基于自适应加权平均多传感器的泥石流灾害预报模型

Based on adaptive weighted average multi-sensor debris flow prediction model

针对泥石流监测中多传感器数据融合问题,提出了一种分批估计的自适应加权平均多传感器数据融合算法,该算法将同类无线传感器采集的数据融合。并建立粒子群算法优化极限学习机来预测泥石流的发生概率,仿真结果表明将融合后的数据输入改进极限学习机的泥石流灾害预报模型中,可以有效提高模型预报准确率以及模型训练时间。

Aiming at the problem of multi-sensor data fusion in debris flow monitoring,an adaptive weighted average multi-sensor data fusion algorithm is proposed for batch estimation,which will affect the rainfall,soil moisture content,pore water pressure,slope,infrasound,ditch bed ratio drop and relative height difference of debris flow.The particle swarm optimization limit learning machine is established to predict the probability of debris flow,and the activation function of the ELM is optimized.The input of the improved PSO-SELM is the selected disaster factor and the output is the probability of debris flow.The simulation results show that the fused data can be input into the debris flow disaster prediction model based on the improved limit learning machine,and the fused data can be compared with the data without data fusion.The prediction accuracy and training time of the model can be improved effectively.