Non-linear Error Modeling for MEMS-based IMUs
| dc.contributor.advisor | El-Sheimy, Naser | |
| dc.contributor.advisor | Sesay, Abu B. | |
| dc.contributor.author | Radi, Ahmed | |
| dc.contributor.committeemember | Nassar, Sameh | |
| dc.contributor.committeemember | Noureldin, Aboelmagd | |
| dc.contributor.committeemember | Ghannouchi, Fadhel M. | |
| dc.contributor.committeemember | Hamad, Ahmed M. | |
| dc.date | 2019-06 | |
| dc.date.accessioned | 2018-12-17T23:05:42Z | |
| dc.date.available | 2018-12-17T23:05:42Z | |
| dc.date.issued | 2018-12-14 | |
| dc.description.abstract | The precise estimation of the position, velocity and orientation of a moving object with and without reception of satellite signals using low-cost sensors has always been a challenging task. Current navigation market is dominated by integrating satellite positioning, such as Global Navigation Satellite System (GNSS) with Inertial Navigation Systems (INSs) through Bayesian filters; e.g. Kalman Filter (KF). During satellite positioning signal outages, navigation information is provided using the inertial sensors, i.e. the gyroscopes and accelerometers of an Inertial Measurement Unit (IMU). Thus, the overall quality of integrated navigation systems is driven by inertial sensors errors. This thesis aims at improving inertial sensor stochastic error modeling to obtain better accuracy, especially in INS stand-alone mode. A common approach to model inertial sensor stochastic errors (sometimes known as stochastic noise) is a 1st order Gauss-Markov (GM) process where its parameters are estimated using the Autocorrelation sequence of the sensor static measurements output collected at room temperature. However, the stand-alone 1st order GM model has proven not to be the best model for several inertial sensors. Consequently. in this thesis different and better noise characterization approaches are proposed, developed and used for analyzing such inertial sensor stochastic noise. The stochastic characteristics of low-cost Micro-Electro Mechanical Systems (MEMS)-based inertial sensor errors and their changes according to temperature and platform dynamics variation using two different approaches, namely Allan Variance (AV) and Generalized Method of Wavelet Moments (GMWM), are investigated. Advantages and limitations of each method concerning the ability to 1) identify the latent random processes associated with the detected error model and 2) accurately estimate the parameters of each random process; are highlighted and used to provide justifications for the developments brought afterword. A new wavelet variance-based framework, as an extension to the standard GMWM, for multi-signal inertial sensor calibration is proposed and developed in this thesis, namely Multi-Signal GMWM (MS-GMWM) that allows to model complex composite stochastic processes. The proposed approach not only can improve the modeling of stochastic sensor errors by using multiple replicates from a calibration procedure but also allows to understand the properties of these stochastic errors to perform more efficient calibration and, consequently, improve the navigation performance. In addition, a Graphical User Interface (GUI) algorithm is developed to make the MS-GMWM available to the general user and to facilitate the calibration procedures of inertial sensor errors using several complex stochastic error models. The KF design accounting for inertial sensor complex stochastic error models is investigated including detailed mathematical explanation of both the prediction and update stages. A novel environmentally-dependent (i.e. taking into account dynamics and temperature changes) adaptive integrated navigation algorithm is developed in this thesis, which is adapted to switch between different stochastic error parameters values according to 1) the inertial sensor temperature and 2) the platform dynamics to limit the overall environmental-dependent effects. The performance of the constructed stochastic error models, when operated through the proposed adaptive integrated algorithm in the designed GUI platform filter presented with optional adaptivity features, is evaluated using field real INS/GNSS data with induced GNSS signal outages. Compared to the traditional 1st order GM model, results showed that considering more complex error models, based on dynamics and thermal data analysis, improves the positioning errors during GNSS signal outages by 32.36 - 51.19%, which shows the significant effect of the proposed algorithms in this thesis. | en_US |
| dc.identifier.citation | Radi, A. (2018). Non-linear Error Modeling for MEMS-based IMUs (Doctoral thesis, University of Calgary, Calgary, Canada). Retrieved from https://prism.ucalgary.ca. | en_US |
| dc.identifier.doi | http://dx.doi.org/10.11575/PRISM/34983 | |
| dc.identifier.uri | http://hdl.handle.net/1880/109363 | |
| dc.language.iso | en | en_US |
| dc.publisher.faculty | Schulich School of Engineering | en_US |
| dc.publisher.institution | University of Calgary | en |
| dc.rights | University of Calgary graduate students retain copyright ownership and moral rights for their thesis. You may use this material in any way that is permitted by the Copyright Act or through licensing that has been assigned to the document. For uses that are not allowable under copyright legislation or licensing, you are required to seek permission. | en_US |
| dc.subject | Stochastic Error Modeling, Inertial Sensors Calibration, Inertial Measurement Units, Allan Variance, Generalized Method of Wavelet Moments, Environmentally Adaptive INS/GNSS Algorithm, MEMS | en_US |
| dc.subject.classification | Engineering | en_US |
| dc.subject.classification | Engineering--Electronics and Electrical | en_US |
| dc.title | Non-linear Error Modeling for MEMS-based IMUs | en_US |
| dc.type | doctoral thesis | en_US |
| thesis.degree.discipline | Interdisciplinary Graduate Program | en_US |
| thesis.degree.grantor | University of Calgary | en_US |
| thesis.degree.name | Doctor of Philosophy (PhD) | en_US |
| ucalgary.item.requestcopy | true |