Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Eigenvalues, Zeros and Poles x Ax Bu y Cx Du , x(0) x0 x Rn , y Rr , (1) u Rm Definition: Roots of the characteristic polynomial of a system are called eigenvalues of the system. characteristic p( ) det( I A) ( 1 )( 2 )...( n ) polynomial eigenvalues i , i 1,2,...,n . nxn adj ( sI A) Ri X (s) [sI A] x0 x0 x0 det[ sI A] i 1 ( s i ) n 1 n x(t ) i eit , i 1 i ˆ Ri x0 nx1 Transfer function of the system (1): Y (s) H (s) C ( sI A) 1 B D U (s) C[ adj ( sI A)]B D[det( sI A)] [det( sI A)] In general H (s ) is a matrix. H ij (s ) is then the transfer function from the input U j (s ) to the output Yi (s ) . After cancelling the common terms in nominator and denominator, one gets W (s) H ij (s ) ( s) Definition: Poles of H ij (s ) (or the system (1)) are the roots of (s ). Zeros of H ij (s ) (or the system (1)) are the roots of W (s ). Result: Poles of a system are a subset of system eigenvalues. Stability x (t ) Ax(t ) Bu (t ) y (t ) Cx(t ) Du (t ) Zero input stability: x (t ) Ax(t ), x(t0 ) x0 (2) Definition: (Equilibrium) A constant solution x (t ) xd , t t 0 of the system (2) is called an equilibrium of the system x f (x) . How to find xd ? What is xd for linear systems? Solve 0 f ( xd ) ! If A is invertible then xd is zero. Otherwise, there are infinitely many equilibria. Solve 0 Axd . Definition (Norm): Let V be a vector space (a space with appropriate addition and multiplication by scalars operations). Norm of a vector is defined as a positive valued function that satisfies. x 0 x0 . :V R x x x y x y , x, y V Generalization of the notion ........... Definition: (Lyapunov stability) Let xd be an equilibrium of the system x (t ) f ( x (t )) The equilibrium is called Lyapunov stable if for every arbitrary small 0 , there exists a ( ) such that x(t0 ) xd ( , t0 ) x(t ) xd , t t0 Definition: (Asymptotic stability) Let xd be a Lyapunov stable equilibrium of the system x (t ) f Then xd is asymptotically stable if there exists a 0 such that x(t0 ) - xd < d ( x(t )) lim x(t ) xd 0, t t0 t Theorem: The zero equilibrium of the system x (t ) Ax(t ) is Lyapunov stable if and only if (t , t0 ) c, t t0 Proof: (if) Solution: x(t ) (t , t0 ) x0 , t t0 x(t ) (t , t0 ) x0 (t , t0 ) x0 Norm property (t , t0 ) c, t t0 x(t ) (t , t0 ) x0 (t , t0 ) x0 c x0 ( , t0 ) ̂ (only if) c x(t ) Let xd 0 be Lyapunov stable but (t , t 0 ) not bounded. Then there exist a ji (t , t 0 ) which is not bounded. Choose x0 0 ... 0 1 0 i. ... 0 T Then, x(t ) (t , t0 ) x0 ji (t , t0 ) ji (t , t0 ) c x(t ) This result is contraditory. Therefore, (t , t0 ) c Theorem: The zero equilibrium of the system x (t ) Ax(t ) is Lyapunov stable if and only if lim (t , t 0 ) 0, t t 0 t Proof: Similar to the previous proof. Theorem: 1) For the system let the x (t ) Ax(t ) , eigenvalues of the system be i , i 1,2,... n . Then, x (t ) Ax(t ) is Lyapunov stable Re i 0, i 1,2,...., n and eigenvalues with Re i 0 have multiplicity one. 2) x (t ) Ax(t ) is asymptotically stable Re i 0, i 1,2,...., n x (t ) Ax(t ) Bu (t ), x(t0 ) 0 y (t ) Cx (t ) Du (t ) Definition: (BIBO sability) The system is bounded input bounded output (BIBO) stable if and only if output of the system is bounded for all bounded inputs. Theorem: is BIBO stable All poles of have negative real part. Theorem: is asymptotically stable is BIBO stable. Example: Find the equilibria and analyze the stability for the following system! é 2 -x x + x 2 1 1 -1 ê x= ê -x2 x1 ë ù ú ú û Example: Find the transfer function for the following system! Is the system a) Lyapunov stable? b) asymptotically stable? c) BIBO stable? 2 x t 0 0 y t 1 1 2 1 1 3 1 xt 1 u t 0 0 1 1 xt