计算机工程与应用 ›› 2019, Vol. 55 ›› Issue (12): 140-144.DOI: 10.3778/j.issn.1002-8331.1803-0099

• 模式识别与人工智能 • 上一篇    下一篇

PSO优化多核RVM的模拟电路故障预测

颜学龙,陈  卓   

  1. 桂林电子科技大学 CAT实验室,广西 桂林 541004
  • 出版日期:2019-06-15 发布日期:2019-06-13

Analog Circuit Fault Prognostic Utilizing Particle Swarm Optimized Multi-Kernel Relevance Vector Machine

YAN Xuelong, CHEN Zhuo   

  1. The CAT Lab of Guilin University of Electronic Technology, Guilin, Guangxi 541004, China
  • Online:2019-06-15 Published:2019-06-13

摘要: 针对模拟电路健康管理的特点,提出了一种基于PSO优化多核RVM的模拟电路故障预测方法。利用参数分析得到电路的输出频域响应作为特征,计算其与电路无故障标准响应的欧氏距离来表征电路元件健康值,将多个核函数线性组合,并用PSO优化多核RVM参数后的模型实现对各个时间点元件的健康值变化轨迹进行预测。仿真结果表明,该方法在小样本情况下,预测效果优于单一核函数的RVM模型,适用于健康管理中实时预测,具有较好的实用性。

关键词: 相关向量机, 核函数, 欧氏距离, 模拟电路, 粒子群寻优

Abstract: Aimed at the characteristics of prognostic and health management in analog circuits, a new method for analog circuit fault prediction based on particle swarm optimized multi-kernel RVM is proposed in this paper. Firstly, fault features are extracted by parameter analysis of the circuit, then the health value of circuit components is represented through calculating the Euclidean distance. Finally, combining different kernel function linearly, the trend of health value trajectory with respect to time points can be predicted by particle swarm optimization, the proposed approach is appropriate for real time prediction in prognostic and health management and is better than the single kernel RVM model in the case of small sample. Simulation results validate the good practicability and effectiveness of the proposed method.

Key words: Relevance Vector Machine(RVM), kernel function, Euclidean distance, analog circuits, Particle Swarm Optimization(PSO)