传感器非线性校正的遗传支持向量机方法.pdfVIP

第 25 卷 第 1 期 电子测量与仪器学报 Vol. 25 No. 1 · 56 · JOURNAL OF ELECTRONIC MEASUREMENT AND INSTRUMENT 2011 年 1 月 DOI: 10.3724/SP.J.1187.2011.00056 传感器非线性校正的遗传支持向量机方法* 刘 涛 王 华 (兰州理工大学机电工程学院, 兰州 730050) 摘 要: 针对传感器非线性校正中现有的较为常用的神经网络法的不足和支持向量机参数难确定的问题, 提出了一种遗传 算法和支持向量机相结合的方法, 阐述了支持向量机的非线性校正原理和遗传算法优化支持向量机参数的实现过程, 并分别采 用 BP 神经网络法和遗传支持向量机方法对压力传感器进行非线性校正。实验结果表明: BP 神经网络法使得传感器的最大相对 波动由初始的 22.2%降低到 1.12%; 而遗传支持向量机方法使其降低到 0.04%, 显著改善了传感器的性能, 取得了较好的效果。 关键词: 遗传算法；支持向量机；压力传感器；非线性校正 中图分类号: TP212 文献标识码: A 国家标准学科分类代码: 460.4030 Nonlinear correction of sensor using genetic algorithm and support vector machine Liu Tao Wang Hua (College of Mechanical and Electrical Engineering, Lanzhou University of Technology, Lanzhou 730050, China) Abstract: In order to overcome the disadvantages of neural networks more commonly used in nonlinear correction of sensors and the problems that the parameters of Support Vector Machine are difficult to determine, a method based on Ge- netic Algorithm and Support Vector Machine is presented. The nonlinear correction principle of this method is explained. The realization process of Genetic Algorithm optimizing the parameters of Support Vector Machine is introduced. A pres- sure sensor is adjusted with BP neural network and the method of Genetic Algorithm and Support Vector Machine, respec- tively. The experiment results show that BP neural network reduces the maximum relative fluctuation from the initial 22.2% to 1.12%. Moreover, the method of Genetic Algorithm and Support Vector Machine reduces the maximum relative fluctua- tion to 0.04%, therefore it evidently improves the performance of the sensor and achieves better result. Keywords: genetic algorithm (GA); sup