Space-Time Derivative-Based Prediction: A Novel Trickling Mechanism for WSN
César Huegel Richa, Antônio Augusto Fröhlich, Giovani Gracioli · 2017
Time series prediction techniques reduce the number of messages generated at the application level, saving energy spent in the communication and, consequently, extending the network lifetime. Trickle is a well-known time series prediction mechanism commonly used to decrease the number of transmitted messages in Wireless Sensor Networks (WSN) and thus save energy. This paper presents the Space-Time Derivative-Based Prediction (ST-DBP), a novel Trickling mechanism to suppress data transmission in space-time regions in WSNs. We integrate ST-DBP with the Trustful Space-Time Protocol (TSTP), an application-oriented, cross-layer communication protocol, and compare two variations of the ST-DBP with the original DBP using real data from a Solar Farm in terms of suppression data ratio. Our results show that the two variations of the ST-DBP outperform the original DBP.