Battery Service-Life Enhancement Using Temporal Data Partitioning Mechanism for Sustainable IoT Applications

Majority of event-driven IoT applications in wireless multimedia sensor networks (WMSN) nodes often acquire redundant data of the same target or event, which show a very high degree of temporal correlation, and causes high consumption of energy due to transmission of redundant information; resulting a massive depletion of battery sources of the sensor node and reduces its service-life. To mitigate this issue, this paper presents a novel twofold encoding scheme to tackle the problem of highly temporally correlated data acquisition by sensor devices and redundant data transmission in smart IoT applications. The scheme first employs the Kruskal–Wallis Hypothesis test to partition the data, followed by encoding the data using the Redundant Binary Number System (RBNS) produces high proportion of 0 s in the encoded string. This in turn increases Silent Symbol period during transmission. To further conserve energy, a hybrid FSK-ASK modulation and demodulation technique is employed during communication with a non-coherent receiver, leading to a significant reduction in transmitter energy. We simulate the proposed raw sensor data encoding technique exploiting runs of individual encoded symbols of real-life sensor data source with commercially available transceiver Atmel ATR 2406. The simulation result shows about 0.041 mW h energy consumption per day employing Atmel ATR 2406 transceiver utilizing our proposed scheme which is about 58.0–83.6% less consumption compared to popular communication schemes like CNS, PCA, NCS, and ALDC schemes, respectively. The proposed idea also extended the CR2032 lithium coin cell battery life of the sensor module to 2.41 years which is 58–83% more compared to existing schemes, and resulted in a reduced CO 2 emission rate of 0.036 mg/day, making the scheme environmentally sustainable.