• ๋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




Finite-time singular stability, Discrete-time singular systems, Linear matrix inequality, Parameter uncertainty

1. ์„œ ๋ก 

๊ธฐ์กด์˜ ๋ฆฌ์•„ํ‘ธ๋…ธํ”„ ์•ˆ์ •์„ฑ์€ ๋ฌดํ•œ์‹œ๊ฐ„์—์„œ ์‹œ์Šคํ…œ์˜ ์ ๊ทผ์  ์•ˆ์ •์„ฑ(asymptotic stability)์— ๊ด€์‹ฌ์„ ๊ฐ€์กŒ์œผ๋‚˜, ์‹ค์ œ ์‹œ์Šคํ…œ์˜ ์‘์šฉ๋ฌธ์ œ์— ์žˆ์–ด์„œ๋Š” ์ •ํ•ด์ง„ ์‹œ๊ฐ„์•ˆ์— ์‹œ์Šคํ…œ์˜ ๋™ํŠน์„ฑ ๋“ฑ์˜ ์ž‘๋™์„ ๋‹ค๋ฃจ๊ธฐ ๋•Œ๋ฌธ์— ์œ ํ•œ์‹œ๊ฐ„์— ๋Œ€ํ•œ ์ œ์–ด์‹œ์Šคํ…œ ํ•ด์„๊ณผ ์„ค๊ณ„๋ฌธ์ œ์— ๋Œ€ํ•œ ์—ฐ๊ตฌ๊ฐ€ ํ™œ๋ฐœํžˆ ์ง„ํ–‰๋˜๊ณ  ์žˆ๋‹ค(1). ๊ทธ๋Ÿฌ๋ฏ€๋กœ ์œ ํ•œ์‹œ๊ฐ„ ์•ˆ์ •์„ฑ์€ ์‹ค์ œ ์‹œ์Šคํ…œ์˜ ์„ฑ๋Šฅ์„ ๋‹ค๋ฃจ๋Š” ์ธก๋ฉด์—์„œ ๋”์šฑ ์ ์ ˆํ•˜๋‹ค๊ณ  ํ•  ์ˆ˜ ์žˆ๋‹ค. ์œ ํ•œ์‹œ๊ฐ„ ์•ˆ์ •์„ฑ(finite-time stability)์— ๋Œ€ํ•œ ์—ฐ๊ตฌ๋Š” 1950๋…„๋Œ€์— ๊ฐœ๋…์ด ์ฒ˜์Œ์œผ๋กœ ์†Œ๊ฐœ๋˜๋ฉด์„œ ๋ฏธ๋ฆฌ ์„ค์ •ํ•œ ์œ ๊ณ„(bound)์™€ ์œ ํ•œ์‹œ๊ฐ„ ๊ตฌ๊ฐ„์„ ๋‹ค๋ฃจ๋Š” ์—ฐ๊ตฌ๊ฒฐ๊ณผ๊ฐ€ ๋‚˜์™”๋‹ค(2). Amato ๋“ฑ(3,4)์€ ์—ฐ์†์‹œ๊ฐ„์—์„œ ๋ณ€์ˆ˜ ๋ถˆํ™•์‹ค์„ฑ์„ ๊ฐ€์ง€๋Š” ์„ ํ˜•์‹œ์Šคํ…œ์˜ ์œ ํ•œ์‹œ๊ฐ„ ๊ฐ•์ธ ์•ˆ์ •์„ฑ์„ ๋‹ค๋ฃจ์—ˆ๊ณ , ์ด์‚ฐ์‹œ๊ฐ„์—์„œ๋Š” ์ด์‚ฐ์‹œ๊ฐ„ ์‹œ์Šคํ…œ์— ๋Œ€ํ•œ ์œ ํ•œ์‹œ๊ฐ„ ์ œ์–ด์— ๋Œ€ํ•œ ํ•ด์„๊ณผ ์„ค๊ณ„์กฐ๊ฑด์— ๋Œ€ํ•œ ๊ฒฐ๊ณผ๋ฅผ ์ œ์‹œํ•˜์—ฌ ์—ฐ๊ตฌ๋ฒ”์œ„๋ฅผ ํ™•๋Œ€ํ•˜์˜€๋‹ค. ํ•˜์ง€๋งŒ, ๋‹ค๋ฃจ๋Š” ์‹œ์Šคํ…œ์ด ๋น„ํŠน์ด ์‹œ์Šคํ…œ์ด์—ˆ๋‹ค. ์ƒํƒœ๊ณต๊ฐ„ ๋ชจ๋ธ์€ ๋งค์šฐ ์œ ์šฉํ•˜์ง€๋งŒ ์ƒํƒœ ๋ณ€์ˆ˜๊ฐ€ ๋ชจ๋“  ๋ฌผ๋ฆฌ์  ์˜๋ฏธ๋ฅผ ํฌํ•จํ•˜์ง€๋Š” ๋ชปํ•œ๋‹ค. ๋”ฐ๋ผ์„œ, ํŠน์ดํ˜„์ƒ์€ ์„ ํ˜• ๋™์ ์‹œ์Šคํ…œ์˜ ์ž์—ฐ์Šค๋Ÿฌ์šด ํ˜•ํƒœ์ด๊ณ , ๋ฌผ๋ฆฌ์  ๋ณ€์ˆ˜๋“ค ์‚ฌ์ด์— ์กด์žฌํ•˜๋Š” ๋Œ€์ˆ˜ ์ œ์•ฝ์กฐ๊ฑด์„ ํ‘œํ˜„ํ•˜๋Š” ์ด๋ก ์ ์ธ ๋ฉด์ด๋‚˜ ์‹ค์šฉ์ ์ธ ๋ฉด์—์„œ ์ค‘์š”ํ•œ ๋™์  ์‹œ์Šคํ…œ์ด๋‹ค(5). ๋˜ํ•œ, ํŠน์ด์‹œ์Šคํ…œ์˜ ์•ˆ์ •์„ฑ๊ณผ ์ œ์–ด๋ฌธ์ œ๋Š” ํŠน์ด์‹œ์Šคํ…œ์˜ ํŠน๋ณ„ํ•œ ์„ฑ์งˆ๋กœ ์ธํ•˜์—ฌ ๋Œ€๊ทœ๋ชจ ์‹œ์Šคํ…œ, ํŠน์ด ์„ญ๋™ ์ด๋ก , ์ œ์•ฝ์กฐ๊ฑด์ด ์žˆ๋Š” ๊ธฐ๊ณ„์‹œ์Šคํ…œ ๋“ฑ์— ๊ด‘๋ฒ”์œ„ํ•˜๊ฒŒ ์ ์šฉ๋˜๊ธฐ ๋•Œ๋ฌธ์— ์ด์‚ฐ์‹œ๊ฐ„ ์˜์—ญ์—์„œ ์œ ํ•œ์‹œ๊ฐ„ ํŠน์ด์‹œ์Šคํ…œ์— ๋Œ€ํ•œ ์—ฐ๊ตฌ(6-10)๊ฐ€ ํ™œ๋ฐœํžˆ ์ง„ํ–‰๋˜๊ณ  ์žˆ๋‹ค.

์ด์‚ฐ์‹œ๊ฐ„ ํŠน์ด์‹œ์Šคํ…œ์— ๋Œ€ํ•œ ์œ ํ•œ์‹œ๊ฐ„ ์•ˆ์ •์„ฑ ํ•ด์„ ๋ฌธ์ œ์™€ ์œ ํ•œ์‹œ๊ฐ„ ์ƒํƒœ๊ถคํ™˜ ์ œ์–ด๊ธฐ ์„ค๊ณ„๋ฐฉ๋ฒ•์— ๋Œ€ํ•˜์—ฌ Antic ๋“ฑ(6)์ด ํ–‰๋ ฌ๋ถ€๋“ฑ์‹์˜ ์ถฉ๋ถ„์กฐ๊ฑด์„ ์ œ์‹œํ•˜์˜€๋‹ค. ๊ทธ๋Ÿฌ๋‚˜, ์ œ์–ด๊ธฐ ์„ค๊ณ„๋ฐฉ๋ฒ•์˜ ์ถฉ๋ถ„์กฐ๊ฑด์— ๋น„์„ ํ˜• ๋ณ€์ˆ˜๋“ค์ด ํฌํ•จ๋˜์–ด ์žˆ์–ด์„œ ๊ตฌํ•˜๊ณ ์ž ํ•˜๋Š” ๋ณ€์ˆ˜์˜ ์ธก๋ฉด์—์„œ ์„ ํ˜•ํ–‰๋ ฌ๋ถ€๋“ฑ์‹์˜ ํ˜•ํƒœ๊ฐ€ ์•„๋‹ˆ๋ฏ€๋กœ ์ตœ์ ํ™” ๋ฌธ์ œ๋ฅผ ํ•ด๊ฒฐํ•˜๊ธฐ ์–ด๋ ต๋‹ค๋Š” ๋ฌธ์ œ์ ์ด ์žˆ๋‹ค. Wo์™€ Han(7)์€ ์ด์‚ฐ์‹œ๊ฐ„ ์„ ํ˜• ํŠน์ด์‹œ์Šคํ…œ์— ๋Œ€ํ•œ ์œ ํ•œ์‹œ๊ฐ„ ์•ˆ์ •์„ฑ์„ ๋งŒ์กฑํ•˜๊ธฐ ์œ„ํ•œ ์–‘์˜ ์ค€์ •๋ถ€ํ˜ธ(positive semidefinite) ์•ˆ์ •ํ™” ์กฐ๊ฑด์„ ์ œ์‹œํ•˜์˜€์ง€๋งŒ, ๋“ฑํ˜ธ๋ฅผ ํฌํ•จํ•˜๋Š” ์กฐ๊ฑด์€ ํ•ด๋ฅผ ๊ตฌํ•˜๊ธฐ ์‰ฝ์ง€ ์•Š๋‹ค๋Š” ๋‹จ์ ์ด ์žˆ๋‹ค. ๋˜ํ•œ, Ma ๋“ฑ(8)์€ ์‹œ๊ฐ„์ง€์—ฐ๊ณผ ๊ตฌ๋™๊ธฐ ํฌํ™”๋ฅผ ๊ฐ€์ง€๋Š” ์ด์‚ฐ์‹œ๊ฐ„ ๋งˆ์ฝ”ํ”„ ์ ํ”„ ํŠน์ด์‹œ์Šคํ…œ์— ๋Œ€ํ•œ ๊ฐ•์ธ ์œ ํ•œ์‹œ๊ฐ„ Hโˆž์ œ์–ด๊ธฐ ์„ค๊ณ„๋ฐฉ๋ฒ•์„ ์ œ์‹œํ•˜์˜€์œผ๋‚˜, ์ œ์–ด๊ธฐ์˜ ์กด์žฌ ์กฐ๊ฑด์— ๋“ฑํ˜ธ๊ฐ€ ํฌํ•จ๋œ ์ค€์ •๋ถ€ํ˜ธ ํ–‰๋ ฌ๋ถ€๋“ฑ์‹ ํ˜•ํƒœ์ด๋ฏ€๋กœ ์ตœ์ ํ™” ๋ฌธ์ œ๋ฅผ ํ•ด๊ฒฐํ•˜๊ธฐ ์‰ฝ์ง€ ์•Š์€ ์ ์ด ์žˆ๋‹ค. ๊ทธ๋ฆฌ๊ณ , Ma ๋“ฑ(9)์€ ์ด์‚ฐ์‹œ๊ฐ„ ๋งˆ์ฝ”ํ”„ ์ ํ”„ ํŠน์ด์‹œ์Šคํ…œ์— ๋Œ€ํ•œ ์œ ํ•œ์‹œ๊ฐ„ ์‚ฐ์ผ์„ฑ ์ œ์–ด๊ธฐ(dissipative controller) ์„ค๊ณ„ ๋ฐฉ๋ฒ•์„ ์ œ์‹œํ•˜์˜€์œผ๋‚˜, ์ถฉ๋ถ„์กฐ๊ฑด์— ๋“ฑํ˜ธ๊ฐ€ ํฌํ•จ๋œ ์Œ์˜ ์ค€์ •๋ถ€ํ˜ธ ์กฐ๊ฑด์ด ํฌํ•จ๋˜์–ด ์žˆ์–ด์„œ ์ˆ˜์น˜์ ์œผ๋กœ ํ•ด๋ฅผ ๊ตฌํ•˜๊ธฐ ์‰ฝ์ง€ ์•Š๋‹ค. ์ตœ๊ทผ, Wang ๋“ฑ(10)์€ ํ•ด๊ฐ€ ์กด์žฌํ•˜๊ธฐ ์œ„ํ•œ ์กฐ๊ฑด์—์„œ ๋“ฑํ˜ธ๊ฐ€ ํฌํ•จ๋œ ์ค€์ •๋ถ€ํ˜ธ ํ–‰๋ ฌ๋ถ€๋“ฑ์‹์˜ ๋ฌธ์ œ์ (6-9)์„ ํ•ด๊ฒฐํ•˜๊ธฐ ์œ„ํ•˜์—ฌ ๋น ๋ฅธ ๋ถ€์‹œ์Šคํ…œ(fast sub- system)๊ณผ ๋Š๋ฆฐ ๋ถ€์‹œ์Šคํ…œ(slow subsystem) ์‚ฌ์ด์˜ ๋Œ€์ˆ˜๊ด€๊ณ„๋ฅผ ํ‘œํ˜„ํ•˜๋Š” ์ถ”๊ฐ€์ ์ธ ํ–‰๋ ฌ(additional matrix)์„ ์‚ฌ์šฉํ•˜์—ฌ ์ด์‚ฐ์‹œ๊ฐ„ ํŠน์ด์‹œ์Šคํ…œ์— ๋Œ€ํ•œ ๊ฐ•์ธ ์œ ํ•œ์‹œ๊ฐ„ ์•ˆ์ •์„ฑ ์กฐ๊ฑด๊ณผ ์ƒํƒœ๊ถคํ™˜ ์ œ์–ด๊ธฐ ์„ค๊ณ„๋ฐฉ๋ฒ•์„ ๋‹ค๋ฃจ์—ˆ๋‹ค. ์ด์‚ฐ์‹œ๊ฐ„ ํŠน์ด์‹œ์Šคํ…œ์ด ์ •๊ทœ์ (regular)์ด๊ณ  ์ธ๊ณผ์ (causal)์ด๋ฉฐ ์œ ํ•œ์‹œ๊ฐ„ ์•ˆ์ •์ด ๋˜๊ธฐ ์œ„ํ•œ ์กฐ๊ฑด์„ ๊ตฌํ•˜๋Š” ๋ณ€์ˆ˜์˜ ์ธก๋ฉด์—์„œ ๋ณผ๋ก์ตœ์ ํ™”(convex optimiza- tion)๊ฐ€ ๊ฐ€๋Šฅํ•œ ์„ ํ˜•ํ–‰๋ ฌ๋ถ€๋“ฑ์‹์œผ๋กœ ์ œ์‹œํ•˜์˜€๋‹ค. ๊ทธ๋Ÿฌ๋‚˜, ๊ฐ•์ธ ์œ ํ•œ์‹œ๊ฐ„ ํŠน์ด์•ˆ์ •ํ™”ํ•˜๊ฒŒ ํ•˜๋Š” ์ƒํƒœ๊ถคํ™˜ ์ œ์–ด๊ธฐ๋ฅผ ๊ตฌํ•˜๋Š” ์ถฉ๋ถ„์กฐ๊ฑด์—์„œ ์ œ์–ด๊ธฐ์˜ ๋ช…ํ™•ํ•œ ํ˜•ํƒœ๋Š” ์ œ์‹œํ•˜์˜€์œผ๋‚˜ ๊ตฌํ•˜๋ ค๋Š” ๋ณ€์ˆ˜์˜ ๊ฒฌ์ง€์—์„œ ๋ณผ๋ก์ตœ์ ํ™”๋กœ ํ‘œํ˜„๋˜์ง€ ์•Š์•„์„œ ํ•ด๋ฅผ ๊ตฌํ•˜๊ธฐ ์‰ฝ์ง€ ์•Š์•˜๋‹ค. ๋˜ํ•œ, ์ œ์•ˆํ•œ ์ œ์–ด๊ธฐ์˜ ํ˜•ํƒœ๊ฐ€ ์ง€๋ฃจํ•œ ๊ณผ์ •(tedious procedure)์ด ํ•„์š”ํ•˜๊ณ  ๊ตฌํ•˜๋ ค๋Š” ๋ณ€์ˆ˜์˜ ํ•ด๋ฅผ ์–ป๊ธฐ๊ฐ€ ์–ด๋ ค์› ๋‹ค. ๋”ฐ๋ผ์„œ, ๋ณธ ๋…ผ๋ฌธ์˜ ๋ชฉ์ ์€ Wang ๋“ฑ(10)์ด ๋‹ค๋ฃจ์—ˆ๋˜ ๋ณ€์ˆ˜ ๋ถˆํ™•์‹ค์„ฑ์„ ๊ฐ€์ง€๋Š” ์ด์‚ฐ์‹œ๊ฐ„ ํŠน์ด์‹œ์Šคํ…œ์— ๋Œ€ํ•ด, ์œ ํ•œ์‹œ๊ฐ„ ํŠน์ด์•ˆ์ •์„ฑ ์กฐ๊ฑด๊ณผ ์ฃผ์–ด์ง„ ์‹œ์Šคํ…œ์ด ์ •๊ทœ์ (regular)์ด๊ณ  ์ธ๊ณผ์ (causal)์ด๋ฉฐ ๊ฐ•์ธ ์œ ํ•œ์‹œ๊ฐ„ ํŠน์ด์•ˆ์ •์„ฑ์„ ๋ณด์žฅํ•˜๋Š” ์ƒํƒœ๊ถคํ™˜ ์ œ์–ด๊ธฐ ์„ค๊ณ„๋ฐฉ๋ฒ•์„ ๊ตฌํ•˜๋ ค๋Š” ๋ณ€์ˆ˜์˜ ์ธก๋ฉด์—์„œ ๋ณผ๋ก์ตœ์ ํ™”๊ฐ€ ๊ฐ€๋Šฅํ•œ ์„ ํ˜•ํ–‰๋ ฌ๋ถ€๋“ฑ์‹ ๊ธฐ๋ฒ•์œผ๋กœ ํ‘œํ˜„ํ•˜์—ฌ ๊ธฐ์กด์˜ ๋ฌธ์ œ์ ์„ ๊ทน๋ณตํ•˜๋Š” ๊ฒƒ์ด๋‹ค. ์ด๋ฅผ ์œ„ํ•˜์—ฌ ๋ณธ ๋…ผ๋ฌธ์—์„œ๋Š” ๋จผ์ €, ๋“ฑ๊ฐ€์ ์œผ๋กœ ๋ณ€ํ˜•ํ•œ ํ๋ฃจํ”„ ์‹œ์Šคํ…œ์„ ์ด์šฉํ•˜์—ฌ ์ •๊ทœ์„ฑ, ์ธ๊ณผ์„ฑ, ์œ ํ•œ์‹œ๊ฐ„ ํŠน์ด์•ˆ์ •์„ฑ์„ ๋งŒ์กฑํ•˜๋Š” ์ƒˆ๋กœ์šด ์ถฉ๋ถ„์กฐ๊ฑด์„ ์ œ์‹œํ•œ๋‹ค. ๊ทธ๋ฆฌ๊ณ , ๊ตฌํ•œ ์ถฉ๋ถ„์กฐ๊ฑด๊ณผ ๋‹ค๋ฃจ๋Š” ํŠน์ด์‹œ์Šคํ…œ์˜ ๋“ฑ๊ฐ€ ์„ฑ์งˆ(equivalent property)์„ ์ด์šฉํ•˜์—ฌ ๊ฐ•์ธ ์œ ํ•œ์‹œ๊ฐ„ ์ƒํƒœ๊ถคํ™˜ ์ œ์–ด๊ธฐ ์„ค๊ณ„๋ฐฉ๋ฒ•์„ ์ œ์•ˆํ•œ๋‹ค.

๋ณธ ๋…ผ๋ฌธ์—์„œ๋Š” ์ด์‚ฐ์‹œ๊ฐ„ ๋ถˆํ™•์‹ค ํŠน์ด์‹œ์Šคํ…œ์— ๋Œ€ํ•˜์—ฌ ์ •๊ทœ์ ์ด๊ณ  ์ธ๊ณผ์ ์ด๋ฉฐ ์œ ํ•œ์‹œ๊ฐ„ ํŠน์ด์•ˆ์ •์„ฑ์„ ๋งŒ์กฑํ•˜๋Š” ์ถฉ๋ถ„์กฐ๊ฑด์„ ์ƒˆ๋กœ์šด ์ ‘๊ทผ๋ฐฉ์‹์œผ๋กœ ์ œ์•ˆํ•œ๋‹ค. ๋˜ํ•œ, ๊ตฌํ•œ ์กฐ๊ฑด์œผ๋กœ๋ถ€ํ„ฐ ๊ฐ•์ธ ์œ ํ•œ์‹œ๊ฐ„ ํŠน์ด์•ˆ์ •์„ฑ์„ ๋ณด์žฅํ•˜๋Š” ์ƒํƒœ๊ถคํ™˜ ์ œ์–ด๊ธฐ ์„ค๊ณ„๋ฐฉ๋ฒ•์„ ๊ตฌํ•˜๋ ค๋Š” ๋ณ€์ˆ˜์˜ ์ธก๋ฉด์—์„œ ํ•œ๋ฒˆ์— ํ•ด๋ฅผ ๊ตฌํ•  ์ˆ˜ ์žˆ๋Š” ์ถฉ๋ถ„์กฐ๊ฑด์˜ ํ˜•ํƒœ๋กœ ์ œ์‹œํ•œ๋‹ค. ์ œ์•ˆํ•œ ์ œ์–ด๊ธฐ ์„ค๊ณ„๋ฐฉ๋ฒ•์˜ ํƒ€๋‹น์„ฑ์„ ํ™•์ธํ•˜๊ธฐ ์œ„ํ•˜์—ฌ ๋ถˆ์•ˆ์ •ํ•œ ๊ฐœ๋ฃจํ”„ ์‹œ์Šคํ…œ์„ ๊ฐ€์ง€๋Š” ์ˆ˜์น˜ ์˜ˆ์ œ๋ฅผ ๋‹ค๋ฃฌ๋‹ค. ์ œ์•ˆํ•œ ๋ฐฉ๋ฒ•์€ ๋‹ค์–‘ํ•œ ์ œ์–ด๊ธฐ์™€ ํ•„ํ„ฐ ์„ค๊ณ„๋ฐฉ๋ฒ•์œผ๋กœ ํ™•์žฅ ๊ฐ€๋Šฅํ•˜๋‹ค.

๋ณธ ๋…ผ๋ฌธ์—์„œ ์‚ฌ์šฉํ•˜๋Š” ํ‘œ๊ธฐ๋Š” ์ผ๋ฐ˜์ ์ธ ๊ธฐํ˜ธ๋ฅผ ์‚ฌ์šฉํ•œ๋‹ค. $I$, $0$๊ณผ $bold R^{r}$์€ ์ ์ ˆํ•œ ์ฐจ์›์„ ๊ฐ€์ง€๋Š” ๋‹จ์œ„ํ–‰๋ ฌ, ์˜ํ–‰๋ ฌ๊ณผ $r\times 1$ ์ฐจ์›์„ ๊ฐ€์ง€๋Š” ์‹ค์ˆ˜ ๋ฒกํ„ฐ๋ฅผ ๊ฐ๊ฐ ์˜๋ฏธํ•œ๋‹ค. $P>0$์€ ์–‘์˜ ์ •๋ถ€ํ˜ธ ํ–‰๋ ฌ(positive definite matrix)์ด๊ณ , $Ex(k+1)=Ax(k)$๋Š” $(E,\:A)$๋กœ ํ‘œํ˜„ํ•œ๋‹ค. $\ast$๋Š” ๋Œ€์นญํ–‰๋ ฌ(symmetric matrix)์˜ ์ฃผ ๋Œ€๊ฐ์„  ์•„๋ž˜์— ๋†“์ด๋Š” ์š”์†Œ, $diag\{\bullet\}$๋Š” ์ฃผ ๋Œ€๊ฐ์„ ์—๋งŒ ๊ฐ’์„ ๊ฐ€์ง€๋Š” ๋Œ€๊ฐํ–‰๋ ฌ(diagonal matrix), $< X > =X + X^{T}$, $\lambda_{\max}(\bullet)$๋Š” ๊ฐ€์žฅ ํฐ ๊ณ ์œ ๊ฐ’์ด๊ณ  $\lambda_{\min}(\bullet)$๋Š” ๊ฐ€์žฅ ์ž‘์€ ๊ณ ์œ ๊ฐ’์„ ์˜๋ฏธํ•œ๋‹ค.

2. ๊ฐ•์ธ ์œ ํ•œ์‹œ๊ฐ„ Hโˆž ์ƒํƒœ๊ถคํ™˜ ์ œ์–ด

๋ณ€์ˆ˜ ๋ถˆํ™•์‹ค์„ฑ์„ ๊ฐ€์ง€๋Š” ์ด์‚ฐ์‹œ๊ฐ„ ํŠน์ด์‹œ์Šคํ…œ

(1)
\begin{align*} Ex(k+1)& = &(A +\Delta A(k))x(k)+(B+\Delta B(k))u(k) \end{align*}

์„ ๋‹ค๋ฃฌ๋‹ค. ์—ฌ๊ธฐ์„œ, $x(k)\in bold R^{n}$๋Š” ์ƒํƒœ๋ณ€์ˆ˜, $u(k)\in bold R^{m}$๋Š” ์ œ์–ด์ž…๋ ฅ, $E$๋Š” $rank(E)=r\le n$์„ ๋งŒ์กฑํ•˜๋Š” ํŠน์ดํ–‰๋ ฌ(singular matrix)์ด๊ณ , ๋ชจ๋“  ์‹œ์Šคํ…œ ํ–‰๋ ฌ์€ ์ ์ ˆํ•œ ์ฐจ์›์„ ๊ฐ€์ง„๋‹ค. $\Delta A(k)$์™€ $\Delta B(k)$๋Š”

(2)
$\Delta A(k)=M_{a}F(k)N_{a}$, $\Delta B(k)=M_{b}F(k)N_{b}$

๋ฅผ ๋งŒ์กฑํ•˜๋Š” ๋ชจ๋ฅด๋Š” ํ–‰๋ ฌ์ด๋‹ค. ์—ฌ๊ธฐ์„œ, $M_{a}$, $M_{b}$, $N_{a}$, $N_{b}$๋Š” ์ ์ ˆํ•œ ์ฐจ์›์„ ๊ฐ€์ง€๋Š” ์ƒ์ˆ˜ํ–‰๋ ฌ์ด๊ณ , $F(k)$๋Š” $F^{T}(k)F(k)\le I$์— ์˜ํ•ด ์œ ๊ณ„๋˜๋Š” ๋ชจ๋ฅด๋Š” ํ–‰๋ ฌ์ด๋‹ค. ๋ณธ ๋…ผ๋ฌธ์˜ ๋ชฉ์ ์€ ๋ณ€์ˆ˜ ๋ถˆํ™•์‹ค์„ฑ์„ ๊ฐ€์ง€๋Š” ์ด์‚ฐ์‹œ๊ฐ„ ํŠน์ด์‹œ์Šคํ…œ (1)์— ๋Œ€ํ•˜์—ฌ ์ •๊ทœ์ ์ด๊ณ  ์ธ๊ณผ์ ์ด๋ฉฐ ๊ฐ•์ธ ์œ ํ•œ์‹œ๊ฐ„ ํŠน์ด์•ˆ์ •์„ฑ์„ ๋งŒ์กฑํ•˜๋Š” ์ƒํƒœ๊ถคํ™˜ ์ œ์–ด๊ธฐ

(3)
$u(k)=Kx(k)$ (3)

์„ ์„ค๊ณ„ํ•˜๋Š” ๊ฒƒ์ด๋‹ค.

์ •์˜ 1(11): ์ด์‚ฐ์‹œ๊ฐ„ ํŠน์ด์‹œ์Šคํ…œ $(E,\:A)$์— ๋Œ€ํ•˜์—ฌ,

(i) $\det(z E-A)$์ด ํ•ญ๋“ฑ์ ์œผ๋กœ ์˜(identically zero)์ด ์•„๋‹ˆ๋ฉด, ์ด์‚ฐ์‹œ๊ฐ„ ํŠน์ด์‹œ์Šคํ…œ์€ ์ •๊ทœ์ (regular)์ด๋‹ค.

(ii) $rank(E)=\deg(\det(z E-A))$์ด๋ฉด, ์ด์‚ฐ์‹œ๊ฐ„ ํŠน์ด์‹œ์Šคํ…œ $(E,\:A)$๋Š” ์ธ๊ณผ์ (causal)์ด๋‹ค.

์ •์˜ 2(11): (์œ ํ•œ์‹œ๊ฐ„ ํŠน์ด์•ˆ์ •์„ฑ(Finite-time Singular Stability)) ์ด์‚ฐ์‹œ๊ฐ„ ํŠน์ด์‹œ์Šคํ…œ $(E,\:A)$๊ฐ€

(4)
$x^{T}(0)E^{T}REx(0)\le c_{1}\Rightarrow x^{T}(k)E^{T}R E x(k)< c_{2},\:\forall k\in\{1,\:2,\:\cdots ,\: N\}$

๋ฅผ ๋งŒ์กฑํ•˜๋ฉด, ์œ ํ•œ์‹œ๊ฐ„ ํŠน์ด์•ˆ์ •ํ•˜๋‹ค.

์ •์˜ 3(10): (๊ฐ•์ธ ์œ ํ•œ์‹œ๊ฐ„ ํŠน์ด์•ˆ์ •์„ฑ(Robust Finite-time Singular Stability)) ์ด์‚ฐ์‹œ๊ฐ„ ๋ถˆํ™•์‹ค ํŠน์ด์‹œ์Šคํ…œ (1)์ด

(5)
$x^{T}(0)E^{T}REx(0)\le c_{1}\Rightarrow x^{T}(k)E^{T}REx(k)<c_{2},\:\forall k\in\{1,\:2,\:\cdots ,\:N\}$

๋ฅผ ๋งŒ์กฑํ•˜๋ฉด, ๊ฐ•์ธ ์œ ํ•œ์‹œ๊ฐ„ ํŠน์ด์•ˆ์ •ํ•˜๋‹ค.

์ •์˜ 2์™€ ์ •์˜ 3์—์„œ ์–‘์˜ ์‹ค์ˆ˜ $c_{1}$๊ณผ $c_{2}$๋Š” $c_{1}<c_{2}$๋ฅผ ๋งŒ์กฑํ•˜๊ณ , $R$์€ ์–‘์˜ ์ •๋ถ€ํ˜ธ ํ–‰๋ ฌ์ด๊ณ  $N$์€ ์ฃผ์–ด์ง„ ์–‘์˜ ์ •์ˆ˜์ด๋‹ค. ์•„๋ž˜ ์ •๋ฆฌ 1์—์„œ๋Š” ๊ณต์นญ์‹œ์Šคํ…œ(nominal system) $(E,\:A)$์— ๋Œ€ํ•ด ์ •์˜ 1๊ณผ ์ •์˜ 2๋ฅผ ๋งŒ์กฑํ•˜๋Š” ์œ ํ•œ์‹œ๊ฐ„ ํŠน์ด์•ˆ์ •์„ฑ ์กฐ๊ฑด์„ ์ œ์‹œํ•œ๋‹ค.

์ •๋ฆฌ 1: ์ฃผ์–ด์ง„ ์–‘์˜ ์‹ค์ˆ˜ $c_{2}>c_{1}$, $\alpha >1$, ์–‘์˜ ์ •์ˆ˜ $N$๊ณผ $R>0$์— ๋Œ€ํ•˜์—ฌ, ์•„๋ž˜์˜ ์„ ํ˜•ํ–‰๋ ฌ๋ถ€๋“ฑ์‹

(6)
$\begin{bmatrix}\Lambda_{1}&(A-E)^{T}X +Z\Phi^{T}-X &E^{T}P &(A-E)^{T}X \\ \ast &-2X&P &-X \\ \ast &\ast &-P&0\\ \ast &\ast &\ast &-X\end{bmatrix}<0$

(7)
$\theta I <\widetilde P <I$, $0<\theta <1$

(8)
$\alpha^{N}c_{1}-\theta c_{2}<0$

์„ ๋งŒ์กฑํ•˜๋Š” ์–‘์˜ ์ •๋ถ€ํ˜ธ ํ–‰๋ ฌ $P$, $X$, ํ–‰๋ ฌ $Z$์™€ ์–‘์˜ ์ƒ์ˆ˜ $\theta$๊ฐ€ ์กด์žฌํ•˜๋ฉด, $(E,\:A)$๊ฐ€ ์ •์˜ 1๊ณผ ์ •์˜ 2๋ฅผ ๋งŒ์กฑํ•˜๋Š” ์œ ํ•œ์‹œ๊ฐ„ ํŠน์ด์•ˆ์ •ํ•˜๋‹ค. ์—ฌ๊ธฐ์„œ, $\Lambda_{1}=\left <(A-E)^{T}X\right > -\alpha E^{T}PE$, $\Phi$๋Š” $E^{T}\Phi =0$์„ ๋งŒ์กฑํ•˜๋Š” ํ–‰๋ ฌ์ด๊ณ , $\widetilde P =R^{-1/2}P R^{-1/2}$์ด๋‹ค.

์ฆ๋ช…: ๋ณ€์ˆ˜ $y(k)=x(k+1)-x(k)$๋กœ ์„ค์ •ํ•˜๋ฉด

(9)
$\bar{E}\bar{x}(k+1)=\bar{A}\bar{x}(k)$

๊ฐ€ ๋˜๊ณ , ์—ฌ๊ธฐ์„œ, $\bar{E}=\begin{bmatrix}E & 0\\0& 0\end{bmatrix}$, $\bar{A}=\begin{bmatrix}E& I\\A-E& -I\end{bmatrix}$, $\bar{x}(k)=\begin{bmatrix}x(k)\\Ey(k)\end{bmatrix}$์ด๋‹ค. ๋ฆฌ์•„ํ‘ธ๋…ธํ”„ ํ•จ์ˆ˜ $V(\bar{x}(k))=\bar{x}^{T}(k)\bar{E}^{T}\bar{P}\bar{E}\bar{x}(k)$๋ฅผ ๋‘๊ณ  $\Delta V(\bar{x}(k))-(\alpha -1)V(\bar{x}(k))< 0$์ด๋ ค๋ฉด, $\alpha >1$์ด๋ฏ€๋กœ

(10)
$V(\bar{x}(k+1))-\alpha V(\bar{x}(k))<0,\: \forall k\in\{0,\:1,\:\cdots ,\:N\}$

์ด ๋˜๊ณ , $\Delta V(\bar{x}(k))$๋Š” $V(\bar{x}(k))$์˜ ์ „๋ฐฉํ–ฅ ์ฐจ๋ถ„(forward difference)์ด๋‹ค. ๋˜ํ•œ, $\bar{E}^{T}\bar{\Phi}=0$์œผ๋กœ ๋‘๋ฉด

(11)
$2\bar{x}^{T}(k+1)\bar{E}^{T}\bar{\Phi}\bar{Z}^{T}\bar{x}(k)=0$

$2\bar{x}^{T}(k+1)\bar{E}^{T}\bar{\Phi}\bar{Z}^{T}\bar{x}(k)=0$ ์ด๊ณ , ์‹(9)-(11)์—์„œ $\bar{x}^{T}(k)\left(\bar{A}^{T}\bar{P}\bar{A}-\alpha\bar{E}^{T}\bar{P}\bar{E}+\left <\bar{A}^{T}\bar{\Phi}\bar{Z}^{T}\right >\right)x(k)<0$์ด๋ฏ€๋กœ

(12)
$\bar{A}^{T}\bar{P}\bar{A}-\alpha\bar{E}^{T}\bar{P}\bar{E}+\left <\bar{A}^{T}\bar{\Phi}\bar{Z}^{T}\right > <0$

๋ฅผ ๋งŒ์กฑํ•˜์—ฌ์•ผ ํ•œ๋‹ค. ์‹(9)๊ฐ€ ์ •๊ทœ์ ์ด๊ณ  ์ธ๊ณผ์ ์ธ ๊ฒƒ์€ $(E,\:A)$๊ฐ€ ์ •๊ทœ์ ์ด๊ณ  ์ธ๊ณผ์ ์ธ ๊ฒƒ๊ณผ ๋™์ผํ•œ ๊ฒƒ์€ ์ •์˜ 1๋กœ๋ถ€ํ„ฐ ์ง์ ‘ ๋ณด์ผ ์ˆ˜ ์žˆ์œผ๋ฏ€๋กœ, ๋จผ์ € ์‹(9)๊ฐ€ ์ •๊ทœ์ ์ด๊ณ  ์ธ๊ณผ์ ์ž„์„ ์ฆ๋ช…ํ•œ๋‹ค. $\bar{E}$๊ฐ€ ํŠน์ดํ–‰๋ ฌ์ด๋ฏ€๋กœ $U\bar{E}V =\begin{bmatrix}I_{r}&0\\0&0\end{bmatrix}$, $U\bar{A}V =\begin{bmatrix}\bar{A}_{11}&\bar{A}_{12}\\\bar{A}_{21}&\bar{A}_{22}\end{bmatrix}$,$U^{-T}\bar{P}U^{-1}=\begin{bmatrix}\bar{P}_{11}&\bar{P}_{12}\\\ast &\bar{P}_{22}\end{bmatrix}$, $V^{T}\bar{Z}=\begin{bmatrix}\bar{Z_{1}}\\\bar{Z_{2}}\end{bmatrix}$, $U^{T}\bar{\Phi}=\begin{bmatrix}0\\\bar{\Phi}_{2}\end{bmatrix}$๋ฅผ ๋งŒ์กฑํ•˜๋Š” ๋น„ํŠน์ดํ–‰๋ ฌ $U$์™€ $V$๊ฐ€ ์กด์žฌํ•œ๋‹ค. ์‹(12)์˜ ์ขŒ์ธก๊ณผ ์šฐ์ธก์— $U^{T}$์™€ $U$๋ฅผ ๊ณฑํ•ด์ฃผ๋ฉด, $\begin{bmatrix}\star &\star \\\star &\left <\bar{A}_{22}^{T}\bar{\Phi}_{2}^{T}\bar{Z}_{2}^{T}\right >\end{bmatrix}<0$์„ ๋งŒ์กฑํ•˜์—ฌ์•ผ ํ•˜๊ณ , $\star$๋Š” ์ฆ๋ช…์—์„œ ํ•„์š” ์—†๋Š” ๋ถ€๋ถ„์„ ์˜๋ฏธํ•œ๋‹ค. ๋”ฐ๋ผ์„œ, Xu์™€ Lam(12)์˜ ๊ฒฐ๊ณผ๋กœ๋ถ€ํ„ฐ $\bar{A}_{22}$๊ฐ€ ๋น„ํŠน์ด ํ–‰๋ ฌ(nonsingular matrix)์ด๋ฉด ์‹(9)๊ฐ€ ์ •๊ทœ์ ์ด๊ณ  ์ธ๊ณผ์ ์ž„์„ ๋ณด์ผ ์ˆ˜ ์žˆ๋‹ค. ์‹(12)์—์„œ ์Šˆ์–ด ์—ฌ์ˆ˜(Schur complement) ์ •๋ฆฌ(13)๋ฅผ ์ด์šฉํ•˜๋ฉด

(13)
$\begin{bmatrix}\left <\bar{A}^{T}\bar{\Phi}\bar{Z}^{T}\right > -\alpha\bar{E}^{T}\bar{P}\bar{E}&\bar{A}^{T}\bar{P}\\\ast &-\bar{P}\end{bmatrix}<0$

์ด ๋œ๋‹ค. ๋ณ€์ˆ˜๋“ค์„ $\bar{P}=\begin{bmatrix}P& 0\\\ast & X\end{bmatrix}$, $\bar{\Phi}=\begin{bmatrix}\Phi & 0\\\ast & X\end{bmatrix}$, $\bar{Z}=\begin{bmatrix}Z & I \\0& I\end{bmatrix}$๋กœ ๋‘๋ฉด, ์‹(6)์ด ๋œ๋‹ค. $V(\bar{x}(k))=\bar{x}^{T}(k)\bar{E}^{T}\bar{P}\bar{E}\bar{x}(k)$์—์„œ ์ •์˜ํ•œ ๋ณ€์ˆ˜๋ฅผ ๋Œ€์ž…ํ•˜๋ฉด, $V(\bar{x}(k))=x^{T}(k)E^{T}P Ex(k)\equiv V(x(k))$๊ฐ€ ๋œ๋‹ค. $\Delta V(x(k))-(\alpha -1)V(x(k))< 0$๋กœ๋ถ€ํ„ฐ

(14)
$V(x(k))<\alpha V(x(k-1))<\alpha^{2}V(x(k-2))<\cdots <\alpha^{k}V(x(0))$

๋ฅผ ์œ ์ถ”ํ•  ์ˆ˜ ์žˆ๋‹ค. ๋˜ํ•œ, ์„ค์ •ํ•œ ๋ฆฌ์•„ํ‘ธ๋…ธํ”„ ํ•จ์ˆ˜์˜ ์ดˆ๊ธฐ๊ฐ’์€

(15)
$V(x(0))=x^{T}(0)E^{T}PEx(0)\le c_{1}\lambda_{\max}(\widetilde P)$

๊ฐ€ ๋˜๊ณ , ์‹(14), (15)์™€ $V(x(k))\ge\lambda_{\min}(\widetilde P)x^{T}(k)E^{T}REx(k)$๋กœ๋ถ€ํ„ฐ

(16)
$x^{T}(k)E^{T}REx(k)\le\alpha^{k}c_{1}\dfrac{\lambda_{\max}(\widetilde P)}{\lambda_{\min}(\widetilde P)}<c_{2}$

์˜ ๊ด€๊ณ„๋ฅผ ๊ตฌํ•  ์ˆ˜ ์žˆ๋‹ค. ์—ฌ๊ธฐ์„œ, ์‹(7)๊ณผ (8)์€ Wang ๋“ฑ(10)์ด ์ „๊ฐœํ•œ ๋‚ด์šฉ์ฒ˜๋Ÿผ ์ผ๋ฐ˜์„ฑ์„ ์žƒ์ง€ ์•Š๊ณ (without loss of generality) ์‹(7)์„ ๊ฐ€์ •ํ•˜๋ฉด, $\theta <\lambda_{\min}(\widetilde P)<\lambda_{\max}(\widetilde P)<1$์ด ๋œ๋‹ค. ๊ทธ๋Ÿฌ๋ฏ€๋กœ, ์‹(16)์—์„œ $\alpha^{N}c_{1}\lambda_{\max}(\widetilde P)<\alpha^{N}c_{1}$์ด ๋˜๊ณ , $\theta c_{2}<\lambda_{\min}(\widetilde P)c_{2}$๊ฐ€ ๋˜๊ธฐ ๋•Œ๋ฌธ์— ์‹(8)์„ ๊ตฌํ•  ์ˆ˜ ์žˆ๋‹ค. ๋”ฐ๋ผ์„œ, ์ •๋ฆฌ 1์„ ๋งŒ์กฑํ•˜๋Š” ํ•ด๊ฐ€ ์กด์žฌํ•˜๋ฉด ์ด์‚ฐ์‹œ๊ฐ„ ํŠน์ด์‹œ์Šคํ…œ $(E,\:A)$๋Š” ์ •์˜ 1๊ณผ ์ •์˜ 2๋ฅผ ๋งŒ์กฑํ•˜๋Š” ์œ ํ•œ์‹œ๊ฐ„ ํŠน์ด์•ˆ์ •ํ•˜๋‹ค.

์ •๋ฆฌ 1์—์„œ ๊ตฌํ•œ ์œ ํ•œ์‹œ๊ฐ„ ํŠน์ด์•ˆ์ •์„ฑ ์กฐ๊ฑด์„ ๊ธฐ๋ฐ˜์œผ๋กœ ์ด์‚ฐ์‹œ๊ฐ„ ๋ถˆํ™•์‹ค ํŠน์ด์‹œ์Šคํ…œ (1)์— ๋Œ€ํ•ด ๊ฐ•์ธ ์œ ํ•œ์‹œ๊ฐ„ ํŠน์ด์•ˆ์ •์„ฑ์„ ๋งŒ์กฑํ•˜๋Š” ์‹(3)์˜ ์ƒํƒœ๊ถคํ™˜ ์ œ์–ด๊ธฐ ์„ค๊ณ„๋ฐฉ๋ฒ•์„ ์ •๋ฆฌ 2์—์„œ ์ œ์•ˆํ•œ๋‹ค.

์ •๋ฆฌ 2: ์ฃผ์–ด์ง„ ์–‘์˜ ์‹ค์ˆ˜ $c_{2}>c_{1}$, $\alpha >1$, ์–‘์˜ ์ •์ˆ˜ $N$๊ณผ $R>0$์— ๋Œ€ํ•˜์—ฌ, ์•„๋ž˜์˜ ์„ ํ˜•ํ–‰๋ ฌ๋ถ€๋“ฑ์‹

(17)
$\begin{bmatrix}\Omega_{1}&\Omega_{2}&EP&(A-E)X +BY &(N_{a}X)^{T}&(N_{b}Y)^{T\\ \ast}&-2X&P&-X&(N_{a}X)^{T}&(N_{b}Y)^{T\\ \ast}&\ast &-P&0&0&0\\ \ast &\ast &\ast &-X&(N_{a}X)^{T}&(N_{b}Y)^{T\\\ast}&\ast &\ast &\ast &-\beta_{1}I &0\\\ast &\ast &\ast &\ast &\ast &-\beta_{2}I \end{bmatrix}<0$

(18)
$\theta I <\widetilde P <I$, $0<\theta <1$

(19)
$\alpha^{N}c_{1}-\theta c_{2}<0$

๋ฅผ ๋งŒ์กฑํ•˜๋Š” ์–‘์˜ ์ •๋ถ€ํ˜ธ ํ–‰๋ ฌ $P$, $X$, ํ–‰๋ ฌ $Y$, $Z$์™€ ์–‘์˜ ์ƒ์ˆ˜ $\theta$, $\beta_{1}$, $\beta_{2}$๊ฐ€ ์กด์žฌํ•˜๋ฉด, ์ƒํƒœ๊ถคํ™˜ ์ œ์–ด๊ธฐ $u(k)=Y X^{-1}x(k)$๋Š” ๋ณ€์ˆ˜ ๋ถˆํ™•์‹ค์„ฑ์„ ๊ฐ€์ง€๋Š” ์ด์‚ฐ์‹œ๊ฐ„ ํŠน์ด์‹œ์Šคํ…œ (1)์— ๋Œ€ํ•˜์—ฌ ์ •์˜ 1๊ณผ ์ •์˜ 3์„ ๋งŒ์กฑํ•˜๋Š” ๊ฐ•์ธ ์œ ํ•œ์‹œ๊ฐ„ ํŠน์ด์•ˆ์ •ํ•˜๋‹ค. ์—ฌ๊ธฐ์„œ, $\Phi$๋Š” $E\Phi =0$์„ ๋งŒ์กฑํ•˜๋Š” ํ–‰๋ ฌ์ด๊ณ ,

$\Omega_{1}= <(A-E)X >+ < BY >-\alpha EPE^{T}+\beta_{1}M_{a}M_{a}^{T}+\beta_{2}M_{b}M_{b}^{T}$, $\Omega_{2}=(A-E)X+BY+Z\Phi^{T}-X$,$\widetilde P =R^{-1/2}P R^{-1/2}$, $\beta_{i}=\epsilon_{i}^{-1}(i=1,\:2)$ ์ด๋‹ค.

์ฆ๋ช…: ์‹(1)์—์„œ $A_{k}=A+\Delta A(k)$, $B_{k}=B+\Delta B(k)$๋ผ ๋‘๋ฉด, ์ƒํƒœ๊ถคํ™˜ ์ œ์–ด๊ธฐ ์‹(3)์œผ๋กœ๋ถ€ํ„ฐ ํ๋ฃจํ”„์‹œ์Šคํ…œ์€

(20)
$Ex(k+1)= A_{c}x(k)$ (20)

๊ณผ ๊ฐ™๊ณ , $A_{c}=A_{k}+ B_{k}K$์ด๋‹ค. ์‹(20)์„ ์‹(6)์— ๋Œ€์ž…ํ•˜๋ฉด

(21)
$\begin{bmatrix}\Theta_{1}&(A_{c}-E)^{T}X+Z\Phi^{T}-X&E^{T}P&(A_{c}-E)^{T}X\\ \ast &-2X&P&-X\\ \ast &\ast &-P&0\\ \ast &\ast &\ast &-X\end{bmatrix}<0$

์ด ๋œ๋‹ค. ์—ฌ๊ธฐ์„œ, $\Theta_{1}=\left <(A_{c}-E)^{T}X\right > -\alpha E^{T}PE$์ด๋‹ค. ๋˜ํ•œ, $\det(z E-A_{c})=\det(z E^{T}-A_{c}^{T})$์ด๋ฏ€๋กœ $(E,\:A_{c})$๊ฐ€ ์ •๊ทœ์ ์ด๊ณ  ์ธ๊ณผ์ ์ด๊ธฐ ์œ„ํ•œ ํ•„์š”์ถฉ๋ถ„์กฐ๊ฑด์€ $(E^{T},\: A_{c}^{T})$์ด ์ •๊ทœ์ ์ด๊ณ  ์ธ๊ณผ์ ์ด๋‹ค. ๋˜ํ•œ, $\det(z E-A_{c})=0$์˜ ํ•ด๋Š” $\det(z E^{T}-A_{c}^{T})=0$์˜ ํ•ด์™€ ๋™์ผํ•˜๊ธฐ ๋•Œ๋ฌธ์— ์‹(20)์˜ ์œ ํ•œ์‹œ๊ฐ„ ์•ˆ์ •์„ฑ์€ $(E^{T},\: A_{c}^{T})$์˜ ์œ ํ•œ์‹œ๊ฐ„ ์•ˆ์ •์„ฑ๊ณผ ๋™์ผํ•œ ์กฐ๊ฑด์ด๋‹ค. ๋”ฐ๋ผ์„œ, $K=YX^{-1}$๋กœ ๋‘๊ณ , ์‹(21)์—์„œ $E$์™€ $A_{c}$๋ฅผ $E^{T}$์™€ $A_{c}^{T}$๋กœ ๋Œ€์ž…ํ•˜์—ฌ ์ •๋ฆฌํ•˜๋ฉด

(22)
$\begin{bmatrix}\Sigma_{1}&(A_{k}-E)X+B_{k}Y+Z\Phi^{T}-X&EP&(A_{k}-E)X+B_{k}Y\\\ast &-2X&P&-X\\\ast &\ast &-P&0\\\ast &\ast &\ast &-X\end{bmatrix}<0$

์™€ ๊ฐ™๊ณ , $\Sigma_{1}=\left <(A_{k}-E)X\right > +\left < B_{k}Y\right > -\alpha EPE^{T}$์ด๋‹ค. ์‹(22)์—์„œ ์‹(2)๋ฅผ ๋Œ€์ž…ํ•ด์„œ ์ •๋ฆฌํ•˜๋ฉด

(23)
\begin{align*} \begin{bmatrix}\Sigma_{2}&(A-E)X+BY+Z\Phi^{T}-X&EP&(A-E)X+BY\\\ast &-2X&P&-X\\\ast &\ast &-P&0\\\ast &\ast &\ast &-X\end{bmatrix}\\ +\left <\begin{bmatrix}M_{a}\\0\\0\\0\end{bmatrix}F(k)\begin{bmatrix}N_{a}X&N_{a}X&0&N_{a}X\end{bmatrix}\right > \\+\left <\begin{bmatrix}M_{b}\\0\\0\\0\end{bmatrix}F(k)\begin{bmatrix}N_{b}Y&N_{b}Y&0&N_{b}Y\end{bmatrix}\right > <0 \end{align*}

์ด ๋˜๊ณ , $\Sigma_{2}= <(A-E)X > + < BY > -\alpha EPE^{T}$์ด๋‹ค. ์—ฌ๊ธฐ์„œ, $F(k)^{T}F(k)\le I$์ด๋ฏ€๋กœ ์‹(23)์˜ 2๋ฒˆ์งธ์™€ 3๋ฒˆ์งธ๋Š” ์ˆ˜์‹์€

(24)
\begin{align*} \left <\begin{bmatrix}M_{a}\\0\\0\\0\end{bmatrix}F(k)\begin{bmatrix}N_{a}X&N_{a}X&0&N_{a}X\end{bmatrix}\right >\\\le\epsilon_{1}^{-1}\begin{bmatrix}M_{a}\\0\\0\\0\end{bmatrix}\begin{bmatrix}M_{a}^{T}&0&0&0\end{bmatrix}+\epsilon_{1}\begin{bmatrix}(N_{a}X)^{T}\\(N_{a}X)^{T}\\0\\(N_{a}X)^{T}\end{bmatrix}\begin{bmatrix}N_{a}X&N_{a}X&0&N_{a}X\end{bmatrix} \end{align*}

(25)
\begin{align*} \left <\begin{bmatrix}M_{b}\\0\\0\\0\end{bmatrix}F(k)\begin{bmatrix}N_{b}Y&N_{b}Y&0&N_{b}Y\end{bmatrix}\right > \\\le\epsilon_{2}^{-1}\begin{bmatrix}M_{b}\\0\\0\\0\end{bmatrix}\begin{bmatrix}M_{b}^{T}&0&0&0\end{bmatrix}+\epsilon_{2}\begin{bmatrix}(N_{b}Y)^{T}\\(N_{b}Y)^{T}\\0\\(N_{b}Y)^{T}\end{bmatrix}\begin{bmatrix}N_{b}Y&N_{b}Y&0&N_{b}Y\end{bmatrix} \end{align*}

๋ฅผ ๋งŒ์กฑํ•˜๋Š” ์–‘์˜ ์‹ค์ˆ˜ $\epsilon_{1}$๊ณผ $\epsilon_{2}$๊ฐ€ ์กด์žฌํ•˜๋ฏ€๋กœ, ์‹(23)์— ์‹(24)์™€ 25์˜ ๊ด€๊ณ„๋ฅผ ๋Œ€์ž…ํ•˜์—ฌ ์ •๋ฆฌํ•˜๋ฉด ์‹(17)์„ ์–ป์„ ์ˆ˜ ์žˆ๋‹ค. ์‹(18)๊ณผ (19)๋Š” ์ •๋ฆฌ 1์—์„œ ์ง์ ‘์ ์œผ๋กœ ๊ตฌํ•˜์—ฌ์ง„๋‹ค. ๋”ฐ๋ผ์„œ, $u(k)=Kx(k)=YX^{-1}x(k)$์˜ ์ƒํƒœ๊ถคํ™˜ ์ œ์–ด๊ธฐ๋Š” ์ด์‚ฐ์‹œ๊ฐ„ ๋ถˆํ™•์‹ค ํŠน์ด์‹œ์Šคํ…œ (1)์ด ์ •๊ทœ์ ์ด๊ณ  ์ธ๊ณผ์ ์ด๋ฉฐ ๊ฐ•์ธ ์œ ํ•œ์‹œ๊ฐ„ ํŠน์ด์•ˆ์ •์„ฑ์„ ๋งŒ์กฑํ•˜๋„๋ก ํ•œ๋‹ค.

์ •๋ฆฌ 2์˜ ์‹(19)์—์„œ $\alpha^{N}$์€ $N$์ด ๋ฌดํ•œ๋Œ€๋กœ ๊ฐˆ์ˆ˜๋ก ์œ ํ•œ์‹œ๊ฐ„์—์„œ ๋ฌดํ•œ์‹œ๊ฐ„ ๋ฌธ์ œ๋กœ ๋ณ€๊ฒฝ๋˜์–ด์ง€๋ฉฐ, $\alpha >1$์ด๋ฏ€๋กœ $\alpha$๋Š” 1์— ๊ฐ€๊นŒ์›Œ์ ธ์•ผ ์ˆ˜๋ ดํ•˜๊ฒŒ ๋œ๋‹ค. ์ตœ๊ทผ Wang ๋“ฑ(10)์˜ ์œ ํ•œ์‹œ๊ฐ„ ํŠน์ด์•ˆ์ •์„ฑ ๋ฌธ์ œ์—์„œ ์œ ํ•œ์‹œ๊ฐ„ ์•ˆ์ •์„ฑ ์กฐ๊ฑด์€ ์„ ํ˜•ํ–‰๋ ฌ๋ถ€๋“ฑ์‹์œผ๋กœ ์ œ์‹œํ•˜์˜€์ง€๋งŒ, ๊ฐ•์ธ ์œ ํ•œ์‹œ๊ฐ„ ํŠน์ด์•ˆ์ •ํ™”ํ•˜๊ฒŒ ํ•˜๋Š” ์ƒํƒœ๊ถคํ™˜ ์ œ์–ด๊ธฐ๋ฅผ ๊ตฌํ•˜๋Š” ์ถฉ๋ถ„์กฐ๊ฑด์€ ๊ตฌํ•˜๋ ค๋Š” ๋ณ€์ˆ˜์˜ ๊ฒฌ์ง€์—์„œ ๋ณผ๋ก์ตœ์ ํ™”๋กœ ํ‘œํ˜„๋˜์ง€ ์•Š์•„์„œ ํ•ด๋ฅผ ๊ตฌํ•˜๊ธฐ ์‰ฝ์ง€ ์•Š์•˜๋‹ค. ๋˜ํ•œ, ์ œ์•ˆํ•œ ์ œ์–ด๊ธฐ์˜ ํ˜•ํƒœ๋ฅผ ์ œ์‹œํ•˜์˜€๋‹ค๊ณ  ํ•˜๋‚˜ ์ง€๋ฃจํ•œ ๊ณผ์ •์ด ํ•„์š”ํ–ˆ๋‹ค. ํ•˜์ง€๋งŒ, ๋ณธ ๋…ผ๋ฌธ์—์„œ ์ œ์‹œํ•˜๋Š” ์œ ํ•œ์‹œ๊ฐ„ ํŠน์ด์•ˆ์ •์„ฑ ์กฐ๊ฑด์˜ ์ •๋ฆฌ 1๊ณผ ๊ฐ•์ธ ์œ ํ•œ์‹œ๊ฐ„ ํŠน์ด์•ˆ์ •ํ™”ํ•˜๊ฒŒ ํ•˜๋Š” ์ƒํƒœ๊ถคํ™˜ ์ œ์–ด๊ธฐ ์„ค๊ณ„๋ฐฉ๋ฒ•์ธ ์ •๋ฆฌ 2๋Š” ๊ตฌํ•˜๋ ค๋Š” ๋ชจ๋“  ๋ณ€์ˆ˜์˜ ๊ฒฌ์ง€์—์„œ ์„ ํ˜•ํ–‰๋ ฌ๋ถ€๋“ฑ์‹ ์กฐ๊ฑด์œผ๋กœ ํ‘œํ˜„ํ•˜๋ฏ€๋กœ ํ•ด๋ฅผ ํ•œ๋ฒˆ์— ๊ตฌํ•  ์ˆ˜ ์žˆ๋‹ค. ๋˜ํ•œ, ์ •๋ฆฌ 2์—์„œ $E=I$๊ฐ€ ๋˜๋ฉด ๋น„ํŠน์ด์‹œ์Šคํ…œ์— ๋Œ€ํ•œ ๊ฐ•์ธ ์œ ํ•œ์‹œ๊ฐ„ ์•ˆ์ •์„ฑ์„ ๋งŒ์กฑํ•˜๋Š” ์ƒํƒœ๊ถคํ™˜ ์ œ์–ด๊ธฐ๋ฅผ ์„ค๊ณ„ํ•  ์ˆ˜ ์žˆ์œผ๋ฏ€๋กœ ์ผ๋ฐ˜์ ์ธ ์ œ์–ด๊ธฐ ์„ค๊ณ„ ์•Œ๊ณ ๋ฆฌ๋“ฌ์ด๋‹ค.

3. ์ˆ˜์น˜ ์˜ˆ์ œ

์ œ์•ˆํ•œ ์•Œ๊ณ ๋ฆฌ๋“ฌ์˜ ํƒ€๋‹น์„ฑ์„ ๋ณด์—ฌ์ฃผ๊ธฐ ์œ„ํ•˜์—ฌ ๊ฐœ๋ฃจํ”„ ์‹œ์Šคํ…œ์ด ๋ถˆ์•ˆ์ •ํ•œ ๋ณ€์ˆ˜ ๋ถˆํ™•์‹ค์„ฑ์„ ๊ฐ€์ง€๋Š” ์ด์‚ฐ์‹œ๊ฐ„ ํŠน์ด์‹œ์Šคํ…œ

(26)
\begin{align*} \begin{bmatrix}1&0&0\\0&1&0\\0&0&0\end{bmatrix}x(k+1) &=&\left\{\begin{bmatrix}1.2&0&1\\1&0.1&0\\1&-0.3&-0.6\end{bmatrix}+\begin{bmatrix}0.1\\0.1\\0.1\end{bmatrix}F(k)\begin{bmatrix}0.2&0.2&0.1\end{bmatrix}\right\}x(k) \\&&+\left\{\begin{bmatrix}0&1\\1&1\\1&0\end{bmatrix}+\begin{bmatrix}0.2\\0.2\\0.2\end{bmatrix}F(k)\begin{bmatrix}0.2&0.1\end{bmatrix}\right\}u(k) \end{align*}

์„ ๊ณ ๋ คํ•œ๋‹ค. $F(k)=\sin(k)$์™€ $u(k)=0$์ธ ์‹(26)์˜ ๊ฐœ๋ฃจํ”„ ์‹œ์Šคํ…œ์— ๋Œ€ํ•œ ์ƒํƒœ์˜ ๊ถค์ ์€ ์‹œ๊ฐ„์ด ์ฆ๊ฐ€ํ• ์ˆ˜๋ก ๋ฐœ์‚ฐํ•˜๊ณ  ์žˆ์Œ์„ ๊ทธ๋ฆผ 1์—์„œ ๋ณด์—ฌ์ค€๋‹ค. ์—ฌ๊ธฐ์„œ, $c_{1}=2$, $c_{2}=5$, $N=40$, $\alpha =1.0001$, $R=diag\{1,\:1,\:1\}$๋กœ ์„ค์ •ํ•˜๊ณ , $E\Phi =0$์„ ๋งŒ์กฑํ•˜๋Š” $\Phi =\begin{bmatrix}0& 0& 1\end{bmatrix}^{T}$๋กœ ๋‘๋ฉด, ์ •๋ฆฌ 2๋ฅผ ๋งŒ์กฑํ•˜๋Š” ํ•ด๋Š”

(27)
$P=\begin{bmatrix}0.6591 & -0.0809& -0.1485 \\\ast & 0.9124& 0.0171\\\ast &\ast & 0.5900\end{bmatrix}$, $X=\begin{bmatrix}1.9944 & 0.1007& -1.1707\\\ast & 2.7084& 0.3883\\\ast &\ast & 1.5618\end{bmatrix}$, $Y=\begin{bmatrix}-1.4960& 0.5230& 0.8354\\-0.8436& -0.5711& -0.0695\end{bmatrix}$, $Z=\begin{bmatrix}-0.0048\\0.0121\\0.5235\end{bmatrix}$, $\epsilon_{1}= 0.2885$, $\epsilon_{2}=0.5735$, $\theta =0.4511$

๊ณผ ๊ฐ™์ด ํ•œ๋ฒˆ์— ๊ตฌํ•ด์ง„๋‹ค. ๋”ฐ๋ผ์„œ, ๋ณธ ๋…ผ๋ฌธ์˜ ๋ชฉ์ ์ด ๊ตฌํ•˜๋ ค๋Š” ๋ณ€์ˆ˜ ์ธก๋ฉด์—์„œ ์„ ํ˜•ํ–‰๋ ฌ๋ถ€๋“ฑ์‹์œผ๋กœ ํ‘œํ˜„ํ•œ ์ •๋ฆฌ 2์—์„œ ์‹(27)์˜ ํ•ด๋ฅผ ํ•œ๋ฒˆ์— ๊ตฌํ•˜๋Š” ๊ฒƒ์ด๋‹ค. ๋˜ํ•œ, ์‹(3)์˜ ์ƒํƒœ๊ถคํ™˜ ์ œ์–ด๊ธฐ๋„ ์‹(27)๋กœ๋ถ€ํ„ฐ

(28)
$u(k)=Kx(k)=YX^{-1}x(k)=\begin{bmatrix}-0.8667&0.2507&-0.1771\\-0.7681&-0.0968&-0.5962\end{bmatrix}x(k)$

๊ณผ ๊ฐ™์ด ์ง์ ‘ ๊ตฌํ•ด์ง„๋‹ค. ์‹(26)๊ณผ ์‹(28)๋กœ๋ถ€ํ„ฐ ๊ตฌํ•œ ํ๋ฃจํ”„ ์‹œ์Šคํ…œ์˜ ๊ฐ•์ธ ์œ ํ•œ์‹œ๊ฐ„ ํŠน์ด์•ˆ์ •์„ฑ์˜ ์‹œ๋ฎฌ๋ ˆ์ด์…˜ ๊ฒฐ๊ณผ๋ฅผ ๋ณด์—ฌ์ฃผ๊ธฐ ์œ„ํ•˜์—ฌ $F(k)=\sin(k)$, ์ดˆ๊ธฐ์กฐ๊ฑด์„ $x(0)=\begin{bmatrix}1& -0.5& 0.7\end{bmatrix}^{T}$์™€ ๊ฐ™์ด ๋‘๋ฉด, ์ดˆ๊ธฐ์กฐ๊ฑด์— ๋Œ€ํ•˜์—ฌ $x^{T}(0)E^{T}R E x(0)\le c_{1}=2$๋ฅผ ๋งŒ์กฑํ•œ๋‹ค. ๊ทธ๋ฆผ 2์™€ 3์—์„œ๋Š” ํ๋ฃจํ”„ ์‹œ์Šคํ…œ์— ๋Œ€ํ•œ ์ƒํƒœ์˜ ๊ถค์ ๊ณผ $x^{T}(k)E^{T}REx(k)$์— ๋Œ€ํ•œ ๊ถค์ ์„ ๊ฐ๊ฐ ๋ณด์—ฌ์ค€๋‹ค. ๋”ฐ๋ผ์„œ, $k\in\{1,\:2,\:\cdots ,\:N\}$์— ๋Œ€ํ•ด $x^{T}(k)E^{T}REx(k)<c_{2}=5$๋ฅผ ๋งŒ์กฑํ•˜๋ฏ€๋กœ ์ œ์•ˆํ•œ ์ƒํƒœ๊ถคํ™˜ ์ œ์–ด๊ธฐ ์‹(28)์€ ๋ณ€์ˆ˜ ๋ถˆํ™•์‹ค์„ฑ์„ ๊ฐ€์ง€๋Š” ์ด์‚ฐ์‹œ๊ฐ„ ํŠน์ด์‹œ์Šคํ…œ (26)์— ๋Œ€ํ•ด ๊ฐ•์ธ ์œ ํ•œ์‹œ๊ฐ„ ํŠน์ด์•ˆ์ •ํ•˜๊ฒŒ ํ•œ๋‹ค.

๊ทธ๋ฆผ. 1. ๊ฐœ๋ฃจํ”„ ์‹œ์Šคํ…œ์˜ ์ƒํƒœ ๊ถค์ 

Fig. 1. The state trajectories of open-loop system.

../../Resources/kiee/KIEE.2020.69.12.1929/fig1.png

๊ทธ๋ฆผ. 2. ํ๋ฃจํ”„ ์‹œ์Šคํ…œ์˜ ์ƒํƒœ ๊ถค์ 

Fig. 2. The state trajectories of closed-loop system.

../../Resources/kiee/KIEE.2020.69.12.1929/fig2.png

๊ทธ๋ฆผ. 3. $x^{T}(k)E^{T}RE x(k)$์˜ ๊ถค์ 

Fig. 3. The trajectory of $x^{T}(k)E^{T}RE x(k)$.

../../Resources/kiee/KIEE.2020.69.12.1929/fig3.png

4. ๊ฒฐ ๋ก 

๋ณธ ๋…ผ๋ฌธ์—์„œ๋Š” ๋ณ€์ˆ˜ ๋ถˆํ™•์‹ค์„ฑ์„ ๊ฐ€์ง€๋Š” ์ด์‚ฐ์‹œ๊ฐ„ ํŠน์ด์‹œ์Šคํ…œ ๋Œ€ํ•œ ์œ ํ•œ์‹œ๊ฐ„ ํŠน์ด์•ˆ์ •์„ฑ ์กฐ๊ฑด๊ณผ ๊ฐ•์ธ ์œ ํ•œ์‹œ๊ฐ„ ํŠน์ด์•ˆ์ •์„ฑ์„ ๋งŒ์กฑํ•˜๋Š” ์ƒํƒœ๊ถคํ™˜ ์ œ์–ด๊ธฐ ์„ค๊ณ„๋ฐฉ๋ฒ•์„ ๊ตฌํ•˜๋ ค๋Š” ๋ณ€์ˆ˜์˜ ๊ฒฌ์ง€์—์„œ ๋ณผ๋ก์ตœ์ ํ™”๊ฐ€ ๊ฐ€๋Šฅํ•œ ์„ ํ˜•ํ–‰๋ ฌ๋ถ€๋“ฑ์‹์œผ๋กœ ์ œ์•ˆํ•˜์˜€๋‹ค. ๊ธฐ์กด์˜ ๊ฒฐ๊ณผ๋“ค์ด ๋“ฑํ˜ธ์กฐ๊ฑด์„ ํฌํ•จํ•˜๋Š” ์ค€์ •๋ถ€ํ˜ธ ๋ฌธ์ œ๋ฅผ ๋‹ค๋ฃจ๊ณ  ์žˆ์–ด์„œ ํ•ด๋ฅผ ๊ตฌํ•˜๊ธฐ ์–ด๋ ต๋‹ค๋Š” ๋ฌธ์ œ์ ์„ ์‹œ์Šคํ…œ์˜ ๋“ฑ๊ฐ€์„ฑ์งˆ์„ ์ด์šฉํ•œ ์ƒˆ๋กœ์šด ์ ‘๊ทผ๋ฐฉ๋ฒ•์œผ๋กœ ํ•ด๊ฒฐํ•˜์˜€๋‹ค. ์ œ์•ˆํ•œ ์ •๋ฆฌ 1๊ณผ ์ •๋ฆฌ 2๋Š” ํŠน์ด์‹œ์Šคํ…œ ๋ฟ๋งŒ ์•„๋‹ˆ๋ผ ๋น„ํŠน์ด์‹œ์Šคํ…œ์— ๋Œ€ํ•ด์„œ๋„ ์ ์šฉ๊ฐ€๋Šฅํ•˜๋ฏ€๋กœ ์ผ๋ฐ˜์ ์ธ ์กฐ๊ฑด์ด๋‹ค. ๊ฐœ๋ฃจํ”„ ์‹œ์Šคํ…œ์ด ๋ถˆ์•ˆ์ •ํ•œ ์ˆ˜์น˜์˜ˆ์ œ์™€ ์‹œ๋ฎฌ๋ ˆ์ด์…˜์„ ํ†ตํ•˜์—ฌ ์ œ์•ˆํ•œ ์กฐ๊ฑด์˜ ํƒ€๋‹น์„ฑ์„ ํ™•์ธํ•˜์˜€๋‹ค. ์ œ์•ˆํ•œ ์ƒˆ๋กœ์šด ์ƒํƒœ๊ถคํ™˜ ์ œ์–ด๊ธฐ ์„ค๊ณ„๋ฐฉ๋ฒ•์€ ํŠน์ด์‹œ์Šคํ…œ์„ ๋‹ค๋ฃจ๋Š” ๋‹ค์–‘ํ•œ ๋ถ„์•ผ์— ํ™•์žฅ๊ฐ€๋Šฅํ•˜๋‹ค.

References

1 
F. Amato, R. Ambrosino, M. Ariola, C. Cosentino, G. De Tommasi, 2014, Finite-time Stability and Control, Lecture Notes in Control and Information Sciences 453, SpringerGoogle Search
2 
G. Kamenkov, 1953, On stability of motion over a finite interval of time, Journal of Applied Mathematics and Mechanics, Vol. 17, No. 1, pp. 529-540Google Search
3 
F. Amato, M. Ariola, P. Dorato, 1998, Robust finite-time stabilization of linear systems depending on parameter uncertainties, Proc. 37th IEEE Conference on Decision & Control, pp. 1207-1208DOI
4 
F. Amato, M. Ariola, 2005, Finite-time control of discrete-time linear systems, IEEE Transactions on Automatic Control, Vol. 50, No. 5, pp. 724-729DOI
5 
S. Xu, J. Lam, 2006, Robust Control and Filtering of Singular Systems, Lecture Notes in Control and Information Sciences 332, SpringerGoogle Search
6 
D. S. Antic, S. B. Stojanovic, D. Lj. Debeljkovic, 2012, Finite-time stability and stabilization of singular discrete time-delay systems, XI International SAUM Conference on Systems, automatic control and measurements, pp. 160-163Google Search
7 
S. Wo, X. Han, 2014, Finite-time stability analysis of discrete-time linear singular systems, Mathematical Pro- blems in Engineering, ID 579863DOI
8 
Y. Ma, X. Jia, D. Liu, 2016, Robust finite-time control for discrete-time singular Markovian jump systems with time-varying delay and actuator saturation, Applied Mathematics and Computation, Vol. 286, pp. 213-227DOI
9 
Y. Ma, X. Jia, D. Liu, 2018, Finite-time dissipative control for singular discrete-time Markovian jump systems with actuator saturation and partly unknown transition rates, Applied Mathematics and Modelling, Vol. 53, pp. 49-70DOI
10 
J. Wang, H. Wu, X. Ji, X. Liu, 2020, Robust finite-time stabilization for uncertain discrete-time linear singular systems, IEEE Access, Vol. 8, pp. 100645-100650DOI
11 
L. Dai, 1989, Singular Control Systems, Berlin, Springer-VerlagGoogle Search
12 
S. Xu, J. Lam, 2005, Robust stability and stabilization of discrete singular systems: an equivalent characterization, IEEE Transactions on Automatic Control, Vol. 49, No. 4, pp. 568-574DOI
13 
S. Boyd, L. E. Ghaoui, E. Feron, V. Balakrishnan, 1994, Linear Matrix Inequalities in System and Control Theory, SIAM Studies in Applied MathematicsGoogle Search

์ €์ž์†Œ๊ฐœ

Jonghae Kim
../../Resources/kiee/KIEE.2020.69.12.1929/au1.png

He received the B.S., M.S., and Ph.D. degrees in Electronics from the Kyungpook National University, Korea, in 1993, 1995, and 1998, respectively. He was with STRC at Kyungpook National University from 1998 to 2002. Dr. Kim was a Research Fellow at Osaka University, Japan, from March 2000 to March 2001. Also, he was a Visiting Research Scholar at the Georgia Institute of Technology, USA, during the period, Jan. 2010~Feb. 2011. In 2002, he has joined the Department of Electronic Engi- neering, Sun Moon University, Korea, and currently he is a Professor at the Department. His research interests include robust control, robust filtering, signal processing, and industrial application systems.