Application of Improved Kalman Filter in High Precision Electronic Scales

Yanming Wang, Haifeng Chen

Abstract: Aiming at the problem of large error and poor stability caused by the nonlinear error in the input-output characteristics of the resistance strain type weighing sensor and the random interference noise in the system acquiring data, a Kalman filter processing method combined with piecewise linear interpolation calibration is proposed. The method not only eliminates the accumulated error, but also achieves the fitting and calibration of the characteristic curve of the nonlinear weighing relationship. What's more, the precision is improved while the requirements of the system to the environment are broadened. The test results show that the electronic scale has a range of 1kg and a precision of 0.02g, and the repeated test error is less than 0.03g, which achieves the expected effect.