CH Selection and Compressive Sensing-Based Data Aggregation in WSN Using Hybrid Golden Circle-Inspired Optimization

Abstract

The arbitrary distribution of sensor nodes and irregularity of the routing path led to unordered data, which is complex to handle in a wireless sensor network (WSN). To increase WSN lifetime, data aggregation models are developed to minimize energy consumption or ease the computational burden of nodes. The compressive sensing (CS) provides a new technique for prolonging the WSN lifetime. A hybrid optimized model is devised for cluster head (CH) selection and CS-based data aggregation in WSN. The method aids to balance the energy amidst different nodes and elevated the lifetime of the network. The hybrid golden circle inspired optimization (HGCIO) is considered for cluster head (CH) selection, which aids in selecting the CH. The CH selection is done based on fitness functions like distance, energy, link quality, and delay. The routing is implemented with HGCIO to transmit the data projections using the CH to sink and evenly disperse the energy amidst various nodes. After that, compressive sensing is implemented with the Bayesian linear model. The convolutional neural network-long short term memory (CNN-LSTM) is employed for the data aggregation process. The proposed HGCIO-based CNN-LSTM provided the finest efficiency with a delay of 0.156 s, an energy of 0.353 J, a prediction error of 0.044, and a packet delivery ratio (PDR) of 76.309%.