Time filter

Source Type

Beijing, China

Yu X.-T.,ASIT Co. | Zhang L.,ASIT Co. | Chen G.-F.,Chinese Peoples Liberation Army | Zhou F.,ASIT Co. | Xu Z.-S.,ASIT Co.
Zhongguo Guanxing Jishu Xuebao/Journal of Chinese Inertial Technology | Year: 2010

There exits a large error in accelerometer static model when using linear approximate model to identify it. Support vector regression (SVR) has good regression performance, especially in small-sample, nonlinear and high-dimensional feature space. So a new model identification method based on SVR is proposed to identify the accelerometer static model. Particle swarm optimization (PSO) algorithm is adopted to optimize the parameters of SVR, which cost less time compared with the trial-and-error method. The 12-position static rolling test of Quartz flexure accelerometer is made using precision optical dividing head. The experiment results demonstrate that the identification accuracy of static model by the proposed method is approximately more than double the one by the least square method. Source

Yu X.-T.,ASIT Co. | Zhang L.,ASIT Co. | Guo L.-R.,ASIT Co. | Zhou F.,ASIT Co. | Yu H.,ASIT Co.
Zhongguo Guanxing Jishu Xuebao/Journal of Chinese Inertial Technology | Year: 2011

Environment temperature is an important factor for the precision of accelerometer. The mechanism of temperature's influence on bias and the scale factor of accelerometer is analyzed. An identification method for the temperature model of bias and scale factor of quartz flexure accelerometer is proposed which is based on least squares wavelet support vector regression (LS-WSVR). The experiments under different temperature conditions were conducted to verify the good regression performance of the proposed LS-WSVR, and the experimental data was used as the training data of LS-WSVR. The experiments of temperature with fixation point were conducted, and the experiment results show that the accuracy of accelerometer compensated with LS-WSVR model is better than that of using least square model or having no compensation. Source

Discover hidden collaborations