P2.9.33 Unsupervised adjustment of centers in RBF networks for sensor drift compensation
- 14th International Meeting on Chemical Sensors - IMCS 2012
2012-05-20 - 2012-05-23
- P2.9 Technology and Application
- N. Kim, H. Byun, K. Kwon - School of Electronics, Information & Communication Engineering, Kangwon National University (Korea), K. Persaud - School of Chemical Engineering and Analytical Science, The University of Manchester (U.K.), J. Lim - Biomedical Research Institute, Kyungpook National University (Korea)
- 1802 - 1804
In our previous research for sensor drift compensation, the unsupervised signal processing approach of readjusting the weights of Radial Basis Function Network (RBFN) based on probability distribution functions (PDFs) has shown a possibility to solve the sensor drift problems, but it was not satisfactory still showing deteriorated distributions in some gases. In this paper, a new readjustment method for another parameter, center of RBFN based on PDFs is proposed for sensor drift compensation. Compared to the case of weight readjustment only, the proposed method yields significantly decreased dispersions for most gases and even for the ones deteriorated after weight readjustment. This proves the proposed method for additional center readjustment to be more effective in sensor drift compensation.