Signal Processing and Linear Systems-B.P.Lathi copy

B ut even in a n a symptotically stable system we see

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: matched. Therefore, Ao = bn a nd h (t) = bn8(t) + characteristic modes (2.79) To determine t he c haracteristic mode terms in t he above equation, let us consider a system So whose i nput f (t) a nd t he c orresponding o utput x (t) a re related by Q (D)x(t) = f (t) (2.80) 162 2 T ime-pomain Analysis of Continuous-Time Systems Observe t hat b oth t he systems S a nd So have t he same characteristic polynomial; namely, Q(>.), a nd, consequently, t he s ame characteristic modes. Moreover, So is t he s ame as S w ith P(D) = 1, t hat is, bn = O. Therefore, according t o Eq. (2.79), t he impulse response of So consists of characteristic mode terms only without a n impulse a t t = O. L et us denote this impulse response of So b y Yn(t). Observe t hat Yn (t) c onsists of characteristic modes of S, a nd therefore may be viewed as a zero-input response of S. Now Yn(t) is t he response of So t o i nput 8(t). Therefore, according t o Eq. (2.80) 2.9 S ummary 163 T he right-hand side of Eq. (2.84) is a linear combination of t he derivatives of Yn(t)u(t). Evaluating these derivatives is clumsy and inconvenient because of the presence of u(t). T he derivatives will generate an impulse and its derivatives at the origin. Fortunately when m :s; n [Eq. (2.77)], we c an avoid this difficulty by using t he observation in Eq. (2.79), which asserts t hat a t t = 0 ( the origin), h(t) = bno(t). Therefore, we need not bother t o find h (t) a t t he origin. This simplification means t hat i nstead of deriving P(D)[Yn(t)u(t)], we c an derive P(D)Yn(t) and add t o it the term bno(t), so t hat (2.81a) Q(D)Yn(t) = 8(t) or = bno(t) + [P(D)Yn(t)]u(t) (2.85) (2.81b) or (2.81c) where y~k) (t) r epresents t he k th derivative of Yn(t). T he r ight-hand side contains a single impulse t erm 8(t). T his is possible only if y~n-l) (t) has a u nit j ump discontinuity a t t = 0, so t hat y~n) (t) = 8(t). Moreover, t he lower-order t erms c an not have any j ump d iscontinuity because t his would mean t he presence of t he derivatives of 8(t). Suppose, f or instance, Yn(t) h as a jump discontinuity, t hen i ts derivative Yn(t) c ontains a n i mpulse 8(t), a nd i ts second derivative Yn(t) contains t he first derivative of t he impulse 8 (t), a nd so on. B ut t his is impossible because t he r ight-hand side of Eq. (2.81c) consists of only 8(t). For this reason only y~n-l) (t) c an have a unit j ump d iscontinuity so t hat y~n)(t) is 8(t). T here c an be no j ump discontinuities in any of t he r emaining variables because this would give rise t o higher-order derivatives of 8(t) o n t he l eft-hand side. Therefore Yn(O) = y~l) (0) = '" = y~n-2) (0) = 0 (no discontinuity a t t = 0). Therefore, t he n initial conditions on Yn(t) a re y~n-l)(o) = 1 Yn(O) = y~l)(O) = . .. = y~n-2)(0) =0 (2.82) T his discussion m eans t hat Yn(t) is t he z ero-input response of t he s ystem S s ubject t o i nitial conditions (2.82). We now show t hat for t he same i nput f (t) t o b oth systems, S a nd So, t...
View Full Document

This note was uploaded on 04/14/2013 for the course ENG 350 taught by Professor Bayliss during the Spring '13 term at Northwestern.

Ask a homework question - tutors are online