Abstract:

In order to improve the efficiency of data fusion in the wireless sensor network (WSN) of a uranium tailings pond, reduce redundant data transmission, and extend network lifespan, an innovative data fusion algorithm was proposed, namely the SAPSO-BP data fusion algorithm based on improved SA and PSO optimized BP neural network. The algorithm integrated the global search capability of the SA algorithm with the efficient optimization mechanism of the PSO algorithm, incorporating dynamic inertia weights and mutation operators to enhance global search ability and avoid local optima. Furthermore, the improved algorithm was used to optimize the weight matrix and threshold parameters of the BP neural network, constructing a high-performance multi-sensor data fusion model, which was applied to radionuclide monitoring in uranium tailings ponds. The results show that the SAPSO-BP algorithm outperforms the compared algorithms in terms of data fusion accuracy, network energy consumption, and network lifespan. Compared with the traditional BP algorithm, it reduces mean relative error(MRE) and root mean square error(RMSE)by up to 40% and 45%, respectively, and improves the goodness of fit to 0.908 3. Additionally, it delays the first node death to approximately 1 180 rounds, extends the overall network lifespan to about 1 500 rounds, and achieves lower node energy consumption and a more balanced energy distribution.

Key words: uranium tailings pond, nuclide monitoring, simulated annealing (SA), particle swarm optimization (PSO), back propagation (BP) neural network, wireless sensor network(WSN), data fusion