7 Minimal realization and coprime fraction • 7.1 Introduction • If a transfer function is realizable, what is the smallest possible dimension? • Realizations with the smallest possible dimension are called minimal-dimensional or minimal realizations. 7.2 implications and Coprimeness • Consider β1s3 + β2s2 + β3s + β4 N (s ) = ĝ(s) = D(s) s4 + α1s3 + α2s2 + α3s + α4 • Consider ŷ(s) = N (s) D −1(s) û (s) • Let a pseudo state v̂ (s) = D −1(s) û (s)( or D(s) v̂ (s) = û(s)) • The realization is ⎡ − α1 − α2 ⎢ 1 0 x& = Ax + bu = ⎢ 1 ⎢ 0 ⎢ 0 ⎣ 0 − α3 − α 4 ⎤ ⎡1 ⎤ ⎢ 0⎥ 0 0 ⎥ ⎥x + ⎢ ⎥u 0 0 ⎥ ⎢ 0⎥ ⎢ ⎥ ⎥ 1 0 ⎦ ⎣ 0⎦ y = cx = [β1 β2 β3 β4 ]x • Its controllability matrix can be computed as ⎡1 − α1 α12 − α2 ⎢ 0 1 − α1 C=⎢ ⎢0 0 1 ⎢ 0 0 ⎢⎣0 − α13 + 2α1α2 − α3 ⎤ ⎥ 2 α1 − α2 ⎥ ⎥ − α1 ⎥ 1 ⎥⎦ • Its determinant is 1 for any αi. Hence the realization is called a controllable canonical form. • Theorem 7.1 The controllable canonical form is observable if and only if D(s) and N(s) are coprime. • If the controllable canonical form is a realization of ĝ(s), then we have, by definition, ĝ(s) = c(sI − A ) −1 b • Taking its transpose yields the state equation (a different realization) ⎡ − α1 ⎢− α x& = A′x + c′u = ⎢ 2 ⎢ − α3 ⎢ ⎣ − α4 1 0 0⎤ ⎡ β1 ⎤ ⎢β ⎥ 0 1 0⎥ ⎥x + ⎢ 2 ⎥u 0 0 1⎥ ⎢β3 ⎥ ⎥ ⎢ ⎥ 0 0 0⎦ ⎣β4 ⎦ y = b′x = [1 0 0 0]x • It is called an observable canonical form. • The equivalent transformation x = Px with ⎡0 ⎢0 P=⎢ ⎢0 ⎢ ⎣1 0 0 1⎤ 0 1 0⎥ ⎥ 1 0 0⎥ ⎥ 0 0 0⎦ • will get the different controllable and observable canonical form. • 7.2.1 Minimal realizations • Let R(s) be a greatest common divisor (gcd) of N(s) and D(s). Then, the transfer function can be reduce to (coprime fraction) ĝ(s) = N (s) / D(s). where N (S) = N (s) R (s) and D(s) = D(s)R(s) • We call D(s) a characteristic polynominal of ĝ(s). Its degree is defined as the degree of ĝ(s). • Theorem 7.2 A state equation (A, b, c, d) is a minimal realization of a proper reational function ĝ(s) if and only if (A, b) is controllable and (A, c) is observable or if and only if dim(A) = deg(ĝ(s)) • The Theorem provides a alternative way of checking controllability and observability. • Theorem 7.3 All minimal realizations of ĝ(s) are equivalent. • If a state equation is controllable and observable, then every eigenvalue of A is a pole of ĝ(s) and every pole of ĝ(s) is an eigenvalue of A. • Thus we conclude that if (A, b, c, d) is controllable and observable, then we have Asymptotic stability ⇔ BIBO stability. 7.3 Computing coprime fractions • Let write N (s ) N (s ) = D(s ) D (s ) which implies D(s)( − N (s)) + N (s) D(s) = 0 • Let D(s)=D0+D1s+D2S2+D3s3+D4s4 N(s)=N0+N1s+N2s2+N3s3+N4s4 D (s) = D0 + D1s + D2s2 + D3s3 N (s) = N 0 + N1s + N 2s2 + N 3s3 • Sylverster resultant (Homogeneous linear algebraic equation) ⎡ D0 ⎢D ⎢ 1 ⎢ D2 ⎢ D S := ⎢ 3 ⎢ D4 ⎢ ⎢ 0 ⎢ 0 ⎢ ⎣ 0 N0 0 0 0 0 0 N1 D0 N0 0 0 0 N2 D1 N1 D0 N0 0 N3 D2 N2 D1 N1 D0 N4 D3 N3 D2 N2 D1 0 D4 N4 D3 N3 D2 0 0 0 D4 N4 D3 0 0 0 0 0 D4 0 ⎤ ⎡− N0 ⎤ 0 ⎥ ⎢⎢ D0 ⎥⎥ ⎥ 0 ⎥ ⎢ − N1 ⎥ ⎥ ⎥⎢ N 0 ⎥ ⎢ D1 ⎥ =0 ⎥ ⎢ ⎥ N1 − N 2 ⎥ ⎥⎢ N 2 ⎥ ⎢ D2 ⎥ N3 ⎥⎢− N3 ⎥ ⎥ ⎥⎢ N 4 ⎦ ⎢⎣ D3 ⎥⎦ • D(s) and N(s) are coprime if and only if the Sylverster resultant is nonsingular. • Theorem 7.4 Deg ĝ(s) = number of linearly independent N-columns =: µ and the coefficients of a coprime fraction [− N 0 D0 − N1 D1 . . − N µ Dµ ]′ equals the monic null vector of the matrix that consists of the primary dependent Ncolumn and all its LHS linearly independent columns of S. • 7.3.1 QR Decomposition • Consider an n×m matrix M. Then there exists an n×n orthogonal matrix Q such that QM = R where R is an upper triangular matrix. • Because Q is orthogonal, we have Q −1 = Q ′ =: Q and M = QR 7.4 Balanced realization • The diagonal and modal forms, which are least sensitive to parameter variations, are good candidates for practical implementation. • A different minimal realizations, called a balanced realization. • Consider a stable system x& = Ax + bu y = cx • Then the controllability Gramian Wc and the observability Wo are positive definite if the system is controllable and observable AWc + WcA’ = -bb’ A’Wo + WoA = -c’c • Different minimal realizations of the same transfer function have different controllability and observability. • Theorem 7.5 Let (A, b, c) and ( A, b, c ) be minimal and equivalent. Then WcWo and Wc Wo are similar and their eigenvalues are all real and positive. • Theorem 7.6 A balanced realization For any minimal state equation (A, b, c) an equivalent transformation x = Px such that the equivalent controllability and observability have the property Wc = Wo = Σ 7.5 Realizations from Markov parameters • Consider the strictly proper rational function β1sn −1 + β2sn − 2 + ... + β n −1s + β n ĝ(s) = n s + α1sn −1 + α2sn − 2 + ... + αn −1s + αn • Expend it into an infinite power series as ĝ(s) = h (0) + h (1)s −1 + h ( 2)s −2 + ... ( h(0) = 0 for strictly proper) • The coefficient h(m) are called Markov parameters. • Let g(t) be the inverse Laplace transform of ĝ(s). Then, we have h( m) = d m −1 dt m −1 g( t ) t = 0 • Hankel matrix (finding Markov parameters) h ( 2) h(3) ⎡ h (1) ⎢ h ( 2) h(3) h ( 4) ⎢ T( α, β) = ⎢ h(3) h ( 4) h(5) ⎢ . . ⎢ . ⎢⎣ h( α) h( α + 1) h ( α + 2) h(1) ⎤ h(β + 1) ⎥ ⎥ . h(β + 2) ⎥ ⎥ . . ⎥ . h ( α + β − 1)⎥⎦ . . h(2) = -α1h(1) + β2; h(1) = β1; h(3) = -α1h(2) - α2h(1) + β3; … h(n) = -α1h(n-1)-α2h(n-2)- … -αn-1h(1)+βn • Theorem 7.7 A strictly proper rational function ĝ(s) has degree n if and only if ρT(n, n) = ρT(n+k, n+l) = n where ρ denotes the rank. 7.6 Degree of transfer matrices • Given a proper rational matrix Ĝ(s) , assume that every entry of Ĝ(s)is a coprime fraction. • Definition 7.1 The characteristic polynomial of Ĝ(s) is defined as the least common denominator of all minors of Ĝ(s) . Its degree is defined as the degree of Ĝ(s) . 7.7 Minimal realizations-Matrix case • Theorem 7.M2 A state equation (A, B, C, D) is a minimal realization of a proper rational matrix Ĝ(s)if and only if (A, B) is controllable and (A, C) is observable or if and only if dim A = deg Ĝ(s) • Theorem 7.M3 All minimal realizations of Ĝ ( s ) are equivalent. 7.8 Matrix polynomial fractions • The degree of the scalar transfer function ĝ(s) = N (s) = N (s) D −1(s) = D −1(s) N (s) D(s ) is defined as the degree of D(s) if N(s) and D(s) are coprime fraction. • Every q×p proper rational matrix can be expressed as (right fraction polynomial) Ĝ (s) = N (s) D −1(s) • The expression (left polynomial fraction) Ĝ (s) = D −1(s) N (s) • The right fraction is not unique (The same holds for left fraction) Ĝ (s) = [ N (s) R (s)][ D(s) R (s)]−1 = N (s) D −1(s) • Definition 7.2 A square polynomial matrix M(s) is called a unimodular matrix if its determinant is nonzero and independent of s. • Definition 7.3 A square polynomial matrix R(s) is a greatest common right divisor (gcrd) of D(s) and N(s) if (i) R(s) is a common right divisor of D(s) N(s) (ii) R(s) is a left multiple of every common right divisor of D(s) and N(s). If a gcrd is a unimodular matrix, then D(s) and N(s) are said to be right coprime. • Definition 7.4 Consider Ĝ (s) = N (s) D −1(s) (right coprime) = D −1(s) N (s) (left coprime) Then, its characteristic polynomial is defined as det D(s) or det D(s) and its degree is defined as deg Ĝ(s) = deg detD(s) = deg det D(s) • 7.8.1 Column and row reducedness • Define δciM(s) = degree of ith column of M(s) δriM(s) = degree of ith row of M(s) ⎡s + 1 s3 − 2s + 5 − 1⎤ • For example: M(s) = ⎢ ⎥ 2 s 0 ⎥⎦ ⎢⎣s − 1 δc1 = 1, δc2 = 3, δc3 = 0, δr1 = 3, and δr2 = 2. • Definition 7.5 A nonsingular matrix M(s) is column reduced if deg detM(s) = sum of all column degrees It is row reduced if deg det M(s) = sum of all row degrees • Let δciM(s) = kci and define Hc(s) = diag(skc1, skc2, …). Then the polynomial matrix M(s) can be expressed as M(s) = MhcHc(s) + Mlc(s) Mhc: The column-degree coefficient matrix Mlc(s): The remaining term and its column has degree less than kci. • M(s) is column reduced⇔Mhc is nonsingular. • Row form of M(s) M(s) = Hr(s)Mhr + Mlr(s) Hr(s) = diag(skr1, skr2, …). Mhr: the row-degree coefficient matrix. • M(s) is row reduced⇔Mhr is nonsingular. • Theorem 7.8 Let D(s) is column reduced, Then N(s)D-1(s) is proper (strictly proper) if and only if δciN(s)≤δciD(s) [δciN(s)<δciD(s)] • 7.8.2 Computing matrix coprime fraction • Consider Ĝ(s) expressed as Ĝ (s) = D −1 (s) N (s) = N (s) D −1 (s) • Imply N (s) D(s) = D(s) N (s) • Assuming D (s) = D0 + D1s + D2s2 + D3s3 + D4s4 N (s) = N 0 + N1s + N 2s2 + N 3s3 + N 4s4 D(s) = D0 + D1s + D2s2 + D3s3 N (s) = N 0 + N1s + N 2s2 + N 3s3 • A generalized resultant (the matrix version) ⎡ D0 ⎢ ⎢ D1 ⎢ D2 ⎢ D SM := ⎢ 3 ⎢ D4 ⎢ ⎢ 0 ⎢ 0 ⎢ ⎢⎣ 0 N0 0 0 0 0 0 N1 D0 N0 0 0 0 N2 D1 N1 D0 N0 0 N3 D2 N2 D1 N1 D0 N4 D3 N3 D2 N2 D1 0 D4 N4 D3 N3 D2 0 0 0 0 0 0 D4 0 N4 0 D3 D4 0 ⎤ ⎡− N0 ⎤ ⎥ 0 ⎥ ⎢⎢ D0 ⎥⎥ 0 ⎥ ⎢ − N1 ⎥ ⎥⎢ ⎥ N 0 ⎥ ⎢ D1 ⎥ =0 ⎥ ⎥ ⎢ N1 − N 2 ⎥⎢ ⎥ N 2 ⎥ ⎢ D2 ⎥ N3 ⎥⎢− N3 ⎥ ⎥⎢ ⎥ D N 4 ⎥⎦ ⎣ 3 ⎦ • Theorem 7.M4 Let µi, be the number of linear independent. Then deg Ĝ(s) = µ1 + µ2 + ... + µp and a right coprime fraction obtained by computing monic null vectors. 7.9 Realization from matrix coprime fraction • Define (for µ1 = 4 and µ2 = 2) ⎡sµ1 H (s) := ⎢ ⎢⎣ 0 0 ⎤ ⎡s4 0 ⎤ , and = µ2 ⎥ ⎢ 2⎥ s ⎥⎦ ⎢⎣ 0 s ⎥⎦ ⎡sµ1 −1 0 ⎤ ⎡ s3 ⎢ ⎥ ⎢ 2 . ⎥ ⎢s ⎢ . ⎢ 1 0 ⎥ ⎢s = L(s) := ⎢ µ 2 −1 ⎥ ⎢ s ⎢ 0 ⎥ ⎢1 ⎢ . . ⎥ ⎢0 ⎢ ⎥ ⎢ 1 ⎥⎦ ⎢⎣ 0 ⎢⎣ 0 0⎤ ⎥ 0⎥ 0⎥ ⎥ 0⎥ s⎥ ⎥ 1⎥⎦ • Let ŷ(s) = Ĝ (s) û (s) = N (s) D −1û (s) and define v̂ (s) = D −1(s)û(s) • Then, we have D(s) v̂ (s) = û (s), and ŷ(s) = N(s)v̂(s) • Let define ⎡sµ1 −1 ⎡s3v̂1(s) ⎤ ⎡ x1(s) ⎤ 0 ⎤ ⎥ ⎢ 2 ⎥ ⎢ ⎢ ⎥ x ( s ) s v̂ ( s ) . . 2 ⎥ ⎢ 1 ⎥ ⎢ ⎢ ⎥ ⎢ 1 0 ⎥ ⎡ v̂1(s) ⎤ ⎢ sv̂1(s) ⎥ ⎢ x 3 (s) ⎥ x̂ (s) = L(s) v̂ (s) = ⎢ = ⎥ := ⎢ ⎥ µ 2 −1 ⎥ ⎢ v̂ (s)⎥ ⎢ x ( s ) v̂ ( s ) 0 s ⎥ ⎢ 4 ⎥ ⎢ ⎥⎣ 2 ⎦ ⎢ 1 ⎢ sv̂ (s) ⎥ ⎢ x 5 (s) ⎥ ⎢ . . ⎥ ⎢ ⎥ ⎢ 2 ⎥ ⎢ ⎥ x ( s ) 1 ⎥⎦ ⎢⎣ 0 ⎢⎣ v̂ 2 (s) ⎥⎦ ⎣ 6 ⎦ • Express D(s) as D(S) =DhcH(s) + DlcL(s) • Then we have H (s) v̂ (s) = − D −hc1Dlc x̂ (s) + D −hc1û (s) and β β113 β114 β121 β122 ⎤ ⎡β ŷ(s) = N (s) v̂ (s) = ⎢ 111 112 ⎥ L(s) v̂ (s) β β β β β β 213 214 221 222 ⎦ ⎣ 211 212 ⎡β111 β112 β113 β114 β121 β122 ⎤ =⎢ ⎥ x̂ (s) β β β β β β 213 214 221 222 ⎦ ⎣ 211 212