Optimal sensor placement for the routine monitoring of urban drainage systems: A re-clustering method

Wang, Siyi; Zhang, Xiangwei; Wang, Jiaying; Tao, Tao; Xin, Kunlun; Yan, Hexiang; Li, Shuping

doi:10.1016/j.jenvman.2023.117579

Cited by 4 publications

(1 citation statement)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the domain of optimal sensor placement relying on clustering algorithms, Wang et al [19] harnessed the k-means algorithm to ascertain both the quantity and positions of sensors. Their innovative contribution was the introduction of a re-clustering strategy, which served to enhance the efficacy of k-means clustering results and elevate the overall monitoring efficiency of sensor deployment.…”

Section: Introductionmentioning

confidence: 99%

Three-Way k-Means Model: Dynamic Optimal Sensor Placement for Efficient Environment Monitoring in Pig House

Li,

et al. 2024

Animals

View full text Add to dashboard Cite

Sensors were of paramount importance in the context of poultry and livestock farming, serving as essential tools for monitoring a variety of production management parameters. The effective surveillance and optimal control of the swine facility environment critically depend on the implementation of a robust strategy for situating the optimal number of sensors in precisely the right locations. This study presents a dynamic sensor placement approach for pigsties using the three-way k-means algorithm. The method involves determining candidate sensor combinations through the application of the k-means algorithm and a re-clustering strategy. The optimal sensor locations were then identified using the Joint Entropy-Based Method (JEBM). This approach adjusts sensor positions based on different seasons (summer and winter) to effectively monitor the overall environment of the pigsty. We employ two clustering models, one based on particle swarm optimization and the other on genetic algorithms, along with a re-clustering strategy to identify candidate sensor combinations. The joint entropy-based method (JEBM) helps select the optimal sensor placement. Fused data from the optimal sensor layout undergo a fuzzy fusion process, reducing errors compared to direct averaging. The results show varying sensor needs across seasons, and dynamic placement enhances pigsty environment monitoring. Our approach reduced the number of sensors from 30 to 5 (in summer) and 6 (in winter). The optimal sensor positions for both seasons were integrated. Comparing the selected sensor layout to the average of all sensor readings representing the overall pigsty environment, the RMSE were 0.227–0.294 and the MAPE were 0.172–0.228, respectively, demonstrating the effectiveness of the sensor layout.

show abstract

Section: Introductionmentioning

confidence: 99%