°í±ÞÇÊÅ͸µ

> LECTURE >°í±ÞÇÊÅ͸µ

°í±ÞÇÊÅ͸µÀÌ·Ð (Advanced Filtering Theory)

°­Á³»¿ë

¼±Çü½Ã½ºÅÛÀÇ ÃÖÀûÃßÁ¤ÀÚÀÎ Ä®¸¸ÇÊÅ͸¦ ±âº»À¸·Î ÇÏ¿© ¸ðµ¨ÀÇ ºÒÈ®½Ç¼ºÀ» °í·ÁÇÑ ÀûÀÀÇÊÅÍ ¹× °­ÀÎÇÊÅÍ, °è»ê·®À» ÁÙÀ̱â À§ÇÑ ½´¹ÌÆ®ÇÊÅÍ, ´Ù¾çÇÑ ÃøÁ¤Ä¡¿¡ ´ëÇÑ ³»°íÀåÀ» ³ôÀ̱â À§ÇÑ ºÐ»êÇÊÅ͸¦ ¼Ò°³ÇÑ´Ù. ¶ÇÇÑ ºñ¼±Çü ¸ðµ¨¿¡ ´ëÇÑ È®ÀåÄ®¸¸ÇÊÅ͸¦ ±â¹ÝÀ¸·Î ÃÖ±Ù¿¡ ¸¹Àº ¿¬±¸°¡ µÇ°í ÀÖ´Â ¹«Çâ Ä®¸¸ÇÊÅÍ (Unscented Kalman Filter)¿¡ ´ëÇÑ Æ¯¼ºÀ» ºÐ¼®ÇÑ´Ù. ¸¶Áö¸·À¸·Î »óź¯¼ö°¡ Á¤±ÔºÐÆ÷¸¦ °¡ÁöÁö ¾ÊÀ» ¶§ ÃÖÀûÀÇ ÇÊÅÍ·Î ÃÖ±Ù¿¡ °¢±¤À» ¹Þ°íÀÖ´Â ´Ù¾çÇÑ ÀÔÀÚÇÊÅÍ (Particle Filter)¸¦ ¼Ò°³ÇÑ´Ù. ¼ö¾÷¿¡¼­ ¼Ò°³µÈ ÇÊÅÍ´Â INS/GPS °áÇսýºÅÛ°ú ·¹ÀÌ´õ ÃßÀû¹®Á¦¿¡ Àû¿ëÇÑ´Ù.

Based on the linear optimal Kalman filter, in the first part, some additional important topics such as adaptive and robust Kalman filtering to consider the model uncertainties, the Schmidt Kalman filtering to reduce the computational complexity, and decentralized filtering to improve the fault tolerance are discussed. In the second part, nonlinear filters including the extended Kalman filter (EKF), the unscented Kalman filter (UKF), and the particle filters are presented. The UKF is recently introduced to provide significant improvement over the EKF, and the particle filter is a completely nonlinear state estimator. Finally, various filters covered in this course are applied to INS/GPS and tracking problems.

What to cover

Introduction

Review of Kalman Filter

Linear Filter
- Decentralized / Federated Kalman Filters
- Robust Kalman Filters

Nonlinear Filter
- Extended Kalman Filter
- Gaussian Sum Filters
- Uncented Kalman Filters : Unscented Transformation, Scaled/Reduced Sigma Point
- Particle Filters

Applications

Reference

D. Simon, Optimal State Estimation, Wiley, 2006 B. Ristic, S. Arulampalan and N. Gordon, Beyond the Kalman Filter, Artech House, 2004 R. Brown and P. Hwang, Introduction to Random Signals and Applied Kalman Filtering(3rd Ed.), Wiley, 1997

¼­¿ï´ëÇб³ Ç×¹ýÀüÀڽýºÅÛ¿¬±¸½Ç | ÁÖ¼Ò : ¼­¿ïƯº°½Ã °ü¾Ç±¸ °ü¾Ç·Î 1 ¼­¿ï´ëÇб³ | ±³¼ö ¿¬±¸½Ç : 133µ¿ 603È£ (Tel. 02-880-7308)
´ëÇпø»ý ¿¬±¸½Ç : 133µ¿ 607È£ (Tel. 02-880-1732) | ½ÇÇè½Ç : 302µ¿ 418-1È£ | Copyright by NESL All Rights Reserved.