• ๋Œ€ํ•œ์ „๊ธฐํ•™ํšŒ
Mobile QR Code QR CODE : The Transactions of the Korean Institute of Electrical Engineers
  • COPE
  • kcse
  • ํ•œ๊ตญ๊ณผํ•™๊ธฐ์ˆ ๋‹จ์ฒด์ด์—ฐํ•ฉํšŒ
  • ํ•œ๊ตญํ•™์ˆ ์ง€์ธ์šฉ์ƒ‰์ธ
  • Scopus
  • crossref
  • orcid

  1. (Dept. of Mechatronics Engineering, Chungnam National University, Korea)



Angular acceleration estimation, Data-driven state observer, Complementary filter

1. ์„œ๋ก 

๊ฐ๊ฐ€์†๋„๋Š” ๋™์  ์‹œ์Šคํ…œ์—์„œ ์š”๊ตฌ๋˜๋Š” โ€œ๊ฐ€๋ณ€ ๊ฐ•์„ฑ์„ ๊ฐ–๋Š” ๊ฐ•๊ฑด์„ฑโ€์„ ๊ตฌํ˜„ํ•  ์ˆ˜ ์žˆ๋Š” ํ•˜๋‚˜์˜ ๋ฐฉ๋ฒ•์ด ๋  ์ˆ˜ ์žˆ๋‹ค๋Š” ๊ฐœ๋…์ด ์†Œ๊ฐœ๋˜์—ˆ๊ณ [1] ๊ฐ๊ฐ€์†๋„ ํ”„๋กœํŒŒ์ผ์„ ์ด์šฉํ•œ ๋จธ์‹  ํˆด ์ œ์–ด ๋ฐฉ๋ฒ•์ด ์ œ์•ˆ๋˜์—ˆ๋‹ค[2]. ๊ฐ๊ฐ€์†๋„ ์ •๋ณด๋Š” ๋‹ค์ถ• ์‹œ์Šคํ…œ์˜ ๋…๋ฆฝ์  ์ œ์–ด์— ํ™œ์šฉ๋  ์ˆ˜ ์žˆ์–ด ์ถ• ๊ฐ„์˜ ์ƒํ˜ธ ์ปคํ”Œ๋ง ํšจ๊ณผ๋ฅผ ๊ณ ๋ คํ•˜์ง€ ์•Š๊ณ  ๋…๋ฆฝ๋œ ์ถ• ์ œ์–ด๋ฅผ ๊ตฌํ˜„ํ•  ์ˆ˜ ์žˆ์œผ๋ฉฐ ๋งค๋‹ˆํ“ฐ๋ ˆ์ดํ„ฐ์˜ ์‹œ๊ฐ„ ์ง€์—ฐ ์ œ์–ด ๊ธฐ๋ฒ•(Time delay control)[3,4]๊ณผ ๊ฐ๊ฐ€์†๋„ ๊ธฐ๋ฐ˜์˜ ์™ธ๋ž€๊ด€์ธก๊ธฐ ๊ฐœ๋…์œผ๋กœ ๋ฐœ์ „ํ•ด์™”๋‹ค[5,6].

์ตœ๊ทผ ์„ ๊ฐ€์†๋„๋ฅผ ์ธก์ •ํ•˜๋Š” AHRS (Attitude and Heading Reference System) ์„ผ์„œ๊ฐ€ ๋„๋ฆฌ ๋ณด๊ธ‰๋˜์—ˆ๋‹ค๋Š” ์ ๊ณผ ์ƒคํ”„ํŠธ ์—”์ฝ”๋” ๊ธฐ๋ฐ˜์˜ ๊ตฌ๋™๊ธฐ๊ฐ€ ์ผ๋ฐ˜ํ™” ๋˜์—ˆ๋‹ค๋Š” ์ ์—์„œ ์ด ๋‘ ์„ผ์„œ๋ฅผ ์ด์šฉํ•˜์—ฌ ๊ตฌ๋™๊ธฐ ์ƒํƒœ ์ถ”์ • ๋ฐฉ๋ฒ•์— ๊ด€ํ•œ ์—ฐ๊ตฌ๊ฐ€ ๋กœ๋ด‡ ๋ถ„์•ผ์—์„œ ํ™œ๋ฐœํ•˜๊ฒŒ ์ง„ํ–‰๋˜๊ณ  ์žˆ๋‹ค. ์„ ๊ฐ€์†๋„ ์„ผ์„œ ๋งŒ์„ ์ด์šฉํ•œ ๊ตฌ๋™๊ธฐ์˜ ๊ฐ๋„ ์ƒํƒœ ์ถ”์ • ๋ฐฉ๋ฒ•์ด ์ œ์•ˆ๋˜์—ˆ๊ณ [7] ์ €๊ฐ€ํ˜• ์„ ๊ฐ๊ฐ€์†๋„ ์„ผ์„œ์™€ ์ž์ด๋กœ์Šค์ฝ”ํ”„ ์„ผ์„œ์˜ ํ“จ์ „์— ์˜ํ•œ ๊ตฌ๋™๊ธฐ ๊ฐ๋„ ์ถ”์ • ๋ฐฉ๋ฒ•์ด ์†Œ๊ฐœ๋˜์—ˆ๊ณ [8] ๋ชจํ„ฐ ์—”์ฝ”๋”์™€ ์—”๋“œ-์ดํŽ™ํ„ฐ ์„ ๊ฐ€์†๋„ ์„ผ์„œ ํ“จ์ „์— ์˜ํ•œ ์กฐ์ธํŠธ ์ƒํƒœ ์ถ”์ • ๋ฐฉ๋ฒ•์ด ์–ธ๊ธ‰๋˜์—ˆ์œผ๋ฉฐ[9] 3๊ฐœ์˜ ๊ด€์„ฑ ์„ผ์„œ๋ฅผ ์ด์šฉํ•œ 6์ถ•์˜ ๊ฐ๋„ ์ƒํƒœ ์ถ”์ • ๋ฐฉ๋ฒ•์ด ์ œ์•ˆ๋˜์—ˆ๋‹ค[10].

๋˜ํ•œ ์ด๋™์ฒด์˜ ๋…๋ฆฝ ํœ  ํ† ํฌ ์ œ์–ด ๋ฐฉ๋ฒ•์— ๊ด€ํ•œ ์—ฐ๊ตฌ๊ฐ€ ํ™œ๋ฐœํžˆ ์ง„ํ–‰๋˜๊ณ  ์žˆ๋‹ค. ์š”-๋ชจ๋ฉ˜ํŠธ๋ฅผ ์ง์ ‘ ์ œ์–ดํ•˜๊ธฐ ์œ„ํ•ด์„œ ๊ฐ์†๋„ ๊ด€์ธก๊ธฐ ๊ธฐ๋ฐ˜์˜ ๋…๋ฆฝ ํœ  ํ† ํฌ ์ œ์–ด ๊ตฌ์กฐ๊ฐ€ ์†Œ๊ฐœ๋˜์—ˆ๊ณ [11] ์ „๋ฅ˜ ๊ตฌ๋™๊ธฐ์˜ ๋…๋ฆฝ์  ํ† ํฌ ์ œ์–ด ๋ถ„๋ฐฐ๊ฐ€ ์ฐจ๋Ÿ‰์˜ ๊ฐ€์†๋„์— ์–ด๋–ค ์˜ํ–ฅ์„ ๋ฏธ์น˜๋Š”์ง€์— ๊ด€ํ•œ ์—ฐ๊ตฌ๊ฐ€ ์–ธ๊ธ‰๋˜์—ˆ๊ณ [12] ํผ์ง€ ์ œ์–ด์— ์˜ํ•œ ํ† ํฌ ๋ถ„์‚ฐ ์ œ์–ด ์•Œ๊ณ ๋ฆฌ๋“ฌ์ด ์ œ์•ˆ๋˜์—ˆ๋‹ค[13]. ์Šฌ๋ฆฝ ์•ˆ์ •ํ™”๋ฅผ ์œ„ํ•œ ์Šฌ๋ผ์ด๋”ฉ ๋ชจ๋“œ ์ œ์–ด ๊ธฐ๋ฐ˜ ํ† ํฌ ๋ถ„์‚ฐ ์ œ์–ด ๋ฐฉ๋ฒ•์ด ์ œ์•ˆ๋˜์—ˆ๊ณ [14] ํœ  ๋…๋ฆฝ์ œ์–ด์— ์žˆ์–ด ๋„คํŠธ์›Œํฌ์— ์˜ํ•œ ์‹œ๊ฐ„ ์ง€์—ฐ ๋ฌธ์ œ๊ฐ€ ์†Œ๊ฐœ๋˜์—ˆ๋‹ค[15].

์œ„์—์„œ ์„ค๋ช…ํ•œ ๊ฒƒ์ฒ˜๋Ÿผ ๋กœ๋ด‡๊ณผ ์ด๋™์ฒด ๋“ฑ์—์„œ ๋…๋ฆฝ๋œ ์กฐ์ธํŠธ ๋˜๋Š” ๋…๋ฆฝ๋œ ํœ ์˜ ํ† ํฌ ์ œ์–ด ๋ฐฉ์‹์— ๋Œ€ํ•œ ์—ฐ๊ตฌ๊ฐ€ ํ™œ๋ฐœํ•˜์ง€๋งŒ ์กฐ์ธํŠธ๋‚˜ ํœ ์˜ ๊ฐ๊ฐ€์†๋„ ์ƒํƒœ ์ถ”์ •์— ๊ด€ํ•œ ์—ฐ๊ตฌ๋Š” ์šฐ๋ฆฌ๊ฐ€ ์•„๋Š” ํ•œ ๋“œ๋ฌธ ๊ฒƒ์œผ๋กœ ํŒ๋‹จ๋œ๋‹ค. ๋ณธ ์—ฐ๊ตฌ์—์„œ๋Š” ์ด ์ ์— ์ฃผ๋ชฉํ•ด์„œ ์กฐ์ธํŠธ๋‚˜ ํœ ์˜ ๊ฐ๊ฐ€์†๋„๋ฅผ ์ถ”์ •ํ•  ์ˆ˜ ์žˆ๋Š” ์ƒํƒœ๊ด€์ธก๊ธฐ๋ฅผ ์ œ์•ˆํ•˜๊ณ ์ž ํ•œ๋‹ค.

๋™์  ์‹œ์Šคํ…œ์˜ ์ƒํƒœ๊ด€์ธก๊ธฐ ์„ค๊ณ„๋ฅผ ์œ„ํ•ด์„œ๋Š” ๋™์ ์œผ๋กœ ๋ณ€ํ•˜๋Š” ๋ณ€์ˆ˜์— ๋Œ€ํ•œ ์ธ์‹์ด ๋ฐ˜๋“œ์‹œ ํ•„์š”ํ•˜๊ณ  ๋™์  ์‹œ์Šคํ…œ์˜ ๋ณ€์ˆ˜ ์ธ์‹์„ ์œ„ํ•œ ์ตœ์†Œ์ž์Šน๋ฒ• ์•Œ๊ณ ๋ฆฌ๋“ฌ์ด ์†Œ๊ฐœ๋˜๊ณ  ์žˆ๋‹ค[16,17]. ๋ณธ ์—ฐ๊ตฌ์—์„œ ๋…๋ฆฝ๋œ ์ถ•์€ ์ž…๋ ฅ ํ† ํฌ์™€ ์ถœ๋ ฅ ์—”์ฝ”๋” ์ •๋ณด์— ๊ธฐ๋ฐ˜ํ•ด์„œ ์žฌ๊ท€์ตœ์†Œ์ž์Šน๋ฒ• (Recursive least squared method)์— ์˜ํ•ด ์ด์ฐจ์‹œ์Šคํ…œ์œผ๋กœ ๋ชจ๋ธ๋ง๋œ๋‹ค[18,19]. ์žฌ๊ท€์ตœ์†Œ์ž์Šน๋ฒ•์— ์˜ํ•ด ์ธ์‹๋œ ์ด์ฐจ์‹œ์Šคํ…œ ๋ชจ๋ธ (RLS ๋ชจ๋ธ)๋กœ ๋ถ€ํ„ฐ ๋ฃจ์—”๋ฒ„๊ฑฐ ์ƒํƒœ๊ด€์ธก๊ธฐ๋ฅผ ์„ค๊ณ„ํ•˜๊ธฐ ์œ„ํ•ด์„œ๋Š” ๊ด€์ธก๊ธฐ ๊ฒŒ์ธ ์„ค๊ณ„๊ฐ€ ํ•„์š”ํ•˜๊ณ  ๊ด€์ธก๊ธฐ ๊ฒŒ์ธ์€ RLS ๋ชจ๋ธ๋กœ๋ถ€ํ„ฐ ์œ ๋„๋˜์–ด์•ผ๋งŒ ํ•œ๋‹ค. ์ด์ฐจ์‹œ์Šคํ…œ์€ ๊ฐ๋„์™€ ๊ฐ์†๋„ ์ƒํƒœ๋ฅผ ํฌํ•จํ•˜๊ณ  ์žˆ์œผ๋ฏ€๋กœ ์ด์ฐจ์‹œ์Šคํ…œ์˜ ์ƒํƒœ์ •๋ณด๋กœ๋ถ€ํ„ฐ ๊ฐ๊ฐ€์†๋„ ์ƒํƒœ๋ฅผ ์ถ”์ •ํ•˜๋Š” ๊ฒƒ์ด ํ•„์š”ํ•˜๋‹ค. ์ด๋ฅผ ์œ„ํ•ด ๋ณธ ๋…ผ๋ฌธ์—์„œ๋Š” ํ™•์žฅ๋œ ์ƒํƒœ ๊ด€์ธก๊ธฐ์— ์˜ํ•ด ๊ฐ๊ฐ€์†๋„๋ฅผ ์ถ”์ •ํ•˜๋Š” ๋ฐฉ๋ฒ•์„ ์ œ์‹œํ•œ๋‹ค.

๋…๋ฆฝ๋œ ์ถ•์˜ ๊ฐ๊ฐ€์†๋„๋ฅผ ์ถ”์ •ํ•˜๋Š” ๋˜ ๋‹ค๋ฅธ ๋ฐฉ๋ฒ•์€ ์œ ํ•œ ์ฐจ๋ถ„ ๋ฐฉ๋ฒ•(finite difference method)์„ ์‚ฌ์šฉํ•˜๋Š” ๋ฐฉ๋ฒ•์ด ์žˆ๋‹ค. ํ•˜์ง€๋งŒ ์œ ํ•œ ์ฐจ๋ถ„ ํ•„ํ„ฐ์˜ ๊ฒฝ์šฐ ์—”์ฝ”๋”์— ์˜ํ•œ ๊ฐ๋„ ๋ณ€์œ„๋Ÿ‰์„ ์ƒ˜ํ”Œ๋ง ์‹œ๊ฐ„์˜ ์ œ๊ณฑ์œผ๋กœ ๋‚˜๋ˆ„์–ด์„œ ๊ตฌํ•˜๋ฏ€๋กœ ์žก์Œ์— ์ทจ์•ฝํ•˜๋‹ค. ๋ฐ˜๋ฉด, ๊ฐ๊ฐ€์†๋„ ์ •๋ณด๋ฅผ ์ถ”์ •ํ•˜๋Š” ๋ฃจ์—”๋ฒ„๊ฑฐ ๊ด€์ธก๊ธฐ (ํ™•์žฅ๋œ ๋ฃจ์—”๋ฒ„๊ฑฐ ๊ด€์ธก๊ธฐ)์— ์˜ํ•ด ์ถ”์ •๋œ ๊ฐ๊ฐ€์†๋„ ์ •๋ณด๋Š” ๋น ๋ฅธ ์ƒ˜ํ”Œ๋ง ์‹œ๊ฐ„์ด ์š”๊ตฌ๋œ๋‹ค. ๋Š๋ฆฐ ์ƒ˜ํ”Œ๋ง ์‹œ๊ฐ„์—์„œ RLS ๋ชจ๋ธ์€ ์ดˆ๊ธฐ ์ตœ๊ณ ๊ฐ’ ํšจ๊ณผ๋ฅผ ๊ฐ–๊ฒŒ ๋˜๊ณ  ์ดˆ๊ธฐ ์ตœ๊ณ ๊ฐ’ ํšจ๊ณผ๋Š” RLS ์‹œ์Šคํ…œ ์ธ์‹์ด ์ดˆ๊ธฐ ๊ณผ์ •์—์„œ ์ถฉ๋ถ„ํ•œ ๋ฐ์ดํ„ฐ๋ฅผ ๋ณด์œ ํ•˜์ง€ ๋ชปํ•˜๊ธฐ ๋•Œ๋ฌธ์— ์ผ์ • ์‹œ๊ฐ„ ๋™์•ˆ ์ถ”์ • ๊ฐ๋„๊ฐ€ ๋“œ๋ฆฌํ”„ํŠธ ๋˜๋Š” ๋ฌธ์ œ๋ฅผ ๋ฐœ์ƒ์‹œํ‚ค๊ณ  ์ด ํ˜„์ƒ์€ ๊ด€์ธก๊ธฐ ์ถ”์ • ๊ฐ๊ฐ€์†๋„์— ๋ฐ”์ด์–ด์Šค๋กœ ์ž‘๋™ํ•˜๊ฒŒ ๋˜๊ณ  ์ƒํƒœ ์ถ”์ •์˜ ์ •ํ™•์„ฑ์ด ๋–จ์–ด์ง€๋Š” ๋ฌธ์ œ๊ฐ€ ๋ฐœ์ƒ๋œ๋‹ค.

๋ณธ ๋…ผ๋ฌธ์—์„œ๋Š” ์œ ํ•œ ์ฐจ๋ถ„๋ฒ•์˜ ๊ฐ•์ ๊ณผ ํ™•์žฅ๋œ ๋ฃจ์—”๋ฒ„๊ฑฐ ๊ด€์ธก๊ธฐ์˜ ๊ฐ•์ ์„ ์„œ๋กœ ๋ณด์™„ํ•˜๋Š” ์ƒ๋ณด ํ•„ํ„ฐ๋ง ๋ฐฉ๋ฒ•์„ ์ œ์•ˆํ•œ๋‹ค. ์œ ํ•œ์ฐจ๋ถ„๋ฒ•์— ์˜ํ•œ ์ถ”์ • ๊ฐ๊ฐ€์†๋„์— ๋Œ€ํ•ด์„œ๋Š” ์ €์—ญ ํ†ต๊ณผ ํ•„ํ„ฐ๋ฅผ ์ ์šฉํ•˜๊ณ  ๊ด€์ธก๊ธฐ์— ์˜ํ•ด ์ถ”์ •๋œ ๊ฐ๊ฐ€์†๋„ ์ •๋ณด์—๋Š” ๊ณ ์—ญ ํ†ต๊ณผ ํ•„ํ„ฐ๋ฅผ ์ ์šฉํ•œ๋‹ค. ์‹œ๋ฎฌ๋ ˆ์ด์…˜ ๊ฒฐ๊ณผ๋ฅผ ๋ฐ”ํƒ•์œผ๋กœ ์ ์ ˆํ•œ ์ฐจ๋‹จ์ฃผํŒŒ์ˆ˜๋ฅผ ์‚ฌ์šฉํ•˜๊ฒŒ ๋˜๊ณ  ์ œ์‹œํ•˜๋Š” ์ƒ๋ณด ํ•„ํ„ฐ๋ง ๋ฐฉ๋ฒ•์„ ํ†ตํ•ด RLS ์ดˆ๊ธฐ ์ตœ๋Œ€ ํšจ๊ณผ์™€ ์œ ํ•œ์ฐจ๋ถ„ ํ•„ํ„ฐ๋ง์˜ ๊ณ ์กฐํŒŒ ๋ฆฌํ”Œ ๋ฌธ์ œ ๋“ฑ์ด ํ•จ๊ป˜ ํ•ด์†Œ๋  ์ˆ˜ ์žˆ๋‹ค๋Š” ์ ์„ ๋กœ๋ด‡ํŒ”์„ ๋Œ€์ƒ์œผ๋กœ ํ•œ ์‹คํ—˜ ๊ฒฐ๊ณผ๋ฅผ ํ†ตํ•ด ์ œ์‹œํ•˜๊ณ ์ž ํ•œ๋‹ค.

2. ๋ฌธ์ œ ์ •์˜

2.1 ๋ฌธ์ œ ์ •์˜ I

์ž„์˜์˜ ๋™์  ์‹œ์Šคํ…œ์—์„œ ๊ตฌ๋™์ถ•์˜ ๊ฐ๊ฐ€์†๋„ ์ •๋ณด๋Š” ์ด๋™์ฒด ์‹œ์Šคํ…œ์˜ ์„ ๊ฐ€์†๋„ ์ •๋ณด๋ฅผ ์•Œ์•„๋‚ด๋Š”๋ฐ ํ•„์š”ํ•˜๊ฒŒ ๋œ๋‹ค. ๊ทธ๋ฆผ. 1(a)์—์„œ์™€ ๊ฐ™์€ ๋…๋ฆฝ ๊ตฌ๋™๊ธฐ ๊ตฌ์กฐ์˜ ๋™์  ์‹œ์Šคํ…œ์„ ๊ฐ€์ •ํ•œ๋‹ค. ๊ทธ๋ฆผ. 1(a)์—์„œ $\vec { \tau } _ { A } [ n ]$๋Š” ์ด์‚ฐ์‹œ๊ฐ„์—์„œ ๊ตฌ๋™๊ธฐ ์ž…๋ ฅ ํ† ํฌ์ด๊ณ  $\vec { a } _ { A } [ n ]$๋Š” ์ด์‚ฐ์‹œ๊ฐ„์—์„œ ๊ตฌ๋™๊ธฐ ๊ฐ๊ฐ€์†๋„ ์ƒํƒœ์ด๊ณ  $\vec { f } _ { L } [ n ]$์€ ์ด์‚ฐ์‹œ๊ฐ„์—์„œ ๊ตฌ๋™๊ธฐ์— ์˜ํ•ด ๋ฐœ์ƒ๋˜๋Š” ํž˜์ด๊ณ  $\vec { a } _ { L } [ n ]$์€ ์ด์‚ฐ์‹œ๊ฐ„์—์„œ ๊ตฌ๋™๊ธฐ์— ์˜ํ•ด ๋ฐœ์ƒ๋œ ์‹œ์Šคํ…œ์˜ ์„ ๊ฐ์†๋„์ด๊ณ  $\vec { f } _ { f } [ n ]$๋Š” ์ด์‚ฐ์‹œ๊ฐ„์—์„œ ์ง€๋ฉด์— ์˜ํ•œ ๋งˆ์ฐฐ๋ ฅ์ด๊ณ  $\vec { f } _ { e } [ n ]$๋Š” ์ด์‚ฐ์‹œ๊ฐ„์—์„œ ์™ธ๋ถ€ ํ™˜๊ฒฝ์— ์˜ํ•ด ๋ฐœ์ƒ๋œ ํž˜์ด๊ณ  $\vec { f } _ { d } [ n ]$๋Š” ์ด์‚ฐ์‹œ๊ฐ„์—์„œ ์ „์ฒด ์™ธ๋ž€์ด๊ณ  ๊ทธ๋ฆผ. 1(b)์—์„œ $\vec { r } [ n ]$๋Š” ํ† ํฌ ๋ฐœ์ƒ์ ์œผ๋กœ๋ถ€ํ„ฐ ํž˜ ๋ฐœ์ƒ์ ๊นŒ์ง€์˜ ๋ณ€์œ„ ๋ฒกํ„ฐ์ด๋‹ค. ๊ทธ๋ฆผ. 1(a)์—์„œ ์ง€๋ฉด์˜ ๋งˆ์ฐฐ๋ ฅ์„ ๋ฌด์‹œํ•  ๊ฒฝ์šฐ ๊ทธ๋ฆผ. 1(b)์™€ ๊ฐ™์€ ๊ฐ„๋‹จํ•œ ๋ชจ๋ธ๋ง์ด ๊ฐ€๋Šฅํ•˜๊ฒŒ ๋œ๋‹ค.

๊ทธ๋ฆผ. 1. ์ž„์˜ ๋™์  ์‹œ์Šคํ…œ์˜ ๊ฐ„๋‹จ ๋ชจ๋ธ

Fig. 1. A simple model of Dynamical systems

../../Resources/kiee/KIEE.2019.68.2.342/fig1.png

๊ทธ๋ฆผ. 1(a)์˜ ์‹œ์Šคํ…œ ๋™์—ญํ•™์€ ๋‹ค์Œ๊ณผ ๊ฐ™๋‹ค.

(1a)
$\vec { \tau } _ { A } [ n ] = I [ n ] \vec { a } _ { A } [ n ]$

(1b)
$\vec { f } _ { L } [ n ] + \vec { f } _ { d } [ n ] = \vec { M a } _ { L } [ n ]$

(1a)์€ ๋…๋ฆฝ ํœ  ๊ตฌ๋™๊ธฐ์˜ ๋™์—ญํ•™์ด๊ณ  ์—ฌ๊ธฐ์„œ $I [ n ]$๋Š” ์ด์‚ฐ์‹œ๊ฐ„์—์„œ ๊ด€์„ฑ ์งˆ๋Ÿ‰์ด๋‹ค. (1b)๋Š” ์‹œ์Šคํ…œ์˜ ๋™์—ญํ•™์ด๊ณ  ์—ฌ๊ธฐ์„œ $M$์€ ์‹œ์Šคํ…œ์˜ ์งˆ๋Ÿ‰์ด๋‹ค. $\vec { a } _ { A } [ n ]$๊ณผ $\vec { a } _ { L } [ n ]$ ์‚ฌ์ด์—๋Š” ๋‹ค์Œ์˜ ๊ด€๊ณ„์‹์ด ์กด์žฌํ•œ๋‹ค.

(2)
$\vec { a } _ { L } [ n ] = \vec { a } _ { A } [ n ] \times \vec { r } [ n ] + \vec { v } _ { A } [ n ] \times \left( \vec { v } _ { A } [ n ] \times \vec { r } [ n ] \right)$

๋งŒ์•ฝ $\vec { a } _ { A } [ n ]$์™€ $\vec { v } _ { A } [ n ]$๋ฅผ ์•Œ๊ณ  ์žˆ์„ ๊ฒฝ์šฐ (2) ์‹์— ์˜ํ•ด $\vec { a } _ { L } [ n ]$์„ ์•Œ ์ˆ˜๊ฐ€ ์žˆ๋‹ค. (1a)์™€ (1b)์˜ ํ† ํฌ-ํž˜ ๊ด€๊ณ„๋Š” $\vec { r } \neq 0$ ์กฐ๊ฑด์—์„œ $\vec { \tau } _ { A } [ n ] = \vec { r } \times \vec { f } _ { L } [ n ]$๋ฅผ ๋”ฐ๋ฅด๊ฒŒ ๋œ๋‹ค. ๋”ฐ๋ผ์„œ (1b)์˜ ์™ธ๋ž€์„ ๋‹ค์Œ๊ณผ ๊ฐ™์ด ์ถ”์ •ํ•  ์ˆ˜๊ฐ€ ์žˆ๋‹ค.

(3)
$\vec { f } _ { d } [ n ] = \vec { M a _ { L } } [ n ] - \vec { f } _ { L } [ n ]$

1์ถ•์—์„œ์˜ ๊ฐ„๋‹จํ•œ ๋ฌผ๋ฆฌ์  ๊ด€๊ณ„๋“ค์€ ๋‹ค์ถ•์œผ๋กœ ํ™•์žฅ ์ ์šฉ๋  ์ˆ˜ ์žˆ์œผ๋ฉฐ ๋งŒ์•ฝ ๋‹ค์ถ•์˜ ๊ด€์„ฑ ์งˆ๋Ÿ‰ $I [ n ]$๊ฐ€ ์„œ๋กœ ๋…๋ฆฝ์ ์ธ ๊ฐ’์œผ๋กœ ์ •์˜๋  ๊ฒฝ์šฐ ๊ฐ ๊ฐ์˜ ๋ชจํ„ฐ์ถ•์€ ๋…๋ฆฝ์ ์ธ ์ œ์–ด ๊ตฌ์กฐ๋ฅผ ๊ฐ–๋Š”๋‹ค.

๋”ฐ๋ผ์„œ ๊ฐ ์ถ•์˜ ๊ฐ๊ฐ€์†๋„ ์ •๋ณด๋ฅผ ํš๋“ํ•˜๋Š” ๊ฒƒ์€ ๋™์  ์‹œ์Šคํ…œ์„ ์‰ฝ๊ฒŒ ์ œ์–ดํ•˜๊ธฐ ์œ„ํ•ด ๋ฐ˜๋“œ์‹œ ํ•„์š”ํ•œ ์ ˆ์ฐจ์ด๋‹ค. ๋ณธ ์—ฐ๊ตฌ๋Š” ์ผ๋ฐ˜์ ์ธ ๋ชจํ„ฐ์ถ•์ด ์—”์ฝ”๋”๋ฅผ ์žฅ์ฐฉํ•˜๊ณ  ์žˆ๋‹ค๊ณ  ๊ฐ€์ •ํ•˜๊ณ  ๊ฐ๊ฐ€์†๋„๋ฅผ ์ถ”์ •ํ•˜๋Š” ๋ฐฉ๋ฒ•์„ ์ œ์‹œํ•˜๊ณ ์ž ํ•œ๋‹ค.

2.2 ๋ฌธ์ œ ์ •์˜ II

์‹ค์‹œ๊ฐ„ RLS ๋ฐฉ๋ฒ•์€ ๋™์  ์‹œ์Šคํ…œ์˜ ๋ณ€์ˆ˜ ์ธ์‹์„ ์œ„ํ•ด ๋„๋ฆฌ ์‚ฌ์šฉ๋˜์–ด์ง€๊ณ  ์žˆ๋‹ค. RLS ์•Œ๊ณ ๋ฆฌ๋“ฌ์€ ์žฌ๊ท€์  ๋ฐฉ๋ฒ•์„ ์‚ฌ์šฉํ•˜๊ธฐ ๋•Œ๋ฌธ์— ์‹คํ–‰ ์‹œ๊ฐ„์ด ๋น ๋ฅด๋‹ค๋Š” ์žฅ์ ์„ ๊ฐ–๊ณ  ์žˆ๋‹ค. ํ•˜์ง€๋งŒ RLS ์•Œ๊ณ ๋ฆฌ๋“ฌ์€ ์ดˆ๊ธฐ ํ”ผํ‚น ํšจ๊ณผ์™€ ๊ฐ™์€ ๋‹จ์ ๋„ ํ•จ๊ป˜ ๊ฐ–๊ณ  ์žˆ์–ด ์ดˆ๊ธฐ์— ๋ณ€์ˆ˜ ์ถ”์ • ์˜ค์ฐจ๊ฐ€ ํฌ๊ฒŒ ๋ฐœ์ƒํ•˜๊ฒŒ ๋œ๋‹ค. RLS ์•Œ๊ณ ๋ฆฌ๋“ฌ์ด ์ •์ƒ์ƒํƒœ๋กœ ์ž‘๋™ํ•˜๊ธฐ ์œ„ํ•ด์„œ๋Š” ์ถฉ๋ถ„ํ•œ ๋ฐ์ดํ„ฐ๊ฐ€ ํ•„์š”ํ•˜๊ธฐ ๋•Œ๋ฌธ์ด๋‹ค. ์ด ๋ฌธ์ œ์ ์œผ๋กœ ์ธํ•ด RLS ๋ชจ๋ธ์— ๊ธฐ๋ฐ˜ํ•œ ๊ฐ๊ฐ€์†๋„ ์ƒํƒœ๊ด€์ธก๊ธฐ๋Š” ๋™์ผํ•œ ๋ฌธ์ œ๋ฅผ ํฌํ•จํ•˜๊ฒŒ ๋œ๋‹ค. ๋”ฐ๋ผ์„œ ๋ณธ ์—ฐ๊ตฌ์—์„œ๋Š” RLS ๋ชจ๋ธ ๊ธฐ๋ฐ˜ ๊ฐ๊ฐ€์†๋„ ์ถ”์ • ๊ด€์ธก๊ธฐ๋ฅผ ์„ค๊ณ„ํ•  ๋•Œ ๋ฐœ์ƒํ•  ์ˆ˜ ์žˆ๋Š” ์ดˆ๊ธฐ ํ”ผํ‚น ํšจ๊ณผ๋ฅผ ์–ต์ œํ•˜๋Š” ํ•„ํ„ฐ๋ง ๋ฐฉ๋ฒ•์„ ํ•จ๊ป˜ ์ œ์•ˆํ•˜๊ณ ์ž ํ•œ๋‹ค.

3. RLS ์‹œ์Šคํ…œ ์ธ์‹

3.1 RLS ์‹œ์Šคํ…œ ์ธ์‹

๊ตฌ๋™์ถ• ์ž…๋ ฅ ํ† ํฌ๋Š” ๋‹ค์Œ๊ณผ ๊ฐ™์ด ์ •์˜๋œ๋‹ค.

(4)
$\tau _ { A } [ n ] = k i [ n ]$

์—ฌ๊ธฐ์„œ $k$๋Š” ํ† ํฌ ์ƒ์ˆ˜์ด๊ณ  $I [ n ]$๋Š” ์ด์‚ฐ์‹œ๊ฐ„์—์„œ ์ž…๋ ฅ ์ „๋ฅ˜์ด๋‹ค.

์—”์ฝ”๋”์— ์˜ํ•ด ์ธก์ •๋œ ์ถœ๋ ฅ ๊ฐ๋„๋ฅผ $y [ n ]$ ์ด๋ผ ์ •์˜ํ•˜๊ณ  RLS ์•Œ๊ณ ๋ฆฌ๋“ฌ์„ ํ†ตํ•ด ์ถ”์ •๋œ ์ถœ๋ ฅ ๊ฐ๋„๋ฅผ $\hat { y } [ n ]$ ์ด๋ผ ์ •์˜ํ•  ๊ฒฝ์šฐ ์ถ”์ • ์˜ค์ฐจ๋Š” ๋‹ค์Œ๊ณผ ๊ฐ™์ด ์“ธ ์ˆ˜ ์žˆ๋‹ค.

(5)
$e _ { y } [ n ] = y [ n ] - \hat { y } [ n ]$

RLS ์•Œ๊ณ ๋ฆฌ๋“ฌ์— ์˜ํ•œ ์ž…๋ ฅ ๋ฒกํ„ฐ๋Š” ๋‹ค์Œ๊ณผ ๊ฐ™๋‹ค.

(6)
$r [ n ] = [ k i [ n ] k i [ n - 1 ] k i [ n - 2 ] - y [ n ] - y [ n - 1 ] ] ^ { T }$

RLS ์•Œ๊ณ ๋ฆฌ๋“ฌ์— ์˜ํ•œ ๋ณ€์ˆ˜ ๋ฒกํ„ฐ๋Š” ๋‹ค์Œ๊ณผ ๊ฐ™๋‹ค.

(7)
$\boldsymbol { \alpha } [ n ] = \left[ \alpha _ { 1 } [ n ] \alpha _ { 2 } [ n ] \alpha _ { 3 } [ n ] \alpha _ { 4 } [ n ] \alpha _ { 5 } [ n ] \right] ^ { T }$

$\hat { y } [ n ]$์€ ๋‹ค์Œ๊ณผ ๊ฐ™๋‹ค.

(8)
$\hat { y } [ n ] = r [ n ] ^ { T } \boldsymbol { \alpha } [ n ]$

(5)๋Š” ๋‹ค์Œ๊ณผ ๊ฐ™๋‹ค.

(9)
$e _ { y } [ n ] = y [ n ] - r [ n ] ^ { T } \boldsymbol { \alpha } [ n ]$

RLS ์•Œ๊ณ ๋ฆฌ๋“ฌ์€ ๋น„์šฉํ•จ์ˆ˜ $\frac { 1 } { 2 } \sum _ { n = 1 } ^ { N } e _ { y } [ n ] ^ { 2 }$๋ฅผ ์ตœ์†Œํ™”ํ•˜๋Š” ์•Œ๊ณ ๋ฆฌ๋“ฌ์ด๊ณ  $\frac { d \frac { 1 } { 2 } \sum _ { n = 1 } ^ { N } e _ { y } [ n ] ^ { 2 } } { d \boldsymbol { \alpha } [ n ] } = 0$์„ ๋งŒ์กฑํ•˜๋Š” $\alpha [ n ]$์„ ์ถ”์ •ํ•˜๋Š” ๋ฐฉ์‹์ด๋‹ค. ๋”ฐ๋ผ์„œ RLS์— ์˜ํ•ด ์ธ์‹๋œ ์ด์ฐจ์‹œ์Šคํ…œ ๋ชจ๋ธ์˜ ์ „๋‹ฌํ•จ์ˆ˜ $\frac { \hat { Y } [ z ] } { T _ { A } [ z ] }$๋Š” (6)~(8)๋กœ๋ถ€ํ„ฐ ๋‹ค์Œ๊ณผ ๊ฐ™์ด ์“ธ ์ˆ˜ ์žˆ๋‹ค.

(10)
$\frac { \hat { Y } [ z ] } { T _ { A } [ z ] } = \frac { \alpha _ { 1 } + \alpha _ { 2 } z ^ { - 1 } + \alpha _ { 3 } z ^ { - 2 } } { 1 + \alpha _ { 4 } z ^ { - 1 } + \alpha _ { 5 } z ^ { - 2 } }$

์—ฌ๊ธฐ์„œ $\hat { Y } [ z ]$๋Š” $\hat { y } [ n ]$์— ๋Œ€ํ•œ z-๋ณ€ํ™˜ ๊ฐ’์ด๊ณ  $T _ { A } [ Z ]$๋Š” $\tau _ { A } [ n ]$์˜ z-๋ณ€ํ™˜ ๊ฐ’์ด๋‹ค. ๋ณธ ๋…ผ๋ฌธ์—์„œ๋Š” (10)์„ RLS ๋ชจ๋ธ์ด๋ผ๊ณ  ์ •์˜ํ•œ๋‹ค.

3.2 ์ด์ฐจ์‹œ์Šคํ…œ ์ƒํƒœ๊ด€์ธก๊ธฐ ๊ฒŒ์ธ ์„ค์ •

์ƒํƒœ๊ด€์ธก๊ธฐ๋Š” ๋‹ค์Œ๊ณผ ๊ฐ™์ด ์ •์˜๋  ์ˆ˜ ์žˆ๋‹ค.

(11)
$\boldsymbol { X } [ n + 1 ] = A [ n ] \boldsymbol { X } [ n ] + \boldsymbol { B } [ n ] u [ n ] + \boldsymbol { z } [ n ]$ $y [ n ] = X _ { 1 } [ n ]$

์—ฌ๊ธฐ์„œ $\boldsymbol { X } [ n ] = \left[ X _ { 1 } [ n ] X _ { 2 } [ n ] \right] ^ { T }$์ด๊ณ  $A [ n ] = \left[ \begin{array} { c c } { 0 } & { 1 } \\ { - \alpha _ { 5 } } & { - \alpha _ { 4 } } \end{array} \right]$์ด๊ณ  $B [ n ] = [ 0 \quad 1 ] ^ { T }$์ด๊ณ  $u [ n ] = \alpha _ { 1 } k i [ n ] + \alpha _ { 2 } k i [ n - 1 ] + \alpha _ { 3 } k i [ n - 2 ]$์ด๋‹ค.

$z [ n ] = \left[ z _ { 1 } [ n ] z _ { 2 } [ n ] \right] ^ { T }$๋Š” ์ƒํƒœ ๊ด€์ธก๊ธฐ ๋ณด์ • ๋ฒกํ„ฐ์ด๊ณ  ๋ฃจ์—”๋ฒ„๊ฑฐ ๊ด€์ธก๊ธฐ ์„ค๊ณ„๋ฅผ ์œ„ํ•ด์„œ๋Š” ๋‹ค์Œ๊ณผ ๊ฐ™์ด ์„ค์ •๋œ๋‹ค.

(12)
$\begin{aligned} \boldsymbol { z } [ n ] & = \boldsymbol { L } [ n ] e _ { y } [ n ] \\ \boldsymbol { L } [ n ] & = \left[ L _ { 1 } [ n ] L _ { 2 } [ n ] \right] ^ { T } \end{aligned}$

์—ฌ๊ธฐ์„œ $L [ n ]$์€ ๋ฃจ์—”๋ฒ„๊ฑฐ ๊ด€์ธก๊ธฐ ๊ฒŒ์ธ ๋ฒกํ„ฐ์ด๊ณ  ๊ทน์  ์žฌ๋ฐฐ์น˜ ๋ฐฉ๋ฒ•์„ ํ†ตํ•ด์„œ ์„ค์ •๋  ์ˆ˜ ์žˆ๋‹ค. ํ•˜์ง€๋งŒ ๊ทน์  ์žฌ๋ฐฐ์น˜ ๋ฐฉ๋ฒ•์€ ์—ฐ์†์‹œ๊ฐ„ ์˜์—ญ์—์„œ ์ •์˜๋˜์–ด์žˆ๊ธฐ ๋•Œ๋ฌธ์— ์ด์‚ฐ์‹œ๊ฐ„์—์„œ์˜ ๊ทน์  ์žฌ๋ฐฐ์น˜ ๋ฐฉ๋ฒ•์ด ํ•„์š”ํ•˜๋‹ค. ์ œ์•ˆํ•˜๋Š” ์ด์‚ฐ์‹œ๊ฐ„ ๊ทน์  ์žฌ๋ฐฐ์น˜ ๋ฐฉ๋ฒ•์€ ๋‹ค์Œ๊ณผ ๊ฐ™๋‹ค.

------------------------------------------------------------------------------------------------------------------------------------------------------------------

Remark I: RLS model์„ ์œ„ํ•œ ๊ทน์  ๋ฐฐ์น˜ ๋ฐฉ๋ฒ•

(a) ์ด์‚ฐ์˜์—ญ์—์„œ,

(i)
$\frac { \hat { Y } [ z ] } { T _ { A } [ z ] } = \frac { \alpha _ { 1 } + \alpha _ { 2 } z ^ { - 1 } + \alpha _ { 3 } z ^ { - 2 } } { 1 + \alpha _ { 4 } z ^ { - 1 } + \alpha _ { 5 } z ^ { - 2 } }$

(b) (ii)๋ฅผ ์‚ฌ์šฉํ•˜์—ฌ (i)๋ฅผ (iii)์œผ๋กœ ๋ฐ”๊พผ๋‹ค.

(ii)
$z = e ^ { s T } \approx \frac { 1 + s T _ { s } / 2 } { 1 - s T _ { s } / 2 } = \frac { 1 + s T } { 1 - s T } , T = \frac { T _ { s } } { 2 }$

(iii)
$\frac { \hat { Y } ( s ) } { T _ { A } ( s ) } = \frac { A _ { 1 } s ^ { 2 } + A _ { 2 } s + A _ { 3 } } { s ^ { 2 } + A _ { 4 } s + A _ { 5 } }$

$A _ { 1 } = \frac { \alpha _ { 1 } T ^ { 2 } - \alpha _ { 2 } T ^ { 2 } + \alpha _ { 3 } T ^ { 2 } } { T ^ { 2 } - \alpha _ { 4 } T ^ { 2 } + \alpha _ { 5 } T ^ { 2 } }$ ,

$A _ { 2 } = \frac { 2 \alpha _ { 1 } T ^ { 2 } - 2 \alpha _ { 3 } T } { T ^ { 2 } - \alpha _ { 4 } T ^ { 2 } + \alpha _ { 5 } T ^ { 2 } }$ ,

$A _ { 3 } = \frac { \alpha _ { 1 } + \alpha _ { 2 } + \alpha _ { 3 } } { T ^ { 2 } - \alpha _ { 4 } T ^ { 2 } + \alpha _ { 5 } T ^ { 2 } }$ ,

$A _ { 4 } = \frac { 2 T - 2 \alpha _ { 5 } T } { T ^ { 2 } - \alpha _ { 4 } T ^ { 2 } + \alpha _ { 5 } T ^ { 2 } }$ ,

$A _ { 5 } = \frac { 1 + \alpha _ { 4 } T ^ { 2 } + \alpha _ { 5 } } { T ^ { 2 } - \alpha _ { 4 } T ^ { 2 } + \alpha _ { 5 } T ^ { 2 } }$

(c) (iii)์˜ ๊ทน์ ์ด K๋ฐฐ์˜ ๊ทน์ ์„ ๊ฐ–๋Š”๋‹ค๋ฉด, (iii)์€ (iv)๋กœ ํ‘œํ˜„๋  ์ˆ˜ ์žˆ๋‹ค.

(iv)
$\frac { \hat { Y } ( s ) } { T _ { A } ( s ) } = \frac { A _ { 1 } s ^ { 2 } + A _ { 2 } s + A _ { 3 } } { s ^ { 2 } + K A _ { 4 } s + K ^ { 2 } A _ { 5 } }$

(d) (v)๋ฅผ ์ด์šฉํ•˜์—ฌ (iv)๋ฅผ (vi)๋กœ ๋ณ€ํ™˜ํ•œ๋‹ค.

(v)
$s = \frac { 2 } { T _ { s } } \frac { z - 1 } { z + 1 } = T _ { z } \frac { z - 1 } { z + 1 } , \quad \quad T _ { z } = \frac { 2 } { T _ { s } }$ $\quad \quad $ $\left. \frac { \hat { Y } [ z ] } { T _ { A } [ z ] } \right| _ { o b s } = K ^ { 2 } \frac { \beta _ { 1 } + \beta _ { 2 } z ^ { - 1 } + \beta _ { 3 } z ^ { - 2 } } { 1 + \beta _ { 4 } z ^ { - 1 } + \beta _ { 5 } z ^ { - 2 } }$

์—ฌ๊ธฐ์„œ

(vi)
$\beta _ { 1 } = \frac { 4 \alpha _ { 1 } } { \left( 1 - \alpha _ { 4 } + \alpha _ { 5 } \right) + \left( 2 - 2 \alpha _ { 5 } \right) K + \left( 1 + \alpha _ { 4 } + \alpha _ { 5 } \right) K ^ { 2 } }$ ,

$\beta _ { 2 } = \frac { 4 \alpha _ { 2 } } { \left( 1 - \alpha _ { 4 } + \alpha _ { 5 } \right) + \left( 2 - 2 \alpha _ { 5 } \right) K + \left( 1 + \alpha _ { 4 } + \alpha _ { 5 } \right) K ^ { 2 } }$ ,

$\beta _ { 3 } = \frac { 4 \alpha _ { 3 } } { \left( 1 - \alpha _ { 4 } + \alpha _ { 5 } \right) + \left( 2 - 2 \alpha _ { 5 } \right) K + \left( 1 + \alpha _ { 4 } + \alpha _ { 5 } \right) K ^ { 2 } }$ ,

$\beta _ { 4 } = \frac { \left( - 2 + 2 \alpha _ { 4 } - 2 \alpha _ { 5 } \right) + \left( 2 + 2 \alpha _ { 4 } + 2 \alpha _ { 5 } \right) K ^ { 2 } } { \left( 1 - \alpha _ { 4 } + \alpha _ { 5 } \right) + \left( 2 - 2 \alpha _ { 5 } \right) K + \left( 1 + \alpha _ { 4 } + \alpha _ { 5 } \right) K ^ { 2 } }$ ,

$\beta _ { 5 } = \frac { \left( 1 - \alpha _ { 4 } + \alpha _ { 5 } \right) - \left( 2 - 2 \alpha _ { 5 } \right) K + \left( 1 + \alpha _ { 4 } + \alpha _ { 5 } \right) K ^ { 2 } } { \left( 1 - \alpha _ { 4 } + \alpha _ { 5 } \right) + \left( 2 - 2 \alpha _ { 5 } \right) K + \left( 1 + \alpha _ { 4 } + \alpha _ { 5 } \right) K ^ { 2 } }$ .

$If \quad \quad K = 1 \quad then \quad \alpha _ { N } = \beta _ { N } , \quad \quad N \in [ 1,2,3,4,5 ]$

------------------------------------------------------------------------------------------------------------------------------------------------------------------

์ด์‚ฐ์‹œ๊ฐ„ ๊ทน์  ์žฌ๋ฐฐ์น˜ ๋ฐฉ๋ฒ•์€ RLS ๋ชจ๋ธ๋กœ๋ถ€ํ„ฐ ๋น ๋ฅธ ์‘๋‹ต์„ ์–ป๊ธฐ ์œ„ํ•ด ์‹œ์Šคํ…œ ๊ทน์ ์„ ์žฌ๋ฐฐ์น˜ํ•˜๋Š” ๋ฐฉ๋ฒ•์ด๋‹ค. ๊ด€์ธก๊ธฐ ๋ชจ๋ธ์˜ ํŠน์„ฑ ๋ฐฉ์ •์‹์€ ๋‹ค์Œ๊ณผ ๊ฐ™์ด ์ •์˜ ๋œ๋‹ค.

(13)
$\lambda ^ { 2 } + \beta _ { 4 } \lambda + \beta _ { 5 } = \operatorname { det } \{ \lambda I - ( A [ n ] - L [ n ] C [ n ] ) \} = 0$

์—ฌ๊ธฐ์„œ $C [ n ] = [ {1} \quad {0} ]$์ด๋‹ค. ์‹(13)์œผ๋กœ๋ถ€ํ„ฐ $L [ n ]$์€ ๋‹ค์Œ๊ณผ ๊ฐ™์ด ๊ฒฐ์ •๋œ๋‹ค.

(14)
$L _ { 1 } [ n ] = \beta _ { 4 } [ n ] - \alpha _ { 4 } [ n ]$ , $L _ { 2 } [ n ] = \beta _ { 5 } [ n ] - \alpha _ { 5 } [ n ]$

3.3 ๊ฐ๊ฐ€์†๋„ ์ถ”์ • ์ƒํƒœ๊ด€์ธก๊ธฐ

์œ ํ•œ ์ฐจ๋ถ„๋ฒ•์— ์˜ํ•ด์„œ ์ƒํƒœ๋ฅผ ๋‹ค์Œ๊ณผ ๊ฐ™์ด ์ •์˜ํ•œ๋‹ค.

(15)
$\begin{aligned} X _ { 1 } [ n + 1 ] & = X _ { 1 } [ n ] + T _ { s } X _ { 2 } [ n ] , \\ X _ { 2 } [ n + 1 ] & = X _ { 2 } [ n ] + T _ { s } \left( - \alpha _ { 5 } [ n ] X _ { 1 } [ n ] - \alpha _ { 4 } [ n ] X _ { 2 } [ n ] + u [ n ] \right) \end{aligned}$

(15)์—์„œ $X _ { 2 } [ n ]$์€ ๊ฐ์†๋„ ์ƒํƒœ๋ฅผ ๋‚˜ํƒ€๋‚ด๊ณ  $\left( - \alpha _ { 5 } [ n ] X _ { 1 } [ n ] - \alpha _ { 4 } [ n ] X _ { 2 } [ n ] + u [ n ] \right)$๋Š” ๊ฐ๊ฐ€์†๋„ ์ƒํƒœ๋ฅผ ๋‚˜ํƒ€๋‚ด๊ณ  ์žˆ์Œ์„ ์•Œ ์ˆ˜๊ฐ€ ์žˆ๋‹ค. ๋”ฐ๋ผ์„œ ํ™•์žฅ๋œ ๋ฃจ์—”๋ฒ„๊ฑฐ ๊ด€์ธก๊ธฐ๋ฅผ ๋‹ค์Œ๊ณผ ๊ฐ™์ด ์ œ์•ˆํ•œ๋‹ค.

(16)
$X _ { v } [ n ] = X _ { v } [ n - 1 ] + T _ { s } X _ { a } [ n - 1 ] + L _ { 1 } [ n ] e _ { y } [ n ]$ , $X _ { a } [ n ] = - \alpha _ { 5 } [ n ] X _ { 1 } [ n ] - \alpha _ { 4 } [ n ] X _ { 2 } [ n ] + u [ n ] + L _ { 2 } [ n ] e _ { y } [ n ]$

์—ฌ๊ธฐ์„œ $X _ { v } [ n ]$ ๊ฐ์†๋„ ์ƒํƒœ๋กœ ์ •์˜๋˜๊ณ  $X _ { a } [ n ]$๋Š” ๊ฐ๊ฐ€์†๋„ ์ƒํƒœ๋กœ ์ •์˜๋œ๋‹ค.

3.4 ์ƒ๋ณด ํ•„ํ„ฐ ์„ค๊ณ„

RLS ๋ฐฉ๋ฒ•์€ ์ถฉ๋ถ„ํ•œ ๋ฐ์ดํ„ฐ๋ฅผ ํ•„์š”๋กœ ํ•œ๋‹ค. ์ถฉ๋ถ„ํ•œ ์‚ฌ์ „ ์ง€์‹์ด ๋ถ€์กฑํ•œ ์ƒํƒœ์—์„œ RLS ๋ฐฉ๋ฒ•์€ ์ดˆ๊ธฐ ํ”ผํ‚น ํšจ๊ณผ๋ฅผ ๋ฐœ์ƒ์‹œํ‚จ๋‹ค. ์ด ๋ฌธ์ œ์ ์€ RLS ๊ธฐ๋ฐ˜ ์ƒํƒœ๊ด€์ธก๊ธฐ์˜ ๊ฐ๊ฐ€์†๋„ ์ถ”์ • ๊ฒฐ๊ณผ์—์„œ ์ผ์ •ํ•œ ๋ฐ”์ด์–ด์Šค ์‹ ํ˜ธ๋กœ ๋‚˜ํƒ€๋‚œ๋‹ค. ๊ทธ๋ฆผ. 2๋Š” ๊ทธ๋ฆผ. 1(b)์— ๋Œ€ํ•œ ์‹คํ—˜ ๊ฒฐ๊ณผ๋ฅผ ๋ณด์—ฌ์ค€๋‹ค. ๊ทธ๋ฆผ. 1(b)์—์„œ (2)๋ฅผ ์ด์šฉํ•˜์—ฌ ์„ ๊ฐ€์†๋„๋ฅผ ์ถ”์ •ํ•  ๊ฒฝ์šฐ ์ธก์ • ๊ฐ’์— ๋น„ํ•ด์„œ ์žก์Œ์€ ์ž‘์ง€๋งŒ ์ดˆ๊ธฐ ํ”ผํ‚น ํšจ๊ณผ๋กœ ์ธํ•ด ์ผ์ •ํ•˜๊ฒŒ ๋ฐ”์ด์–ด์Šค๋œ ๊ด€์ธก ๊ฒฐ๊ณผ๊ฐ€ ๋‚˜ํƒ€๋‚จ์„ ๋ณผ ์ˆ˜๊ฐ€ ์žˆ๋‹ค.

๊ทธ๋ฆผ. 2. ํ”ผํ‚น ํšจ๊ณผ์™€ ๋ฐ”์ด์–ด์Šค ๋ฌธ์ œ

Fig. 2. Peaking and bias problems

../../Resources/kiee/KIEE.2019.68.2.342/fig2.png

๋”ฐ๋ผ์„œ ๋ณธ ์—ฐ๊ตฌ์—์„œ๋Š” ์ƒ๋ณด ํ•„ํ„ฐ์— ์˜ํ•œ ๋ฐ”์ด์–ด์Šค ๋ณด์ƒ ๋ฐฉ๋ฒ•์„ ์ œ์•ˆํ•œ๋‹ค. ์ œ์•ˆํ•˜๋Š” ์ƒ๋ณด ํ•„ํ„ฐ ๊ตฌ์กฐ๋Š” ๊ทธ๋ฆผ. 3๊ณผ ๊ฐ™๋‹ค. ๊ทธ๋ฆผ. 3์—์„œ $X _ { a , 2 } [ n ]$๋Š” ์œ ํ•œ์ฐจ๋ถ„๋ฒ•์— ์˜ํ•œ ๊ฐ๊ฐ€์†๋„ ์ƒํƒœ์ด๊ณ  $X _ { a , c o m p } [ n ]$๋Š” ํ“จ์ „๋œ ๊ฐ๊ฐ€์†๋„ ์ƒํƒœ์ด๋‹ค.

๊ทธ๋ฆผ. 3. ์ œ์•ˆํ•˜๋Š” ์ƒ๋ณด ํ•„ํ„ฐ

Fig. 3. Proposed complementary filter

../../Resources/kiee/KIEE.2019.68.2.342/fig3.png

๊ทธ๋ฆผ. 3์—์„œ HPF๋Š” ๋‹ค์Œ๊ณผ ๊ฐ™๋‹ค.

(17)
$H P F = \frac { \frac { 2 } { T _ { s } } \left( 1 - z ^ { - 1 } \right) } { \left( \omega _ { c } + \frac { 2 } { T _ { s } } \right) + \left( \omega _ { c } - \frac { 2 } { T _ { s } } \right) z ^ { - 1 } }$

๊ทธ๋ฆผ. 3์—์„œ LPF๋Š” ๋‹ค์Œ๊ณผ ๊ฐ™๋‹ค.

(18)
$L P F = \frac { \omega _ { c } \left( 1 + z ^ { - 1 } \right) } { \left( \omega _ { c } + \frac { 2 } { T _ { s } } \right) + \left( \omega _ { c } - \frac { 2 } { T _ { s } } \right) z ^ { - 1 } }$

์—ฌ๊ธฐ์„œ $\omega _ { c } = 2 \pi f _ { c }$์ด๊ณ  $f _ { c }$๋Š” ์ฐจ๋‹จ์ฃผํŒŒ์ˆ˜์ด๋‹ค.

4. ์‹œ๋ฎฌ๋ ˆ์ด์…˜ ๊ฒ€์ฆ

๋ณธ ์žฅ์—์„œ๋Š” ์ œ์•ˆํ•œ ๋ฐฉ๋ฒ•๋“ค์— ๋Œ€ํ•œ ์‹œ๋ฎฌ๋ ˆ์ด์…˜ ์—ฐ๊ตฌ ๊ฒฐ๊ณผ๋ฅผ ์ œ์‹œํ•œ๋‹ค.

4.1 ์ด์‚ฐ ์‹œ๊ฐ„ ๊ทน์  ์žฌ๋ฐฐ์น˜ ๋ฐฉ๋ฒ•

๋‘ ๊ทน์ ์ด ๊ฐ๊ฐ 0.3, 0.8์ด๋ผ๊ณ  ๊ฐ€์ •ํ•  ๊ฒฝ์šฐ K๊ฐ’์— ์˜ํ•œ ์˜ํ–ฅ์€ ๊ทธ๋ฆผ. 4์™€ ๊ฐ™๋‹ค. ๊ทธ๋ฆผ. 4๋Š” ์‹œ๊ฐ์ ์œผ๋กœ ํ™•์ธํ•˜๊ธฐ ์‰ฝ๊ฒŒ ์—ฐ์†์‹œ๊ฐ„์˜์—ญ์œผ๋กœ ๋ณ€ํ™˜๋œ ์‘๋‹ต ํŠน์„ฑ์„ ์ œ์‹œํ•œ๋‹ค.

๊ทธ๋ฆผ. 4. ์Šคํ… ์‘๋‹ต

Fig. 4. Step response

../../Resources/kiee/KIEE.2019.68.2.342/fig4.png

๊ทธ๋ฆผ. 4์—์„œ ๋ณด๋ฉด K ๊ฐ’์˜ ์„ค์ •์— ๋”ฐ๋ผ ์‘๋‹ต ์‹œ๊ฐ„์ด ๋ณ€ํ™”๋จ์„ ๋ณผ ์ˆ˜ ์žˆ๋‹ค. K ๊ฐ’์ด 20์ผ ๊ฒฝ์šฐ ์˜ค์‹ค๋ ˆ์ด์…˜์ด ๋ฐœ์ƒํ•จ์„ ์•Œ ์ˆ˜ ์žˆ๋‹ค. ๊ทธ๋ฆผ. 5๋Š” ๊ฐ ๊ฐ์˜ K ๊ฐ’์— ๋”ฐ๋ฅธ ๊ทน์ ์˜ ์žฌ๋ฐฐ์น˜ ์ƒํƒœ๋ฅผ ๋‚˜ํƒ€๋‚ธ๋‹ค. K ๊ฐ’์ด 1๋ณด๋‹ค ์ž‘์„ ๊ฒฝ์šฐ ๊ทน์ ์€ ์šฐ์ธก์œผ๋กœ ์ด๋™ํ•˜๊ฒŒ ๋˜๊ณ  K ๊ฐ’์ด 1๋ณด๋‹ค ํด ๊ฒฝ์šฐ ๊ทน์ ์€ ์ขŒ์ธก์œผ๋กœ ์ด๋™ํ•จ์„ ๋ณผ ์ˆ˜ ์žˆ๋‹ค. 1 ๋ณด๋‹ค ํฐ K ๊ฐ’์€ ๋น ๋ฅธ ์‘๋‹ต ํŠน์„ฑ์„ ๋ณด์ด์ง€๋งŒ ๋งค์šฐ ํฐ ๊ฐ’์—์„œ๋Š” ์˜ค์‹ค๋ ˆ์ด์…˜ ํ˜„์ƒ์ด ๋ฐœ์ƒํ•˜๋ฏ€๋กœ ์ ์ ˆํ•œ K ๊ฐ’์„ ์‚ฌ์šฉํ•ด์•ผ๋งŒ ํ•œ๋‹ค. ๋ณธ ์—ฐ๊ตฌ์—์„œ๋Š” K ๊ฐ’์€ 5๋ฅผ ์‚ฌ์šฉํ•œ๋‹ค.

๊ทธ๋ฆผ. 5. ๊ทน์  ์œ„์น˜ ๋ณ€ํ™”

Fig. 5. Movements of poles

../../Resources/kiee/KIEE.2019.68.2.342/fig5.png

4.2 ์ƒ๋ณด ํ•„ํ„ฐ ํšจ๊ณผ

๋‹ค์Œ์œผ๋กœ ์ƒ๋ณด ํ•„ํ„ฐ์˜ ํšจ๊ณผ๋ฅผ ๋ถ„์„ํ•œ๋‹ค.

๊ทธ๋ฆผ. 6์—์„œ estimator1์€ ์œ ํ•œ์ฐจ๋ถ„๋ฒ•์— ์˜ํ•œ ์ถ”์ • ๊ฐ๊ฐ€์†๋„[21]์ด๊ณ  estimator2๋Š” ๊ด€์ธก๊ธฐ์— ์˜ํ•œ ์ถ”์ • ๊ฐ๊ฐ€์†๋„์ด๊ณ  estimator3๋Š” ์ƒ๋ณด ํ•„ํ„ฐ์— ์˜ํ•œ ์ถ”์ • ๊ฐ๊ฐ€์†๋„์ด๋‹ค. ์ œ์•ˆํ•œ ์ƒ๋ณด ํ•„ํ„ฐ์˜ ํŠน์„ฑ์€ ์ฃผํŒŒ์ˆ˜ ์˜์—ญ ๋ถ„์„์„ ํ†ตํ•ด ํ™•์ธ๋  ์ˆ˜๊ฐ€ ์žˆ๋‹ค. ๊ทธ๋ฆผ. 7์€ ์ฃผํŒŒ์ˆ˜ ์˜์—ญ์—์„œ ์„ธ ๊ฐœ์˜ ์ถ”์ •๊ธฐ๋ฅผ ๋น„๊ตํ•œ ๊ฒฐ๊ณผ์ด๋‹ค. Estimator1์˜ ๊ฒฝ์šฐ ์ €์ฃผํŒŒ์ˆ˜์—์„œ ๊ฐ•๊ฑดํ•œ ํŠน์„ฑ์„ ๋ณด์ด๊ณ  ์žˆ๊ณ  estimator2์˜ ๊ฒฝ์šฐ ๊ณ ์ฃผํŒŒ์ˆ˜์—์„œ ๊ฐ•๊ฑดํ•œ ํŠน์„ฑ์„ ๋ณด์ด๊ณ  ์žˆ์Œ์„ ์•Œ ์ˆ˜ ์žˆ๋‹ค. ๋˜ํ•œ estimator2์˜ ๊ฒฝ์šฐ RLS์— ์˜ํ•œ ์ดˆ๊ธฐ ํ”ผํ‚น ํšจ๊ณผ์˜ ์˜ํ–ฅ์— ์˜ํ•ด์„œ ์ €์ฃผํŒŒ์ˆ˜์—์„œ ๋ฏผ๊ฐํ•˜๊ฒŒ ๋ฐ˜์‘ํ•˜๊ณ  ์žˆ์Œ์„ ์•Œ ์ˆ˜ ์žˆ๋‹ค. ๋ณธ ์‹œ๋ฎฌ๋ ˆ์ด์…˜์—์„œ๋Š” 0.5 Hz์˜ ์ฐจ๋‹จ ์ฃผํŒŒ์ˆ˜๋ฅผ ์‚ฌ์šฉํ•˜์˜€๋‹ค. Estimator3๋Š” ์ €์ฃผํŒŒ์ˆ˜์—์„œ๋Š” estimator1์˜ ํŠน์„ฑ์„ ์ฃผ์š”ํ•˜๊ฒŒ ๋”ฐ๋ฅด๊ณ  ๊ณ ์ฃผํŒŒ์ˆ˜์—์„œ๋Š” estimator2์˜ ํŠน์„ฑ์„ ์ฃผ์š”ํ•˜๊ฒŒ ๋”ฐ๋ฅด๋Š” ์ƒ๋ณด ํšจ๊ณผ๊ฐ€ ๋ฐœ์ƒํ•จ์„ ํ™•์ธํ•  ์ˆ˜๊ฐ€ ์žˆ๋‹ค. ์ ์ ˆํ•œ ์ฐจ๋‹จ์ฃผํŒŒ์ˆ˜ ์„ค์ •์„ ํ†ตํ•ด์„œ RLS์˜ ์ดˆ๊ธฐ ํ”ผํ‚น ํšจ๊ณผ๋ฅผ ์–ต์ œํ•˜๊ณ  ๋™์‹œ์— ์ด์ค‘ ํ•„ํ„ฐ์˜ ๋ฆฌํ”Œ ๋ฌธ์ œ๋ฅผ ๋™์‹œ์— ํ•ด๊ฒฐํ•  ์ˆ˜ ์žˆ์Œ์„ ์•Œ ์ˆ˜ ์žˆ๋‹ค.

๊ทธ๋ฆผ. 6. ์ƒ๋ณด ํ•„ํ„ฐ ํšจ๊ณผโ€“์‹œ๊ฐ„ ์˜์—ญ

Fig. 6. Effect of a complementary filter-time domain

../../Resources/kiee/KIEE.2019.68.2.342/fig6.png

๊ทธ๋ฆผ. 7. ์ƒ๋ณด ํ•„ํ„ฐ ํšจ๊ณผ - ์ฃผํŒŒ์ˆ˜ ์˜์—ญ

Fig. 7. Effect of a complementary filter-frequency domain

../../Resources/kiee/KIEE.2019.68.2.342/fig7.png

5. ์‹คํ—˜ ๊ฒ€์ฆ

์•ž์„œ ์ œ์‹œํ•œ ๊ฐ๊ฐ€์†๋„ ์ƒํƒœ๊ด€์ธก๊ธฐ์™€ ์ƒ๋ณด ํ•„ํ„ฐ๋Š” PC ๊ธฐ๋ฐ˜ VC++ ์†Œํ”„ํŠธ์›จ์–ด์— ์ ์šฉ๋˜์—ˆ๋‹ค. ๊ทธ๋ฆผ. 1(b)์˜ ์‹คํ—˜์„ ์œ„ํ•ด์„œ ๋กœ๋ด‡ํŒ”์„ ์‚ฌ์šฉํ•˜์˜€๋‹ค. ์‹คํ—˜ ํ™˜๊ฒฝ์€ ๊ทธ๋ฆผ. 8๊ณผ ๊ฐ™๋‹ค.

๊ทธ๋ฆผ. 8. ์‹คํ—˜ ํ™˜๊ฒฝ

Fig. 8. Experimental setup

../../Resources/kiee/KIEE.2019.68.2.342/fig8.png

์‹คํ—˜์—์„œ ๋งค๋‹ˆํ“ฐ๋ ˆ์ดํ„ฐ์˜ ์šฐ์ธก ํŒ”์˜ ๋๋‹จ์— Lord MicroStrain ์‚ฌ์˜ 3DM-GX4-25 AHRS ์„ผ์„œ๋ฅผ ์žฅ์ฐฉํ•˜์˜€๋‹ค. ์žฅ์ฐฉ๋œ AHRS ์„ผ์„œ๋Š” 3์ถ•์˜ ์„ ๊ฐ€์†๋„๋ฅผ ์ธก์ •ํ•  ์ˆ˜๊ฐ€ ์žˆ๋‹ค. RLS ์ธ์‹์— ์‚ฌ์šฉ๋œ ํƒ€๊ฒŸ ์ถ•์˜ ์ž…๋ ฅ ๋ฐ์ดํ„ฐ์™€ ์ถœ๋ ฅ ๋ฐ์ดํ„ฐ๋Š” ๋‹ค์Œ๊ณผ ๊ฐ™๋‹ค.

๊ทธ๋ฆผ. 9. RLS ์ž…๋ ฅ๊ณผ ์ถœ๋ ฅ ๋ฐ์ดํ„ฐ

Fig. 9. Input and output data of RLS

../../Resources/kiee/KIEE.2019.68.2.342/fig9.png

๊ฐ๊ฐ€์†๋„ ์ถ”์ • ์‹คํ—˜ ๊ฒฐ๊ณผ๋Š” ๋‹ค์Œ๊ณผ ๊ฐ™๋‹ค.

๊ทธ๋ฆผ. 10. ๊ฐ๊ฐ€์†๋„ ์ถ”์ • ๊ฒฐ๊ณผ

Fig. 10. Angular velocity estimation

../../Resources/kiee/KIEE.2019.68.2.342/fig10.png

๊ทธ๋ฆผ. 10์—์„œ estimator1์€ ์œ ํ•œ์ฐจ๋ถ„๋ฒ•์— ์˜ํ•œ ์ถ”์ •์น˜์ด๊ณ  estimator2๋Š” ๊ด€์ธก๊ธฐ์— ์˜ํ•œ ์ถ”์ •์น˜์ด๊ณ  estimator3๋Š” ์ƒ๋ณดํ•„ํ„ฐ์— ์˜ํ•œ ์ถ”์ • ๊ฒฐ๊ณผ์ด๋‹ค. Estimator1์€ ๊ณ ์กฐํŒŒ ๋ฆฌํ”Œ์— ์˜ํ•œ ์žก์Œ์ด ๋ฐœ์ƒํ•˜๊ณ  estimator2๋Š” ์ €์ฃผํŒŒ ๋Œ€์—ญ์—์„œ ๋ฏผ๊ฐํ•˜๊ณ  estimator3๋Š” ์ €์ฃผํŒŒ ๋Œ€์—ญ์—์„œ๋Š” estimator1์˜ ํŠน์„ฑ์„ ๋”ฐ๋ฅด๊ณ  ๊ณ ์ฃผํŒŒ ๋Œ€์—ญ์—์„œ๋Š” estimator2์˜ ํŠน์„ฑ์„ ๋”ฐ๋ฅด๊ณ  ์žˆ๋‹ค๋Š” ์ ์„ ๋ณผ ์ˆ˜๊ฐ€ ์žˆ๋‹ค. ๋ณธ ์‹คํ—˜์—์„œ๋Š” 0.1 ์ดˆ๋ผ๋Š” ๋Š๋ฆฐ ์ƒ˜ํ”Œ๋ง ์ฃผ๊ธฐ๋ฅผ ์‚ฌ์šฉํ•˜์˜€๋‹ค. ์ƒ˜ํ”Œ๋ง์ฃผ๊ธฐ์˜ ์˜ํ–ฅ์— ์žˆ์–ด์„œ estimator1์˜ ๊ฒฝ์šฐ ์ƒ˜ํ”Œ๋ง ์‹œ๊ฐ„์ด ์งง์„์ˆ˜๋ก ๊ณ ์กฐํŒŒ ๋ฆฌํ”Œ์— ์˜ํ•œ ์žก์Œ ์˜ํ–ฅ์ด ๊ฐ•ํ•ด์ง€๊ฒŒ ๋˜๊ณ  estimator2์˜ ๊ฒฝ์šฐ ์ƒ˜ํ”Œ๋ง ์‹œ๊ฐ„์ด ์งง์„์ˆ˜๋ก ์ €์ฃผํŒŒ ํ”ผํ‚น ํšจ๊ณผ๊ฐ€ ๊ฐ์†Œํ•˜๊ฒŒ ๋œ๋‹ค. ๋”ฐ๋ผ์„œ ๊ด€์ธก๊ธฐ๋ฅผ ์‚ฌ์šฉํ•  ๊ฒฝ์šฐ ์ผ๋ฐ˜์ ์œผ๋กœ ๋น ๋ฅธ ์ƒ˜ํ”Œ๋ง ์ฃผ๊ธฐ๋ฅผ ์š”๊ตฌํ•˜๊ฒŒ ๋œ๋‹ค. ํ•˜์ง€๋งŒ ๋Š๋ฆฐ ์ƒ˜ํ”Œ๋ง ์ฃผ๊ธฐ๋ฅผ ๊ฐ–๋Š” ์‹œ์Šคํ…œ์˜ ๊ฒฝ์šฐ ์ƒ๋ณด ํ•„ํ„ฐ๋ง ๋ฐฉ๋ฒ•์„ ์‚ฌ์šฉํ•ด์„œ ๊ฐ๊ฐ€์†๋„ ์ถ”์ • ์„ฑ๋Šฅ์„ ํ–ฅ์ƒํ•  ์ˆ˜ ์žˆ์„ ๊ฑฐ๋ผ๊ณ  ํŒ๋‹จ๋œ๋‹ค.

๊ฒฐ๊ณผ์ ์œผ๋กœ ์ œ์‹œํ•œ ๋ฐฉ๋ฒ•์— ์˜ํ•œ ์„ ๊ฐ€์†๋„ ์ถ”์ • ๊ฒฐ๊ณผ๋Š” ๊ทธ๋ฆผ. 11๊ณผ ๊ฐ™๋‹ค.

๊ทธ๋ฆผ. 11. ์„ ๊ฐ€์†๋„ ์ถ”์ • ๊ฒฐ๊ณผ

Fig. 11. Linear No.elocity estimation

../../Resources/kiee/KIEE.2019.68.2.342/fig11.png

๊ทธ๋ฆผ. 11์—์„œ ๋ณด๋ฉด AHRS ์„ผ์„œ ์ธก์ • ๊ฒฐ๊ณผ๋Š” ์žก์Œ์„ ๋งค์šฐ ๋งŽ์ด ํฌํ•จํ•˜๊ณ  ์žˆ์Œ์„ ์•Œ ์ˆ˜๊ฐ€ ์žˆ๋‹ค. Estimator1์˜ ๊ฒฝ์šฐ ๊ฐ๊ฐ€์†๋„ ์ถ”์ • ์žก์Œ์ด ์„ ๊ฐ€์†๋„ ์ถ”์ • ๊ฒฐ๊ณผ์— ๋ฐ˜์˜๋˜๊ณ  ์žˆ์Œ์„ ์•Œ ์ˆ˜ ์žˆ๋‹ค. Estimator2์˜ ๊ฒฝ์šฐ ์ดˆ๊ธฐ ํ”ผํ‚น ๋ฌธ์ œ๋กœ ์ธํ•ด ์ถ”์ • ๊ฒฐ๊ณผ๋“ค์— ์˜คํ”„์…‹๋“ค์ด ์กด์žฌํ•  ์ˆ˜ ์žˆ์Œ์„ ์•Œ ์ˆ˜๊ฐ€ ์žˆ๋‹ค. ํ•˜์ง€๋งŒ Estimator2์˜ ๊ฒฝ์šฐ ๊ณ ์ฃผํŒŒ ์žก์Œ์— ๊ฐ•๊ฑดํ•œ ํŠน์„ฑ์„ ๊ฐ–๊ณ  ์žˆ์Œ์„ ์•Œ ์ˆ˜ ์žˆ๋‹ค. Estimator3์˜ ๊ฒฝ์šฐ estimator2์™€ estimator3์˜ ์žฅ์ ์„ ํ“จ์ „ํ•œ ๊ฒฐ๊ณผ์ด๋‹ค.

๊ตฌ๋™์ถ•์˜ ์ƒํƒœ๋ฅผ ์ด์šฉํ•ด์„œ ์‹œ์Šคํ…œ์˜ end-effector์˜ ์ƒํƒœ๋ฅผ ์ถ”์ •ํ•˜๋Š” ๊ฒƒ์€ ๋งค์šฐ ํฅ๋ฏธ๋กœ์šด ์ผ์ด๋‹ค. ๋™์  ์‹œ์Šคํ…œ์˜ ์ƒํƒœ๋ฅผ ์ง์ ‘ ์ธก์ •ํ•  ์ˆ˜ ์—†์„ ๊ฒฝ์šฐ ์ œ์•ˆ๋œ ๊ฐ„์ ‘ ์ธก์ • ๋ฐฉ๋ฒ•์ด ์šฉ์ดํ•˜๊ฒŒ ์‚ฌ์šฉ๋  ์ˆ˜ ์žˆ๊ธฐ ๋•Œ๋ฌธ์ด๋‹ค.

์ตœ์ข…์ ์œผ๋กœ ๊ทธ๋ฆผ. 11์„ ํ†ตํ•ด ํ™•์ธ๋œ 4๊ฐœ์˜ ๋ฐฉ๋ฒ•๋“ค์— ๋Œ€ํ•œ ์žฅ์ ๊ณผ ๋‹จ์ ์„ ๋‹ค์Œ๊ณผ ๊ฐ™์ด ์ •๋ฆฌํ•˜์—ฌ ์ œ์‹œํ•œ๋‹ค.

6. ๊ฒฐ ๋ก 

๋ณธ ๋…ผ๋ฌธ์—์„œ๋Š” ๋™์  ์‹œ์Šคํ…œ์˜ ๊ฐ๊ฐ€์†๋„ ์ถ”์ •์„ ์œ„ํ•œ ์ƒํƒœ๊ด€์ธก๊ธฐ ์„ค๊ณ„ ๋ฐฉ๋ฒ•๊ณผ ํ•„ํ„ฐ๋ฅผ ์ œ์•ˆํ•˜์˜€๋‹ค. ์ƒํƒœ๊ด€์ธก๊ธฐ ์„ค๊ณ„๋ฅผ ์œ„ํ•ด์„œ ๊ตฌ๋™์ถ•์€ ์ด์ฐจ ์‹œ์Šคํ…œ์œผ๋กœ ๋ชจ๋ธ๋ง๋˜์—ˆ๊ณ  ๋ชจ๋ธ ๊ธฐ๋ฐ˜ ์ƒํƒœ๊ด€์ธก๊ธฐ์˜ ๊ฒŒ์ธ๊ฐ’์€ ๊ทน์  ์žฌ๋ฐฐ์น˜ ๋ฐฉ๋ฒ•์„ ํ†ตํ•ด ์„ค๊ณ„๋˜์—ˆ๊ณ  ๊ฐ๊ฐ€์†๋„ ์ถ”์ •์„ ์œ„ํ•ด ์œ ํ•œ ์ฐจ๋ถ„๋ฒ•์„ ์ด์šฉํ•ด์„œ ์ƒํƒœ์‹์„ ํ™•์žฅํ•˜์˜€๋‹ค. ์„ค๊ณ„๋œ ์ƒํƒœ ์ถ”์ •๊ธฐ๋Š” ์‹ค์‹œ๊ฐ„ ์‹œ์Šคํ…œ ์ธ์‹ ๋ชจ๋ธ์„ ์ด์šฉํ•œ๋‹ค๋Š” ์žฅ์ ์„ ๊ฐ–๊ณ  ์žˆ์—ˆ์ง€๋งŒ ๋ฐ์ดํ„ฐ ๊ธฐ๋ฐ˜ ์‹œ์Šคํ…œ ์ธ์‹ ๋ชจ๋ธ์˜ ์ดˆ๊ธฐ ํ”ผํ‚น ํšจ๊ณผ๊ฐ€ ๋ฐœ์ƒํ•˜์˜€๋‹ค. ํ”ผํ‚น ํšจ๊ณผ๋กœ ์ธํ•ด ์ƒํƒœ ์ถ”์ •๊ธฐ์˜ ์ถ”์ • ๊ฒฐ๊ณผ ๊ฐ’์— ์˜คํ”„์…‹ ๋ฌธ์ œ๊ฐ€ ๋ฐœ์ƒํ–ˆ๋‹ค. ์ด ๋ฌธ์ œ ํ•ด๊ฒฐ์„ ์œ„ํ•ด ์œ ํ•œ์ฐจ๋ถ„๋ฒ•์— ์˜ํ•œ ๊ฐ€์†๋„ ์ถ”์ • ๋ฐฉ๋ฒ•๊ณผ ๊ด€์ธก๊ธฐ๋ฅผ ํ“จ์ „ํ•˜๋Š” ์ƒ๋ณด ํ•„ํ„ฐ๋ฅผ ์„ค๊ณ„ํ•˜์˜€๋‹ค. ๋ณธ ๋…ผ๋ฌธ์—์„œ๋Š” ์ ์ ˆํ•œ ์ฐจ๋‹จ์ฃผํŒŒ์ˆ˜ ์„ค์ •์„ ํ†ตํ•ด์„œ ๋‘ ๊ฐœ์˜ ์ถ”์ • ๋ฐฉ๋ฒ•์ด ์„œ๋กœ ๋ณด์™„๋  ์ˆ˜ ์žˆ์Œ์„ ํ™•์ธํ•  ์ˆ˜๊ฐ€ ์žˆ์—ˆ๋‹ค. ์‹คํ—˜ ๊ฒฐ๊ณผ, ์ œ์‹œ๋œ ์ƒ๋ณด ํ•„ํ„ฐ ๋ฐฉ๋ฒ•์„ ํ†ตํ•ด ๊ด€์ธก๊ธฐ์˜ ์ดˆ๊ธฐ ํ”ผํ‚น ํšจ๊ณผ๊ฐ€ ํ˜„์ €ํžˆ ์ค„์–ด๋“ฆ์„ ํ™•์ธํ•  ์ˆ˜ ์žˆ์—ˆ๊ณ  ์œ ํ•œ์ฐจ๋ถ„๋ฒ•์˜ ๊ณ ์กฐํŒŒ ๋ฆฌํ”Œ ๋ฌธ์ œ๋„ ๋™์‹œ์— ํ•ด๊ฒฐ๋  ์ˆ˜ ์žˆ์Œ์„ ํ™•์ธํ•  ์ˆ˜๊ฐ€ ์žˆ์—ˆ๋‹ค. ๋ณธ ์—ฐ๊ตฌ๋Š” 10Hz์˜ ๋‚ฎ์€ ์ƒ˜ํ”Œ๋ง ์ฃผ๊ธฐ์—์„œ ์ˆ˜ํ–‰๋˜์—ˆ๋‹ค. ๊ณ ์†์˜ ์ƒ˜ํ”Œ๋ง ์ฃผ๊ธฐ๋ฅผ ๊ฐ–๋Š” ์‹œ์Šคํ…œ์—์„œ ๋‘ ๊ฐ€์ง€ ๋ฐฉ๋ฒ•์ด ์–ด๋–ค ํ˜•ํƒœ๋กœ ํ“จ์ „๋  ์ˆ˜ ์žˆ๋Š” ๊ฒƒ์ธ๊ฐ€์— ๋Œ€ํ•œ ์—ฐ๊ตฌ๊ฐ€ ํ•„์š”ํ•˜๋‹ค. ๋˜ํ•œ ์ƒ๋ณด ํ•„ํ„ฐ์˜ ์ฐจ๋‹จ์ฃผํŒŒ์ˆ˜๋ฅผ ์‹ค์‹œ๊ฐ„ ์ตœ์ ํ™”ํ•˜๋Š” ์—ฐ๊ตฌ๊ฐ€ ํ•„์š”ํ•˜๋‹ค.

๊ฐ์‚ฌ์˜ ๊ธ€

๋ณธ ์—ฐ๊ตฌ๋Š” 2017๋…„ ํ•œ๊ตญ์—ฐ๊ตฌ์žฌ๋‹จ ๊ธฐ์ดˆ์—ฐ๊ตฌ(NRF-2016 R1A22012031)์˜ ์ง€์›์„ ๋ฐ›์•„ ์ด๋ฃจ์–ด์ง„ ์—ฐ๊ตฌ๋กœ ์ง€์›์— ๊ฐ์‚ฌ๋“œ๋ฆฝ๋‹ˆ๋‹ค.

References

1 
Ohnishi K., Shibata M., Murakami T., 1996, Motion control for advanced mechatronics, IEEE/ASME Transactions on Mechatronics, Vol. 1, No. 1, pp. 56-67DOI
2 
Yao B., Majed M. A., Tomizuka M., 1997, High- performance robust motion control of machine tools: an adaptive robust control approach and comparative experiments, IEEE/ASME Transactions on Mechatronics, Vol. 2, No. 2, pp. 63-76DOI
3 
Lee J., Chang P. H., Jin M., 2017, Adaptive integral sliding mode control with time-delay estimation for robot manipulators, IEEE Transactions on Industrial Electronics, Vol. 64, No. 8, pp. 6796-6804DOI
4 
Baek J., Cho S., Han S., 2017, Practical time-delay control with adaptive gains for trajectory tracking of robot manipulators, IEEE Transactions on Industrial Electronics, Vol. 65, No. 7, pp. 5682-5692DOI
5 
Bae Y. G., Jung S., 2018, Balancing control of a mobile manipulator with two-wheels by an acceleration-based disturbance observer, International Journal of Humanoid Robotics, Vol. 15, No. 3DOI
6 
Jeong S. H., Jung S., Tomizuka M., 2012, Attitude control of a quad-rotor system using an acceleration-based disturbance observer: An empirical approach, IEEE/ ASME International Conference on Advanced Intelligent Mechatronics, pp. 916-921DOI
7 
Quigley M., Brewer R., Soundararaj S. P., Pradeep V., Le Q., Ng A. Y., 2010, Low-cost accelerometers for robotic manipulator perception, IEEE/RSJ International Conference on Intelligent Robots and Systems, pp. 6168-6174DOI
8 
Roan P., Deshpande N., Wang Y., Pitzer B., 2010, Manipulator stae etimation with low cost accelerometers and gyroscopes, IEEE/RSJ International Conference on Intelligent Robots and Systems, pp. 4822-4827DOI
9 
Chen W., Tomizuka M., 2014, Direct joint space state estimation in robots with multiple elastic joints, IEEE/ASME Transactions on Mechatronics, Vol. 19, No. 2, pp. 697-706DOI
10 
Cantelli L., Muscato G., Nunnari M., Spina D., 2015, A joint-angle estimation method for industrial manipulators using inertial sensors, IEEE/ASME Transactions on Mechatronics, Vol. 20, No. 5, pp. 2486-2495DOI
11 
Shino M., Nagai M., 2003, Independent wheel torque control of small-scale electric vehicle for handling and stability improvement, JSAE Review, Vol. 24, No. 4, pp. 449-456DOI
12 
Osborn R. P., Shim T., 2006, Independent control of all- wheel-drive torque distribution, International Journal of Vehicle Mechanics and Mobility, Vol. 44, No. 7, pp. 529-546DOI
13 
Park J., Jeong H., Jang In. G., Hwang S. H., 2015, Torque distribution algorithm for an independently driven electric vehicle using a fuzzy control method, Energies, No. 8, pp. 8537-8561DOI
14 
He H., Peng J., Xiong R., Fan H., 2014, An Acceleration slip regulation strategy for four-wheel drive electric vehicles based on sliding mode control, Energies, Vol. 7, No. 8, pp. 3748-3763DOI
15 
Shuai Z., Zhang H., Wang J., Li J., Ouyang M., 2014, Combined AFS and DYC control of four-wheel- independent-drive electric vehicles over CAN network with time-varying delays, IEEE Transactions on Vehicular Technology, Vol. 63, No. 2, pp. 591-602DOI
16 
Tempel P., Herve P. El., Tempier O., Gouttefarde M., Pott A., 2017, Estimating inertial parameters of suspended cable-driven parallel robots-use case on CoGiRo, IEEE International Conference on Robotics and Automation, pp. 6093-6098DOI
17 
Atkeson C. G., An C. H., Hollerbach J. M., 1986, Estimation of inertia parameters of manipulator loads and links, The International Journal of Robotics Research, Vol. 5, No. 3, pp. 101-119DOI
18 
Lee S. D., Jung S., 2016, An experimental study on real time estimation of a nominal model for a disturbance observer: Recursive least square approach, Journal of Institute of Control, Robotics and Systems, Vol. 22, No. 8, pp. 650-655Google Search
19 
Lee S. D., Jung S., September 20-23, 2016, An identification technique for non-minimum phase systems by a recursive least square method, Proceedings of the SICE Annual Conference 2016,Tsukuba, Japan, pp. 624-626Google Search
20 
Lee S. D., Jung S., 2018, A state and disturbance observer design technique with a recursive least squares method: a second-order system for a CMG, Journal of Institute of Control, Robotics and Systems, Vol. 24, No. 5, pp. 438-444Google Search
21 
Janabi-Sharifi F., Hayward V., Chen C. S. J., 2000, Discrete-time adaptive windowing for velocity estimation, IEEE Transactions on Control Systems Technology, Vol. 8, No. 6, pp. 1003-1009DOI

์ €์ž์†Œ๊ฐœ

์ด ์ƒ ๋• (Sang-Deok Lee)
../../Resources/kiee/KIEE.2019.68.2.342/au1.png

1998๋…„ ์ „๋ถ๋Œ€ํ•™๊ต ์ „์ž๊ณตํ•™๊ณผ ์กธ์—…

1998๋…„~2000๋…„ LG์ •๋ฐ€ ๊ทผ๋ฌด

2003๋…„ ์ „๋ถ๋Œ€ํ•™๊ต ์ „์ž๊ณตํ•™๊ณผ ์„์‚ฌ

2003๋…„~2014๋…„ ์‚ผ์„ฑ์ค‘๊ณต์—…์—ฐ๊ตฌ์›

2018๋…„ ์ถฉ๋‚จ๋Œ€ํ•™๊ต ๋ฉ”์นดํŠธ๋กœ๋‹‰์Šค ๊ณตํ•™๊ณผ ๋ฐ•์‚ฌ

ํ˜„์žฌ ์ถฉ๋‚จ๋Œ€ํ•™๊ต ์ฒจ๋‹จ์ˆ˜์†ก์ฒด ์—ฐ๊ตฌ์›

์ • ์Šฌ (Seul Jung)
../../Resources/kiee/KIEE.2019.68.2.342/au2.png

1988๋…„ ๋ฏธ๊ตญ ์›จ์ธ ์ฃผ๋ฆฝ๋Œ€ ์ „๊ธฐ ๋ฐ ์ปดํ“จํ„ฐ ๊ณตํ•™๊ณผ ์กธ์—…

1991๋…„ ๋ฏธ๊ตญ ์บ˜๋ฆฌํฌ๋‹ˆ์•„๋Œ€ ๋ฐ์ด๋น„์Šค ์ „๊ธฐ ๋ฐ ์ปดํ“จํ„ฐ ๊ณตํ•™๊ณผ ์„์‚ฌ

1996๋…„ ๋™ ๋Œ€ํ•™ ๋ฐ•์‚ฌ ์กธ์—…

1997๋…„~ํ˜„์žฌ ์ถฉ๋‚จ๋Œ€ํ•™๊ต ๋ฉ”์นดํŠธ๋กœ๋‹‰์Šค๊ณตํ•™๊ณผ ๊ต์ˆ˜

๊ด€์‹ฌ๋ถ„์•ผ๋Š” ์ง€๋Šฅ์ œ์–ด ๋ฐ ์ง€๋Šฅ๋กœ๋ด‡ ์‹œ์Šคํ…œ, ๋ฐธ๋Ÿฐ์‹ฑ ์‹œ์Šคํ…œ, ์„œ๋น„์Šค ๋กœ๋ด‡, ์ž์ด๋กœ ๊ตฌ๋™๊ธฐ, ๋“œ๋ก , ๋กœ๋ด‡๊ต์œก, ์ž์œจ์ฃผํ–‰์ž๋™์ฐจ