Mobile sink-based data gathering in wireless sensor networks with obstacles using artificial intelligence algorithms

Abstract

Data gathering is the fundamental task of Wireless Sensor Networks (WSNs). Making a balance between the energy consumption of sensors and the data gathering delay is considered as an important concern in this regard. Finding a solution to this issue becomes more challenging in the presence of obstacles in the field. The present study proposes an effective algorithm for Data Gathering in WSNs with OBstacles, namely DGOB. The algorithm clusters the nodes and exploits a Mobile Sink (MS) for data gathering from the cluster heads to diminish the exhausted energy. Accordingly, it decomposes the original problem into two phases of cluster and MS tour construction. In this study, two methods are presented in the first phase to derive high-quality clusters per round. The first method, employed in the first round, exploits hierarchical agglomerative clustering and ant colony optimization to construct high-quality clusters in the presence of obstacles. The second one, applied in the succeeding rounds, updates the present clusters using Genetic Algorithm (GA). In the second phase of DGOB, an effective tour construction method is introduced based on GA and multi-agent reinforcement learning. The extensive simulation results verify that DGOB improves energy consumption and network lifetime by 34% and 80% compared to the existing approaches.