Performance enhancement of low-cost MEMS inertial sensors using extensive calibration technique

Low cost Micro Electrical Mechanical Systems (MEMS) inertial sensors are well known for their degraded performance if used in a self-contained navigation system. Due to large number of errors that affect the sensor's output. In this paper, an evaluation of performance improvement technique through experimental work for such sensors is introduced; the metric used in this work to evaluate the performance is the position error calculated using sensor's output through inertial navigation mechanization. For such evaluation, it is required to calculate the deterministic errors for gyroscopes and accelerometers that comprise the IMU, and to estimate the random errors using Allan variance technique for different bandwidths. To achieve this goal, a precise deterministic calibration process and random noise model of a chosen MEMS IMU is applied. The calibrated data then is used with the inertial navigation mechanization model, the test results show that the error in position is reduced dramatically when the calibration procedure is applied correctly to remove different types of errors. The estimated noise errors are compensated using Kalman Filter (KF) which led to improve the positioning accuracy.