Support Vector Machine based Adaptive Calibration Technique for Resistance Temperature Detector

This paper proposes an adaptive calibration technique for temperature measurement using Resistance Temperature Detector (RTD) sensor based on Support Vector Machine (SVM). In practise, RTD has nonlinear response characteristics, and its output varies with variation in RTD. Support vector machine approach is used to design an adaptive calibration technique which (i) produces the output to have a linear mapping relationship to achieve RTD nonlinearity compensation, and (ii) to produce output adaptive to variations in temperature coefficients of RTD and reference resistance. The simulation results show that the proposed technique has fulfilled its set objectives. Since root mean square of percentage error is found to be 0.0015.