Chapter 3 CT LTI Systems Updated: 9/16/13 A Continuous-Time System • How do we know the output? X(t) y(t) System LTI Systems • Time Invariant – X(t) y(t) & x(t-to) y(t-to) • Linearity – a1x1(t)+ a2x2(t) a1y1(t)+ a2y2(t) – a1y1(t)+ a2y2(t)= T[a1x1(t)+a2x2(t)] • Meet the description of many physical systems • They can be modeled systematically – Non-LTI systems typically have no general mathematical procedure to obtain solution What is the input-output relationship for LTI-CT Systems? Convolution Integral • An approach (available tool or operation) to describe the input-output relationship for LTI Systems X(t)=d(t) • In a LTI system y(t)=h(t) LTI System d(t) h(t) – Remember h(t) is T[d(t)] – Unit impulse function the impulse response • It is possible to use h(t) to solve for any input-output relationship X(t) y(t) LTI System: h(t) • One way to do it is by using the Convolution Integral Convolution Integral • Remember X(t)=Ad(t-kto) y(t)=Ah(t-kto) LTI System • So what is the general solution for X(t) y(t) LTI System ? Convolution Integral • Any input can be expressed using the unit impulse function Proof: d (t ) d (t ) Sifting Property (t )d (t t )dt (t ) o o x(t ) x( )d (t )d to and integrate by d x(t )d (t to ) x(to )d (t to ) x(t )d (t )d x( )d (t )d x(t ) d (t )d x(t )(1) x(t ) X(t) y(t) LTI System Convolution Integral • Given X(t) y(t) LTI System • We obtain Convolution Integral • That is: A system can be characterized using its impulse response: y(t)=x(t)*h(t) y (t ) T {x(t )} h(t ) T {d (t )} h(t ) T {d (t )} y (t ) T x( )d (t )d Linearity : y (t ) x( )T d (t )d X(t) y(t) LTI System: h(t) y (t ) x( )h(t )d Do not confuse convolution with multiplication! y(t)=x(t)*h(t) By definition Convolution Integral X(t) y(t) LTI System: h(t) Convolution Integral - Properties x(t ) * h(t ) h(t ) * x(t ) [ x(t ) * h1 (t )] * h2 (t ) x(t ) *[h1 (t ) * h2 (t )] • Commutative • Associative • Distributive x(t ) *[h1 (t ) h2 (t )] [ x(t ) * h1 (t )] [ x(t ) * h2 (t )] • Thus, using commutative property: x(t ) x( )h(t )d h( )x(t )d Next: We draw the block diagram representation! Convolution Integral - Properties • • • Commutative Associative Distributive x(t ) * h(t ) h(t ) * x(t ) [ x(t ) * h1 (t )] * h2 (t ) x(t ) *[h1 (t ) * h2 (t )] x(t ) *[h1 (t ) h2 (t )] [ x(t ) * h1 (t )] [ x(t ) * h2 (t )] Simple Example • What if a step unit function is the input of a LTI system? u(t) y(t)=S{u(t)}=s(t) LTI System • S(t) is called the Step Response y (t ) S{u (t )} y (t ) s (t ) h(t ) x(t ) h(t ) u (t ) y (t ) s (t ) u ( )h(t )d h( )u (t )d t h( )d Step response can be obtained by integrating the impulse response! Note : h(t ) y (t ) / dt s (t ) / dt Impulse response can be obtained by differentiating the step response Example 1 • Consider a CT-LTI system. Assume the impulse response of the system is h(t)=e^(-at) for all a>0 and t>0 and input x(t)=u(t). Find the output. y (t ) h(t ) x(t ) h(t ) u (t ) u(t) h(t)=e^-at y (t ) h( )u (t )d (e a y(t) Because t>0 u ( ) u (t )d 1 ( e d (e a t a 0 at 1) Draw x(), h(), h(t-),etc. next slide 1 (1 e at )u (t ) a The fact that a>0 is not an issue! Example 1 – Cont. y(t); for a=3 y(t) * t t y (t ) h(t ) x(t ) h(t ) u (t ) U(-(-t)) t<0 U(-(-t)) t>0 y (t ) h( )u (t )d (e a u ( ) u (t )d Remember we are plotting it over and t is the variable 1 ( e d (e a t a 0 1 (1 e at )u (t ) a at 1) Example 1 – Cont. Graphical Representation (similar) a=1 In this case! http://www.wolframalpha.com/input/?i=convolution+of+two+functions&lk=4&num=4&lk=4&num=4 Example 1 – Cont. Graphical Representation (similar) Note in our case We have u(t) rather than rectangular function! http://www.jhu.edu/signals/convolve/ Example 2 • Consider a CT-LTI system. Assume the impulse response of the system is h(t)=e^-at for all a>0 and t>0 and input x(t)= e^at u(-t). Find the output. x(t) y(t) h(t)=e^-at y (t ) h(t ) x(t ) h(t ) u (t ) y (t ) x( )h(t )d (e a ( u ( ) e a ( t ) u (t ) d Note that for t>0; x(t) =0; so the integration can only be valid up to t=0 ( (e d (e 21a e t 0 e at t 2 a at 2 at 1 at e 2a ( (e d (e 21a [1 0] 21a e t 0 e at 0 2 a at at y (t ) 1 a t e a 0 2a Draw x(), h(), h(t-),etc. next slide Example – Cont. x(t)= e^at u(-t) * y (t ) h(t ) x(t ) h(t ) u (t ) y (t ) h( )x(t )d ? y (t ) h(t ) x(t ) h(t ) u (t ) 1 a t y (t ) e a 0 2a h(t)=e^-at u(t) Another Example notes Properties of CT LTI Systems • When is a CT LTI system memory-less (static) h(t ) Kd (t ) y (t ) Kx(t ) • When does a CT LTI system have an inverse system (invertible)? h(t ) * hi (t ) d (t ) • When is a CT LTI system considered to be causal? Assuming the input is causal: t t 0 0 y(t ) x( )h(t )d h( )x(t )d • When is a CT LTI system considered to be Stable? notes y (t ) h(t ) dt Example • Is this an stable system? h(t ) e 3t u (t ) y (t ) h(t ) dt 0 3t 3t e u ( t ) dt e dt 1 / 3 • What about this? notes h(t ) e3t u (t ) Differential-Equations Models • This is a linear first order differential equation with constant coefficients (assuming a and b are constants) • The general nth order linear DE with constant equations is Is the First-Order DE Linear? • Consider • Does a1x1(t)+ a2x2(t) a1y1(t)+ a2y2(t)? notes • Is it time-invariant? Does input delay results in an output delay by the same amount? notes • Is this a linear system? + X(t) Sum e(t) a Integrator Y(t) Example • Is this a time invariant linear system? R V(t) Ldi(t)/dt + Ri(t)= v(t) a= -R/L b=1/L y(t)=i(t) x(t) = V(t) L Solution of DE • A classical model for the solution of DE is called method of undermined coefficients – yc(t) is called the complementary or natural solution – yp(t) is called the particular or forced solution Solution of DE Thus, for x(t) =constant yp(t)=P x(t) =Ce^-7t yp(t)= Pe^-7t x(t) =2cos(3t) yp(t)=P1cos(3t)+P2sin(3t) Example • Solve – Assume x(t) = 2 and y(0) = 4 yc(t) = Ce^-2t; yp(t) = P; P = 1 y(t) = Ce^-2t + 1 y(0) = 4 C = 3 y(t) = 3e^-2t + 1 – What happens if notes Schaum’s Examples • Chapter 2: – 2, 4-6, 8, 10, 11-14, 18, 19, 48, 52, 53,