Á¤º¸°úÇÐȸ ³í¹®Áö B : ¼ÒÇÁÆ®¿þ¾î ¹× ÀÀ¿ë
Current Result Document :
ÇѱÛÁ¦¸ñ(Korean Title) |
Çൿ±â¹Ý ·Îº¿ÀÇ ÆÛÁöÁ¦¾î±â ¼³°è¸¦ À§ÇÑ ÁøÈÇü Á¢±Ù ¹æ½Ä |
¿µ¹®Á¦¸ñ(English Title) |
An Evolutionary Approach to design Fuzzy Controller for Behavior-based Robot |
ÀúÀÚ(Author) |
À̽ÂÀÍ
Á¶¼º¹è
Seungik Lee
Sungbae Cho
|
¿ø¹®¼ö·Ïó(Citation) |
VOL 24 NO. 12 PP. 1400 ~ 1407 (1997. 12) |
Çѱ۳»¿ë (Korean Abstract) |
»óÅ °ø°£ÀÇ Ç¥ÇöÀ» ÀÌ¿ëÇÏ¿© ÇöÀçÀ§Ä¡¸¦ ÀÎÁöÇÏ°í °æ·Î¸¦ °èȹÇÑ ÈÄ ¸ðÅ͸¦ ±¸µ¿½ÃÅ°´Â ÀüÅëÀûÀÎ ÀΰøÁö´ÉÀÇ Á¢±Ù¹æ½ÄÀ¸·Î´Â ²÷ÀÓ¾øÀÌ º¯ÈÇϴ ȯ°æ¿¡ ÀûÀýÈ÷ ´ëóÇÒ ¼ö ÀÖ´Â À̵¿·Îº¿À» ±¸ÃàÇϱ⠾î·Æ´Ù. À̸¦ ÇØ°áÇϱâ À§Çؼ ÃÖ±Ù¿¡ Áøȹæ½ÄÀ¸·Î ·Îº¿ÀÇ Á¦¾î±â¸¦ ÀûÀýÈ÷ ÇнÀ½ÃÅ°´Â Çൿ±â¹Ý ·Îº¿¿¡ ´ëÇÑ ¿¬±¸°¡ È°¹ßÈ÷ ÁøÇàµÇ°í ÀÖ´Ù. º» ³í¹®¿¡¼´Â ÀûÀÀ´É·ÂÀ» °®´Â Á¦¾î±â¸¦ ±¸¼ºÇϱâ À§ÇÏ¿© ±âº» Á¦¾î±â´Â ÆÛÁö ½Ã½ºÅÛÀ¸·Î ±¸¼ºÇÏ°í ³»ºÎÀÇ ¸Å°³º¯¼ö µîÀ» À¯ÀüÀÚ ÃÖÀûÈ ¹æ½Ä¿¡ ÀÇÇؼ °áÁ¤ÇÏ´Â ¹æ¹ýÀ» Á¦¾ÈÇÑ´Ù. ¶ÇÇÑ ÀÌ·¯ÇÑ ¹æ¹ýÀ» Khepera¶ó°í ÇÏ´Â ½ÇÁ¦ À̵¿ ·Îº¿ÀÇ ½Ã¹Ä·¹ÀÌÅÍ¿¡ Àû¿ëÇÏ¿©, ·Îº¿ ½º½º·Î ¸ñÇ¥ ÁöÁ¡±îÁö µµ´ÞÇÒ ¼ö ÀÖ´Â ±ÔÄ¢À» ¹ß°ßÇØ ³¿À» º¸ÀδÙ. ƯÈ÷ ÃÖÁ¾ÀûÀ¸·Î »ý¼ºµÈ ±ÔÄ¢µéÀ» ºÐ¼®ÇÑ °á°ú ¸ñÇ¥ ÁöÁ¡±îÁöÀÇ µµ´ÞÀ̶ó´Â ¹®Á¦¸¦ ÇØ°áÇϱâ À§ÇØ Àå¾Ö¹° ȸÇdzª Á¤ý¿ìȸÀü µî°ú °°Àº ºÎ¹®Á¦ÀÇ ÇØ°á¹æ¹ýµµ ½º½º·Î ¸¸µé¾î ³ÂÀ½À» ¾Ë ¼ö ÀÖ¾ú´Ù.
|
¿µ¹®³»¿ë (English Abstract) |
It is difficult to construct an adaptive controller for mobile robot in the changing environments with a conventional approach, where a robot has to represent the environment by state-space, plan an optimal path to the goal, and then drive its motors appropriately. To solve this problem, some researchers have recently proposed such a method that a robot might learn the optimal control logic automatically through evolutionary mechanism. In this paper, we propose another method of building a basic controller by fuzzy logic and determining its parameters by genetic optimization for an adaptive controller. The simulation with a robot called Khepera shows that it finds an optimal set of fuzzy rules that leads the robot to the goal point. Especially, an analysis of the fuzzy rules indicates that the robot has obtained several substrategies, such as avoiding obstacles and turning right or left, in the course of evolution to find the optimal controller.
|
Å°¿öµå(Keyword) |
|
ÆÄÀÏ÷ºÎ |
PDF ´Ù¿î·Îµå
|