Autocalibration of MEMS Accelerometers
In this chapter, we analyze the critical aspects of the widely diffused calibration and autocalibration procedures for MEMS accelerometers. After providing a review of the main applications of this kind of sensors, we introduce the different sensor models
- PDF / 532,438 Bytes
- 36 Pages / 439.37 x 666.142 pts Page_size
- 68 Downloads / 318 Views
Autocalibration of MEMS Accelerometers Iuri Frosio, Federico Pedersini, and N. Alberto Borghese
Abstract In this chapter, we analyze the critical aspects of the widely diffused calibration and autocalibration procedures for MEMS accelerometers. After providing a review of the main applications of this kind of sensors, we introduce the different sensor models proposed in literature, highlighting the role of the axis misalignments in the sensor sensitivity matrix. We derive a principled noise model and discuss how noise affects the norm of the measured acceleration vector. Since autocalibration procedures are based on the assumption that the norm of the measured acceleration vector, in static condition, equals the gravity acceleration, we introduce the international gravity formula, which provides a reliable estimate of the gravity acceleration as a function of the local latitude and altitude. We derive then the autocalibration procedure in the context of maximum likelihood estimate and we provide examples of calibrations. For each calibrated sensor, we also illustrate how to derive the accuracy on the estimated parameters through the covariance analysis and how to compute the angles between the sensing axes of the sensor. In the conclusion, we summarize the main aspects involved in the autocalibration of MEMS accelerometers.
3.1
Introduction
The advent of microelectromechanical system (MEMS) technology has allowed miniaturized, high performance and cheap accelerometers to be built using a variety of different approaches [1–3]. These sensors were initially used to detect sudden, critical events like in airbag control [4], but nowadays applications cover a wide range of fields, briefly described in the following. A common factor for many applications is that an accurate measurement of the local acceleration is needed, but MEMS sensors
I. Frosio (*) • F. Pedersini • N. Alberto Borghese Computer Science Department, University of Milan, Via Comelico 39/41, Milano 20135, Italy e-mail: [email protected]; [email protected]; [email protected] D. Zhang (ed.), Advanced Mechatronics and MEMS Devices, Microsystems, DOI 10.1007/978-1-4419-9985-6_3, # Springer Science+Business Media New York 2013
53
54
I. Frosio et al.
are only imprecisely (or not at all) calibrated after production; therefore, a practical and precise sensor calibration procedure is necessary to get satisfying accuracy. MEMS accelerometers are used, in static or quasi-static conditions, as tilt sensors [5] and to reconstruct the movements of human body segments [6–10]. A surveillance application is for instance described in [9], where the analysis of the signal measured by an accelerometer on the trunk of the subject permits to classify human activities and posture transitions with an accuracy higher than 90 %. In [6], a wearable device containing several accelerometers and gyroscopes is used to track and analyze the movements of body segments; the sensor calibration permits to limit the effect of the drift: the experimental results
Data Loading...
