks should be cheap in price,and these sensor nodes mainly are limited in energy supply, the ability ofdata processing, and the sampling frequency. At present, there are manyscholars making research on the energy efficiency problem, andintroducing compressive sensing technology to data gathering process inwireless sensor networks, which proves to have a good effect on reducingthe amount of data transmitting through the network and help to prolongthe lifetime of the wireless sensor network.In this paper we make use of compressing sensing technology onwireless sensor networks, and reduce the energy consumed by transmissionand sampling process by reducing both the amount of data transmission inthe network and the average sampling frequency of the sensor nodes, sothat the lifetime of the network can be extended. The research work is asfollows:(1) Study the topology of data collection in wireless sensor network, and make full use of the characteristic of the sampling matrix, so that thesensor node selects the node with a similar transmission state as its parentnode in constructing the data gathering tree in the data delivery process. Inthis way we can ensure some sensor nodes, which don’t need to transmitits own sensed data to its parent node, remain still in the dormant state,rather than work as a relay for its child node. We can build a data gatheringtree by selecting the appropriate parent node for each sensor node toreduce the transmission cost in the network. Considering the load balanceproblem may have a bad effect on the lifetime of the network, we adjustthe structure of the data gathering tree to reduce the number of overloadednode in order to further prolong the lifetime of the network.(2) Study the sensor nodes’ sampling frequency. Since the sensing dataprocess requires energy, and the data monitored by the same sensor nodebetween adjacent sampling times have a time redundancy in generalapplication scenario of the wireless sensor networks, we make the sensornodes to determine the temporal redundancy of the signal and then adjusttheir own sampling frequency automatically according to the sensed data.We can then reduce the energy consumed on the sensing module for thesensor nodes by reducing the average sampling frequency of the sensornodes in order to prolong network lifetime.