A Center-of-Gravity and Fuzzy C-Means-Based Clustering Routing Algorithm for Wireless Sensor Networks
In wireless sensor networks with resource-constrained and randomly deployed nodes, extending network lifetime remains a critical challenge. Energy efficiency, as a key determinant of network longevity, has become a central focus in routing protocol design, positioning clustering mechanisms as a widely adopted solution. This paper investigates cluster head election, cluster formation, and data transmission path construction to develop a clustering routing strategy that balances energy consumption and improves efficiency. To overcome limitations in existing cluster head election objective functions that inadequately capture the synergy between node energy and spatial distribution, this paper proposes CGFCM, an adaptive clustering routing protocol integrating a center-of-gravity model with the fuzzy C-means algorithm. The protocol defines a cluster head election objective function jointly considering node energy status and spatial location. A center-of-gravity formulation characterizes node mass based on energy and location, while nodes are classified into internal, boundary, and transitional categories according to cohesion and synergy factors, enhancing topology awareness. In cluster formation, fuzzy C-means clusters internal nodes, and non-internal nodes join via a fitness function. Candidate cluster heads are selected from internal nodes using an energy threshold, from which the optimal cluster head is chosen through an objective function incorporating centroid distance, node mass, residual energy, and node density. Additionally, rotation intervals and relay paths are dynamically adjusted to reduce reconfiguration overhead and alleviate cluster head load. Simulation results show that CGFCM achieves superior performance in balancing network energy consumption and extending network lifetime. In the scenario where the base station is located at (100, 100), the FND of CGFCM is improved by 40.50%, 51.42%, and 17.65% compared with EHCR-FCM, PSO, and EDFIKM algorithms, respectively, which demonstrates its performance advantages under typical deployment conditions.