• Àüü
  • ÀüÀÚ/Àü±â
  • Åë½Å
  • ÄÄÇ»ÅÍ
´Ý±â

»çÀÌÆ®¸Ê

Loading..

Please wait....

±¹³» ÇÐȸÁö

Ȩ Ȩ > ¿¬±¸¹®Çå > ±¹³» ÇÐȸÁö > µ¥ÀÌÅͺ£À̽º ¿¬±¸È¸Áö(SIGDB)

µ¥ÀÌÅͺ£À̽º ¿¬±¸È¸Áö(SIGDB)

Current Result Document :

ÇѱÛÁ¦¸ñ(Korean Title) È¿À²ÀûÀÎ Ä®¸¸ ÇÊÅ͸µÀ» À§ÇÑ ÃøÁ¤ ³ëÀÌÁî Ãßõ
¿µ¹®Á¦¸ñ(English Title) Measurement Noise Recommendation for Efficient Kalman Filtering
ÀúÀÚ(Author) ¹Ú¼¼ºó   ±æ¸í¼±   ¹®¾ç¼¼   Sebin Park   Myeong-Seon Gil   Yang-Sae Moon  
¿ø¹®¼ö·Ïó(Citation) VOL 34 NO. 01 PP. 0025 ~ 0038 (2018. 04)
Çѱ۳»¿ë
(Korean Abstract)
º» ³í¹®¿¡¼­´Â Ä®¸¸ ÇÊÅ͸µÀÇ ÃøÁ¤ ³ëÀÌÁ ÃßõÇÏ´Â ¹æ¹ýÀ» ´Ù·é´Ù. Ä®¸¸ ÇÊÅ͸µÀº ¿øº» µ¥ÀÌÅÍÀÇ ºÎÁ¤È®ÇÑ °ªÀ» º¸Á¤Çϴµ¥ »ç¿ëÇÏ´Â ´ëÇ¥ÀûÀÎ ÇÊÅ͸µ ¾Ë°í¸®ÁòÀ¸·Î, »ç¿ëÀÚ °æÇè¿¡ ÀÇÇØ ÁÖ¾îÁö´Â ³ëÀÌÁî ÆĶó¹ÌÅÍ¿¡ ÀÇÇØ ¼º´ÉÀÌ ´Þ¶óÁø´Ù. À̶§, °æÇèÀÌ ºÎÁ·ÇÑ »ç¿ëÀÚ°¡ À߸øµÈ ³ëÀÌÁî ÆĶó¹ÌÅ͸¦ Àû¿ëÇÒ °æ¿ì ÀÌ·Î ÀÎÇØ ÇÊÅ͸µ ¼º´ÉÀÌ ÀúÇ쵃 ¼ö ÀÖ´Ù. ÀÌ·¯ÇÑ ¹®Á¦¸¦ ÇØ°áÇϱâ À§ÇØ, º» ³í¹®¿¡¼­´Â ÀÔ·Â µ¥ÀÌÅ͸¦ ºÐ¼®ÇÏ¿© Ä®¸¸ ÇÊÅ͸µÀÇ ÁÖ¿ä ³ëÀÌÁî ÆĶó¹ÌÅÍÀÎ ÃøÁ¤ ³ëÀÌÁ ã¾Æ³»°í, À̸¦ Ä®¸¸ ÇÊÅ͸µ¿¡ Àû¿ëÇÏ´Â ¹æ¹ýÀ» Á¦¾ÈÇÑ´Ù. ´ëºÎºÐÀÇ ¼¾¼­ µ¥ÀÌÅÍ´Â ÃøÁ¤ °úÁ¤¿¡¼­ Æ÷ÇԵǴ ±â°èÀÇ ³ëÀÌÁî, Áï ÃøÁ¤ ³ëÀÌÁ °®´Â´Ù. º» ³í¹®¿¡¼­´Â À̵¿Æò±Õ(moving average) ¹× ¿þÀÌºí¸´(Wavelet) º¯È¯À» »ç¿ëÇÏ¿© ¼¾¼­ µ¥ÀÌÅÍÀÇ ³ëÀÌÁ ºÐ¼®ÇÏ°í, À̸¦ ÃøÁ¤ ³ëÀÌÁî·Î È°¿ëÇÑ´Ù. ½ÇÇè °á°ú, ºÐ¼®À» ÅëÇØ ÃøÁ¤ ³ëÀÌÁ °è»êÇÑ ¹æ¹ýÀÌ º¸´Ù Á¤È®ÇÏ°Ô Ä®¸¸ ÇÊÅ͸µÀ» ¼öÇàÇÏ´Â °ÍÀ¸·Î ³ªÅ¸³µ´Ù.
¿µ¹®³»¿ë
(English Abstract)
In this paper, we propose an analytical method for recommending measurement noise in Kalman filtering. The Kalman filtering is a typical filtering algorithm used to correct inaccurate values of original data, and its filtering performance depends on the user-given noise parameters. At this time, if the wrong noise parameters are used due to lack of user experience, it may degrade the filtering performance. Therefore, in this paper, we propose new methods to analyze the input data and find the measurement noise which is the main noise parameter of the Kalman filtering, and apply it to the Kalman filter. The sensor data includes the noise value of the measuring machine, which is called measurement noise. We use two analytical methods of noise analysis of sensor data: moving average and Wavelet transformations. As a result of experiments, it has been found that the proposed method of calculating the measurement noise more accurately performs the Kalman filtering.
Å°¿öµå(Keyword) Ä®¸¸ ÇÊÅÍ   ³ëÀÌÁî Ãßõ   ¼¾¼­ µ¥ÀÌÅÍ Á¤Á¦   À̵¿Æò±Õ   ¿þÀÌºí¸´ º¯È¯   Kalman filter   noise recommendation   sensor data purification   moving average   Wavelet transform  
ÆÄÀÏ÷ºÎ PDF ´Ù¿î·Îµå