Accuracy enhancement of MEMS accelerometer by determining its nonlinear coefficients using centrifuge test

In order to increase accuracy of inertial navigation system, it is necessary to calibrate inertial measurement unit (IMU). In some applications, acceleration applied to accelerometer is larger than its linear operation area, thus input-output relationship deviates from linear form. Main purpose of this paper is to estimate coefficients of nonlinear terms in input-output relationship of accelerometer to increase its measurement accuracy. A centrifuge device is used to generate high acceleration levels as calibration reference. In this paper, linear and nonlinear coefficients of a low cost MEMS accelerometer (x axis) are estimated using continuous calibration method (using earth gravity vector (0-1g)) and centrifuge test (accelerations between 0 and 16g generated by the centrifuge). In addition, adequacy of the nonlinear model is evaluated. The well-known F test is used to compare linear and nonlinear models. Also, errors of linear and nonlinear models are compared using Mean Square Error (MSE). Experimental results showed that nonlinear model has improved measurement accuracy about 11.7%.