1 Dynamic Measurement Errors Prediction Model of 2 Sensors Based on NAPSO - SVM 3

Dynamic measurement error correction is an effective method to improve the sensor 15 precision. Dynamic measurement error prediction is an important part of error correction, support 16 vector machine (SVM) is often used to predicting the dynamic measurement error of sensors. 17 Traditionally, the parameters of SVM were always set by manual, which can not ensure the model’s 18 performance. In this paper, a method of SVM based on an improved particle swarm optimization 19 (NAPSO) is proposed to predict the dynamic measurement error of sensors. Natural selection and 20 Simulated annealing are added in PSO to raise the ability to avoid local optimum. To verify the 21 performance of NAPSO-SVM, three types of algorithms are selected to optimize the SVM’s 22 parameters, they are the particle swarm optimization algorithm (PSO), the improved PSO 23 optimization algorithm (NAPSO), and the glowworm swarm optimization (GSO). The dynamic 24 measurement error data of two sensors are applied as the test data. The root mean squared error 25 and mean absoluter percentage error are employed to evaluate the prediction models’ 26 performances. The experiment results show that the NAPSO-SVM has a better prediction precision 27 and a less prediction errors among the three algorithms, and it is an effective method in predicting 28 dynamic measurement errors of sensors. 29