APPENDIX I

Derivation of the Time
Domain Solution of
State Equations
To Accompany

Control Systems Engineering
4th Edition
By
Norman S. Nise

John Wiley & Sons

To order books please call 1 (800) 225-5945. or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center. mechanical. NJ 07030 USA .Copyright © 2004 by John Wiley & Sons. (978) 750-8400. All rights reserved. 111 River Street Hoboken. ISBN 0-471-44577-0 John Wiley & Sons. 111 River Street. without either the prior written permission of the Publisher. scanning or otherwise. Inc. recording. except as permitted under Sections 107 or 108 of the 1976 United States Copyright Act. Danvers. fax (978) 750-4470. photocopying.. stored in a retrieval system or transmitted in any from or by any means. (201) 748-6011. John Wiley & Sons. Hoboken. 222 Rosewood Drive. Inc. NJ 07030. Inc. electronic. email: PERMREQ@WILEY. fax (201) 748-6008.COM. Requests to the Publisher for permission should be addressed to the Permissions Department. MA 01923. No part of this publication may be reproduced.

4a) (I. Thus. we assume a series solution.2) into (I. assume a homogeneous state equation of the form x (t ) = Ax(t ) Since we want to solve for x.3) Equating like coefficients yields b1 = Ab 0 1 1 b 2 = Ab1 = A 2b 0 2 2 # 1 bk = A k b0 k! 1 bk +1 = A k +1b 0 (k + 1)! # Substituting these values into Eq. (I.2) yields (I.1 Derivation Rather than using the Laplace transformation.1) we get b1 + 2b 2t + " + kbk t k −1 + (k + 1)bk +1t k + " =A(b 0 + b1t + b 2t 2 + " + bk t k + bk +1t k +1 + ") (I.4b) (I.1) (I.4d) .A P P E N D I X I Derivation of the Time Domain Solution of State Equations I. (I. We will find that the final solution consists of two parts that are different from the forced and natural responses First. just as we did in elementary scalar differential equations. x(t ) = b 0 + b1t + b 2t 2 + " + bk t k + bk +1t k +1 + " (I.2) Substituting Eq.4c) (I. we can solve the equations directly in the time domain using a method closely allied to the classical solution of differential equations.

since it performs a transformation on x(0) . (I. (I. 1 In this development we consider the initial time.   1 1 1 x(t ) =  I + At + A 2t 2 + " + A k t k + A k +1t k +1 + "  x(0) 2 (k + 1)! k!   (I.2). from Eq. 2 The state-transition matrix here is for the initial time t0 = 0 . After completing this development. The symbol. the interested reader should consult Appendix J on the accompanying CD-ROM for the more general solution in terms of initial time t0 ≠ 0 . More generally.8). t0 ≠ 0 . is used to denote e At . x(0) .10) We give a special name to e At : it is called the state-transition matrix2. The derivation in Appendix J on the accompanying CD- ROM for t0 ≠ 0 yields x(t ) = e A ( t − t0 ) x(t0 ) . to the state x(t ) at any time. x(0) = b 0 (I.8) where e At is simply a notation for the matrix formed by the right-hand side of Eq.7) Let   1 1 1 e At =  I + At + A 2t 2 + " + A k t k + A k +1t k +1 + "  k! 2 (k + 1)!   (I. Thus.4 Appendix I Derivation of the Time Domain Solution of State Equations 1 1 1 x(t ) = b 0 + Ab 0t + A 2b 0t 2 + " + A k b 0t k + A k +1b 0t k +1 + " 2 k! (k + 1)!   1 1 1 A k +1t k +1 + "  b 0 =  I + At + A 2t 2 + " + A k t k + k! 2 (k + 1)!   (I. taking x from the initial state.7). t0 . or   1 1 1 e at = 1 + at + a 2t 2 + " + a k t k + a k +1t k +1 + "  2 (k + 1)! k!   (I. We use this definition because the right-hand side of Eq. (I.9) Using Eq. to be 0.8) resembles a power series expansion of e at . (I. t.5) But.6)1 Therefore. Φ(t ) . . we have x(t ) = e At x(0) (I.

13) Hence. or nonhomogeneous. differentiating Eq. Also. From Eq.12) There are some properties of Φ(t ) that we will use later when we solve for x(t ) in the text.I.15)  (0) x(0) = Ax(0) Φ (I. (I. (I. or unforced. the first property of Φ(t ) is Φ(0) = I (I.17) In summary.18) where Φ(0) = I (I.1 Derivation Φ(t ) = e At 5 (I.20) and Let us now solve the forced. (I. (I.12) and setting this equal to Eq.16):  (0) = A Φ (I.1) yields  (t ) x(0) = Ax(t ) x (t ) = Φ (I. x(0) = Φ(0)x(0) (I.14) where I is the identity matrix.11) and x(t ) = Φ(t )x(0) (I. at t = 0 .16) which. Given the forced state equation x (t ) = Ax(t ) + Bu (t ) (I. system is x(t ) = Φ(t ) x(0) (I.21) . the second property of Φ(t ) follows from Eq.19)  (0) = A Φ (I.12). the solution to the homogeneous. problem. yields Thus.

(I. Solving for x(t ) in Eq.23) Integrating both sides yields t t 0 0 [e − At x(t )] = e − Αt x(t ) − x(0) = ∫ e − Ατ Bu (τ ) dτ (I. E. B.22) Realizing that the left-hand side is equal to the derivative of the product e − At x(t ) . ..8)). (I.24) we obtain x(t ) = e − Αt x(0) + t ∫e 0 − Α ( t −τ ) Bu (τ ) dτ = Φ(t ) x(0) + t ∫ Φ(t − τ )Bu(τ )dτ 0 (I. 1968. New York. State Space Analysis: An Introduction. McGraw-Hill.6 Appendix I Derivation of the Time Domain Solution of State Equations rearrange and multiply both sides by e − At : e − At [ x (t ) − Ax(t )] = e − At Bu(t ) (I.24) since e − At evaluated at t = 0 is the identity matrix (from Eq. L. we obtain d − At [e x(t )] = e − Αt Bu(t ) dt (I.25) where Φ(t ) = e At by definition. and Bona. Bibliography Timothy. K.