Intelligent odor plume source localization using a simulated-annealing-optimized BP neural network for IoT-enabled industrial environments

In the context of Industry 4.0 and the Internet of Things (IoT), real-time monitoring and rapid response to hazardous environmental leaks are essential for ensuring industrial safety. To address the challenge of accurate odor source localization in complex environments, this paper proposes an improved backpropagation neural network optimized with simulated annealing (SA-BP) for odor plume tracking and positioning. By leveraging the probabilistic jumping capability of simulated annealing to enhance global search performance, and integrating an improved BP network with strong generalization ability, the proposed method effectively mitigates the local-minimum issue in traditional models. The strategy is well-suited for deployment in multi-node wireless sensor networks, where distributed sensing nodes collect odor concentration data and the SA-BP algorithm performs rapid inference of the source location, enabling automated leak localization and risk alert in IoT-enabled industrial environments. Simulation results demonstrate that the SA-BP approach outperforms conventional algorithms in terms of iteration reduction, faster convergence, and higher localization accuracy. The findings confirm the feasibility of the proposed method and highlight its potential to support safety monitoring systems in chemical storage and smart-factory scenarios under the Industry 4.0 framework.