Combined fuzzy-metaheuristic framework for bridge health monitoring using UAV-enabled rechargeable wireless sensor networks

Abstract

It is essential to monitor the health of important infrastructure (e.g., bridges) to maintain their functions. Visual inspections have been conventionally dominant in this regard, although they are susceptible to human errors. Wireless sensor networks (WSNs) provide an automated, convenient, and low-cost option for developing bridge health monitoring (BHM) networks. However, the constant use of WSNs for monitoring the structural and environmental health of bridges can pose a serious challenge due to the limited lifetimes of these networks that depend on the battery lifetimes of sensor nodes. This paper proposes a combined fuzzy-metaheuristic framework to maintain the BHM stability by using rechargeable sensors. This framework benefits from metaheuristic methods and the fuzzy logic to match the sensor network configuration management to the specific conditions of each bridge, recognizing that different bridges share very few common characteristics. Every new bridge is unique; hence, it is difficult to design a BHM paradigm that fits the conditions of all bridges. The proposed framework manages the current network configuration concerning the conditions of each bridge. This framework manages the network topology formation, information relay, and recharge by using a multipurpose objective, tuning control parameters, and controlling network activities in an optimization process. Moreover, unmanned aerial vehicles (UAVs) are employed to recharge sensor nodes under the proposed framework strategies to overcome the energy limitation of sensor nodes. The proposed framework is evaluated on three bridge scenarios: Hardanger Bridge, Bergsøysund Bridge, and New Carquinez Suspension Bridge. Compared to common WSN methods, it demonstrated superior performance under various conditions, including the rate of active and inactive nodes, energy efficiency, survival rate, stability, recharge delay, average node energy, recharge requests, and total packets received. The evaluation results demonstrate that the proposed framework significantly surpasses existing methods in terms of WSN performance metrics. The results show that the proposed framework outperforms existing methods by an average of 32.8 % for the Hardanger Bridge, 53.2 % for the Bergsøysund Bridge, and 31.2 % for the New Carquinez Suspension Bridge.