基于PFP-WRVM的多功能传感器状态自确认研究

Research on state self-validation of multifunctional sensor based on PFP-WRVM

针对传统多功能传感器无法状态自确认的局限性,提出了一种基于小波相关向量机和多项式有限冲激响应预测的测量值及其不确定度的状态自确认方法。首先从模拟量和数字量的信号处理方面,分析其不确定度源,其次建立基于小波相关向量机和多项式有限冲激响应预测的多功能自确认模型,并对正常和故障下确认值及其不确定度传播过程进行详细的理论推导,最后搭建可同时测量温度、湿度、氢气浓度的多功能自确认传感器实验系统。实验结果表明,提出的自确认方法可全面给出正常和故障下多功能传感器的确认值及其不确定度,实现短时故障恢复(相对误差约1%),初步完成多功能传感器的状态自确认。

In order to solve the problem of state self-validation of multifunetional sensor, self-validating this paper proposes a novel approach for the measurement value and its uncertainty based on wavelet relevance vector machine （WRVM） coupled with polynomial FIR （finite impulse response） prediction （PFP）. Firstly, the uncertainty sources are analyzed in terms of the signal processing of analogue and digital quantities. Secondly, a self-validating model is built based on the proposed strategy. Besides, the propagation processes of the validated value and uncer- tainty under normal and faulty states are theoretically deduced in detail. Lastly, a multifunctional self-validating sensor experimental system is set up to measure temperature, humidity and hydrogen concentration simultaneously. Experiment results show that the proposed approach provides a good solution to state self-validation of multifunctional sensor, implements fault recovery with lower relative error of 1% , and accomplishes the preliminary state self-validating function for the validated value and uncertainty.