2019³â Ãß°èÇмú´ëȸ
Current Result Document :
ÇѱÛÁ¦¸ñ(Korean Title) |
Wi-Fi ±â¹Ý ½Ç³» À§Ä¡ÃßÁ¤½Ã½ºÅÛ¿¡¼ ¸Ó½Å·¯´× ÇнÀÀ» À§ÇÑ ½ºÆåÅä±×·¥ ¶óµð¿À ¸Ê ¼³°è |
¿µ¹®Á¦¸ñ(English Title) |
Design of Spectogram Radio Map for Machine Learning Training in Wi-Fi based Indoor Localization System |
ÀúÀÚ(Author) |
±èÅ¿Ï
±èÁ¤¼ö
À̵¿¸í
Tae-Wan Kim
Jungsu Kim
Dong Myung Lee
|
¿ø¹®¼ö·Ïó(Citation) |
VOL 23 NO. 02 PP. 0504 ~ 0505 (2019. 10) |
Çѱ۳»¿ë (Korean Abstract) |
º» ³í¹®ÀÇ °³¹ß¸ñÇ¥ÀÎ ¸Ó½Å·¯´× ÇнÀ ±â¹Ý ½Ç³» À§Ä¡ÃßÁ¤½Ã½ºÅÛÀº Wi-Fi ½ÅÈ£ ¼öÁý´Ü°è, ¸Ó½Å·¯´× ÇнÀ´Ü°è ±×¸®°í À§Ä¡ ÃßÁ¤´Ü°èÀÇ 3´Ü°è·Î ±¸¼ºµÈ´Ù. Wi-Fi ±â¹Ý À§Ä¡ÃßÁ¤ ±â¼ú¿¡´Â ÁÖ·Î ¶óµð¿À ¸Ê (radio signal map)À» ÀÌ¿ëÇÑ ÇΰÅÇÁ¸°Æà ¹æ½Ä (fingerprinting)°ú ½ÅÈ£ ¼¼±â¸¦ ÀÌ¿ëÇÑ »ïº¯Ãø·® ¹æ½Ä (trilateration)ÀÌ ÀÖ´Ù. ¸Ó½Å·¯´× ÇнÀÀ» À§ÇÑ ½ºÆåÅä±×·¥ (spectogram) ¶óµð¿À ¸Ê ¼³°è¸¦ À§ÇÏ¿© ¸ÕÀú ƯÁ¤ ½Ç³» ¿µ¿ªÀ» ÂüÁ¶±¸¿ª (reference points, RPs)À¸·Î ³ª´« ÈÄ, °¢ RP¿¡¼ÀÇ Wi-Fi ½ÅÈ£ÀÇ ¼¼±â¸¦ ½ºÆåÅä±×·¥À¸·Î º¯È¯ÇÑ´Ù. ±× ´ÙÀ½, ÀÌ ½ºÆåÅä±×·¥Àº ¸Ó½Å·¯´× ÇнÀÀ» ¼öÇàÇÑ ÈÄ ¶óµð¿À ¸Ê¿¡ ÀúÀåµÈ´Ù. ¼³°èÇÑ ½ºÆåÅä±×·¥ ¶óµð¿À ¸ÊÀº ½Ç³» À§Ä¡ÃßÁ¤½Ã½ºÅÛÀÇ À§Ä¡ÃßÁ¤ Á¤È®µµ Çâ»ó¿¡ ¸¹Àº µµ¿òÀ» ÁÙ °ÍÀ¸·Î ÆǴܵȴÙ. |
¿µ¹®³»¿ë (English Abstract) |
The indoor localization system based on machine learning training, which is the development goal of this paper, consists of three phases: Wi-Fi signal collection, machine learning training, and location estimation. Wi-Fi based localization techniques mainly include the fingerprinting scheme using radio signal map and the trilateration scheme using signal strength. To design a SPECTOGRAM radio map for machine learning training, first a specific indoor area is divided into reference points (RPs), and then convert the strength of the Wi-Fi signal at each RP into a SPECTOGRAM. It is then stored in a radio map after machine learning training. The designed SPECTOGRAM radio map will be helpful to improve the localization accuracy in indoor localization system. |
Å°¿öµå(Keyword) |
Wi-Fi
Indoor Locaization
Machine Learning
Spectogram
Radio Map
Fingerprinting
|
ÆÄÀÏ÷ºÎ |
PDF ´Ù¿î·Îµå
|