A direct approach for high-quality MEMS based IMU/INS production

Abstract

This paper describes an industrial process for microelectromechanical systems (MEMS) based inertial measurement unit (IMU) and inertial navigation systems (INS) production. Producing high accuracy sensor in a limited time and with a robust process is a universal problem in IMU and INS production. The challenge we face today is finding a calibration and a performance validation processes which will systematically get the most of each IMU.As MEMS based IMU start competing with other technologies based on high-end gyroscopes, the calibration becomes a critical topic to increase the IMU performance. A common technique used on high-end technologies as Fiber-optic Gyroscopes (FOG) is to combine a set of non-specific multi-position observations and a systematic calibration method as a Kalman filter [1–4]. However, this method requires ultra-low-noise gyroscopes with excellent bias stability and repeatability to correctly measure Earth rotation rate. These attributes are hardly found on MEMS. The method also limits the observability of non-linearities and cross-axis sensitivity errors because of low dynamics.The calibration method proposed here is based on a direct process [5] combined with high dynamics. High dynamics help discarding noise and bias stability from a proper measurement. Also, a direct approach allows to master all the process and gives the possibility to separate and compensate sensors manufacturing and calibration tools imperfections.Finally, the performance assessment and acceptance test presented in this paper are used to check the consistency of the direct approach technique by applying high dynamics after calibration and measuring sensor errors and triads misalignment in a whole temperature range as shown in Figure 1.