12843

1061 دكتري

دكتري

الكترونيك

اصفهان: دانشگاه صنعتي اصفهان، دانشكده برق و كامپيوتر

1396

شانزده، [143]ص.: مصور، جدول، نمودار

ص.ع. به فارسي و انگليسي

اصغر غلامي

محمدرضا تابان

جعفر قيصري، محمد دانش

1396/07/10

كتابنامه

برق و كامپيوتر

مهندسي برق و كامپيوتر

ID1061 دكتري

4 2 1 روش پيشنهادي تشخيص و حذف داده دورافتاده 68 4 2 1 1 بهينهسازي پارامترها با استفاده از الگوريتم ژنتيك 98 09 4 2 1 2 ساختار الگوريتم ژنتيك 19 4 3 جمعبندي 39 فصل پنجم نتايج آزمونهاي عملي ناوبري شناورهاي زيرسطحي 5 1 نتايج حذف نويز 49 5 2 نتايج كاليبراسيون 89 5 2 1 محاسبه واريانس آلن در آزمون عملي 89 5 2 2 بررسي اثر تخمين باياس حسگرهاي اينرسي 001 5 3 آزمون عملي الگوريتم MAESKF پيشنهادي 101 5 4 آزمون عملي الگوريتم 105 MAEKF 5 5 آزمون عملي الگوريتم VB MAESKF پيشنهادي 801 5 6 نتايج حذف دادههاي دورافتاده 311 5 6 1 بررسي مقاوم بودن الگوريتم VB MAESKF به داده دورافتاده 311 5 6 2 نتايج حذف دادههاي دورافتاده در روش پيشنهادي 611 5 7 نتايج آزمون سازگاري الگوريتمهاي تلفيق 121 5 8 جمعبندي 221 421 فصل ششم نتيجهگيري و پيشنهادها 6 1 خالصهي پژوهش 421 6 2 دستاوردهاي پژوهش 621 6 3 محدوديتهاي پژوهش 721 6 4 پيشنهادهاي آتي 721 مراجع 821 چكيده انگليسي 731 يازده

Decreasing Error of Integrated Navigation System and Auxiliary Sensors for Underwater Vehicles Narjes Davari Email n davari@ec iut ac ir Date of Submission on May 24 2017 Department of Electrical and Computer Engineering Isfahan University of Technology Isfahan 84156 83111 Isfahan Iran Supervisor Asghar Gholami gholami@cc iut ac ir Advisor Mohammad Reza Taban mrtaban@cc iut ac ir Associate Professor of Department of Electrical and Computer Engineering Isfahan University of Technology Professor of Department of Electrical and Computer Engineering Isfahan University of Technology Abstract This study considers a multi rate multi sensor data integration problem in the linear state space model withtime varying and unknown measurement noise The designed navigation system is composed of a high ratestrapdown inertial navigation system along with low rate auxiliary sensors with different sampling rates Theauxiliary sensors consist of a global positioning system a Doppler velocity log a depthmeter and aninclinometer Using sensors with different sampling rates requires the design of multi rate integrationalgorithms To improve the performance of multi rate error state Kalman filter MESKF for marine navigationsystem a multi rate adaptive error state Kalman filter MAESKF and a variational Bayesian approximationbased MAESKF VB MAESKF are proposed Performance of the proposed algorithm is investigated usingreal measurements Results of two experimental tests show that the average relative root mean square error RMSE of the position estimated by VB MAESKF can be decreased approximately 57 and 36 whencompared to that of MESKF and MAESKF algorithms respectively Key Words Strapdown Inertial Navigation System Multirate Integrated Navigation System Adaptive Kalman Filter Introduction In the conventional integrated navigation systems the statistical information of theprocess and measurement noises are considered constant Parameter adjustment for theintegration algorithm is too time consuming and difficult In real applications due tovariation of vehicle dynamics environmental conditions and imperfect knowledge of thefilter statistical information the process and measurement covariance matrices are unknownand time dependent In practical conditions sensors with different sampling rates are usedwhich requires the design of multi rate integration algorithm such as MESKF In thisresearch data integration methods such as adaptive filters which compensate the errorscaused by the variation of covariance matrices are used Adaptive integration algorithms forseveral sensors with different sampling rates based on the innovation sequence and thevariational Bayesian approximation are proposed and implemented To improveperformance of MESKF this research presents MAESKF based on the innovation sequence For MAESKF algorithm computation of innovation sequence vector and covariance matrix and measurement noise covariance matrix for multi rate sensors are modified In innovation based adaptive estimator the adaptation uses new information from the filter innovationsequence to estimate the states and statistical information This procedure is based on thecovariance matching technique amending the noise statistics by consistent innovation

اصغر غلامي

محمدرضا تابان

جعفر قيصري، محمد دانش