Department of Electrical Engineering University of Arkansas ELEG 3124 SYSTEMS AND SIGNALS Lecture Notes Dr. Jingxian Wu wuj@uark.edu This work is licensed under: Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0) OUTLINE • Chapter 1: Continuous-Time Signals ………………………. 3 • Chapter 2: Continuous-Time Systems ……………………… 45 • Chapter 3: Fourier Series ……………………………………. 84 • Chapter 4: Fourier Transform ……………………………… 122 • Chapter 5: Laplace Transform ……………………………… 170 • Chapter 6: Discrete-time Signals and Systems ……………… 222 2 Department of Electrical Engineering University of Arkansas ELEG 3124 SYSTEMS AND SIGNALS Ch. 1 Continuous-Time Signals Dr. Jingxian Wu wuj@uark.edu OUTLINE • Introduction: what are signals and systems? • Signals • Classifications • Basic Signal Operations • Elementary Signals 4 INTRODUCTION • Examples of signals and systems (Electrical Systems) – Voltage divider • Input signal: x = 5V • Output signal: y = Vout Voltage divider 𝑅2 𝑥) 1 +𝑅2 • The system output is a fraction of the input (𝑦 = 𝑅 – Multimeter • Input: the voltage across the battery • Output: the voltage reading on the LCD display • The system measures the voltage across two points multimeter – Radio or cell phone • Input: electromagnetic signals • Output: audio signals • The system receives electromagnetic signals and convert them to audio signal INTRODUCTION • Examples of signals and systems (Biomedical Systems) – Central nervous system (CNS) • Input signal: a nerve at the finger tip senses the high temperature, and sends a neural signal to the CNS • Output signal: the CNS generates several output signals to various muscles in the hand • The system processes input neural signals, and generate output neural signals based on the input – Retina • Input signal: light • Output signal: neural signals • Photosensitive cells called rods and cones in the retina convert incident light energy into signals that are carried to the brain by the optic nerve. Retina INTRODUCTION • Examples of signals and systems (Biomedical Instrument) – EEG (Electroencephalography) Sensors • Input: brain signals • Output: electrical signals • Converts brain signal into electrical signals EEG signal collection – Magnetic Resonance Imaging (MRI) • Input: when apply an oscillating magnetic field at a certain frequency, the hydrogen atoms in the body will emit radio frequency signal, which will be captured by the MRI machine • Output: images of a certain part of the body • Use strong magnetic fields and radio waves to form images of the body. MRI INTRODUCTION • Signals and Systems – Even though the various signals and systems could be quite different, they share some common properties. – In this course, we will study: • How to represent signal and system? • What are the properties of signals? • What are the properties of systems? • How to process signals with system? – The theories can be applied to any general signals and systems, be it electrical, biomedical, mechanical, or economical, etc. OUTLINE • Introduction: what are signals and systems? • Signals • Classifications • Basic Signal Operations • Elementary Signals 9 SIGNALS AND CLASSIFICATIONS • What is signal? – Physical quantities that carry information and changes with respect to time. – E.g. voice, television picture, telegraph. • Electrical signal – Carry information with electrical parameters (e.g. voltage, current) – All signals can be converted to electrical signals • Speech → Microphone → Electrical Signal → Speaker → Speech audio signal – Signals changes with respect to time 10 SIGNALS AND CLASSIFICATIONS • Mathematical representation of signal: – Signals can be represented as a function of time t s(t ), – Support of signal: t1 t t2 – E.g. s1 (t ) = sin( 2t ) – E.g. s2 (t ) = sin( 2t ) • t1 t t2 − t + 0t s1 (t ) and s2 (t ) are two different signals! – The mathematical representation of signal contains two components: • The expression: s(t ) t1 t t2 • The support: – The support can be skipped if − t + s1 (t ) = sin( 2t ) – E.g. 11 SIGNALS AND CLASSIFICATIONS • Classification of signals: signals can be classified as – – – – – – – Continuous-time signal v.s. discrete-time signal Analog signal v.s. digital signal Finite support v.s. infinite support Even signal v.s. odd signal Periodic signal v.s. Aperiodic signal Power signal v.s. Energy signal …… 12 OUTLINE • Introduction: what are signals and systems? • Signals • Classifications • Basic Signal Operations • Elementary Signals 13 14 SIGNALS: CONTINUOUS-TIME V.S. DISCRETE-TIME • Continuous-time signal – If the signal is defined over continuous-time, then the signal is a continuous-time signal s(t ) = sin( 4t ) • E.g. sinusoidal signal • E.g. voice signal • E.g. Rectangular pulse function p( t ) A A, 0 t 1 p( t ) = 0, otherwise 0 1 Rectangular pulse function t SIGNALS: CONTINUOUS-TIME V.S. DISCRETE-TIME 15 • Discrete-time signal – If the time t can only take discrete values, such as, t = kTs k = 0,1,2, then the signal s(t ) = s(kTs ) is a discrete-time signal – E.g. the monthly average precipitation at Fayetteville, AR (weather.com) Ts = 1 month k = 1, 2, , 12 Monthly average precipitation – What is the value of s(t) at (k −1)Ts t kTs ? • Discrete-time signals are undefined at t kTs !!! • Usually represented as s(k) 16 SIGNALS: ANALOG V.S. DIGITAL • Analog v.s. digital – Continuous-time signal • continuous-time, continuous amplitude→ analog signal – Example: speech signal • Continuous-time, discrete amplitude – Example: traffic light – Discrete-time signal • Discrete-time, discrete-amplitude → digital signal – Example: Telegraph, text, roll a dice 2 1 3 0 2 1 0 3 2 1 0 2 1 0 • Discrete-time, continuous-amplitude – Example: samples of analog signal, average monthly temperature Different types of signals 17 SIGNALS: EVEN V.S. ODD • Even v.s. odd – x(t) is an even signal if: x(t ) = x( −t ) • E.g. x(t ) = cos(2t ) – x(t) is an odd signal if: x( −t ) = − x(t ) • E.g. x(t ) = sin( 2t ) – Some signals are neither even, nor odd x(t ) = cos(2t ), t 0 • E.g. x (t ) = et – Any signal can be decomposed as the sum of an even signal and odd signal y(t ) = ye (t ) + yo (t ) even • proof odd SIGNALS: EVEN V.S. ODD • Example – Find the even and odd decomposition of the following signal x (t ) = et SIGNALS: EVEN V.S. ODD • Example – Find the even and odd decomposition of the following signal t0 2 sin( 4t ), x (t ) = 0 otherwise 19 SIGNALS: PERIODIC V.S. APERIODIC • Periodic signal v.s. aperiodic signal – An analog signal is periodic if • There is a positive real value T such that s(t ) = s(t + nT ) • It is defined for all possible values of t, − t (why?) – Fundamental period T0 : the smallest positive integer T0 that satisfies s (t ) = s (t + nT0 ) • E.g. T1 = 2T0 s(t + nT1 ) = s(t + 2nT0 ) = s(t ) – So T1 is a period of s(t), but it is not the fundamental period of s(t) 20 21 SIGNALS: PERIODIC V.S. APERIODIC • Example – Find the period of – – – – – s(t ) = Acos(0t + ) Amplitude: A Angular frequency: 0 Initial phase: Period: T0 = Linear frequency: f 0 = −t 22 SIGNALS: PERIODIC V.S. APERIODIC • Complex exponential signal – Euler formula: e jx = cos( x) + j sin( x) – Complex exponential signal e j0t = cos(0t ) + j sin( 0t ) – Complex exponential signal is periodic with period T0 = • Proof: Complex exponential signal has same period as sinusoidal signal! 2 0 SIGNALS: PERIODIC V.S. APERIODIC • The sum of two periodic signals – x(t) has a period T1 – y(t) has a period T2 – Define z(t) = a x(t) + b y(t) – Is z(t) periodic? z(t + T ) = ax(t + T ) + by(t + T ) • In order to have x(t)=x(t+T), T must satisfy T = kT1 • In order to have y(t)=y(t+T), T must satisfy T = lT 2 • Therefore, if T = kT = lT 1 2 z (t + T ) = ax(t + kT1 ) + by(t + lT2 ) = ax(t ) + by(t ) = z (t ) • The sum of two periodic signals is periodic if and only if the ratio of the two periods can be expressed as a rational number. T1 l = T2 k • The period of the sum signal is T = kT1 = lT2 23 24 SIGNALS: PERIODIC V.S. APERIODIC • Example x (t ) = sin( – – – – • 3 t) 2 y (t ) = exp( j t) 9 x(t ), y(t ), z (t ) 2 z (t ) = exp( j t ) 9 Find the period of Is 2 x(t ) − 3 y(t )periodic? If periodic, what is the period? Is x(t ) + z(t ) periodic? If periodic, what is the period? Is y(t ) z(t ) periodic? If periodic, what is the period? Aperiodic signal: any signal that is not periodic. 25 SINGALS: ENERGY V.S. POWER • Signal energy – – – – Assume x(t) represents voltage across a resistor with resistance R. Current (Ohm’s law): i(t) = x(t)/R Instantaneous power: p (t ) = x 2 (t ) / R 2 Signal power: the power of signal measured at R = 1 Ohm: p(t ) = x (t ) – Signal energy at: [tn , tn + t ] En p(tn )t p(t ) p (t n ) – Total energy E = lim t →0 p(t n + )t = + 2 n − t p(t )dt E = x(t ) dt tn Instantaneous power − – Review: integration over a signal gives the area under the signal. t SINGALS: ENERGY V.S. POWER t [−,+] • Energy of signal x(t) over E = x(t ) dt 2 − – If 0 E , then x(t) is called an energy signal. • Average power of signal x(t) 1 P = lim T → 2T – If 0 P , • T −T 2 x (t ) dt then x(t) is called a power signal. A signal can be an energy signal, or a power signal, or neither, but not both. 26 27 SINGALS: ENERGY V.S. POWER • Example 1: • • Example 2: • Example 3: x(t ) = A exp( −t ) t0 x(t ) = Asin( 0t + ) x(t ) = (1 + j )e jt 0 t 10 All periodic signals are power signal with average power: 1 P= T T 0 2 x (t ) dt OUTLINE • Introduction: what are signals and systems? • Signals • Classifications • Basic Signal Operations • Elementary Signals 28 29 OPERATIONS: SHIFTING • Shifting operation – x(t − t0 ) : shift the signal x(t) to the right by t0 Shifting to the right by two units – Why right? x(0) = A y(t ) = x(t − t0 ) x(0) = y (t0 ) y(t0 ) = x(t0 − t0 ) = x(0) = A OPERATIONS: SHIFTING • Example – Find t +1 −1 t 0 1 0t 2 x(t ) = − t + 3 2 t 3 0 o.w. x(t + 3) 30 31 OPERATIONS: REFLECTION • Reflection operation – x(−t ) is obtained by reflecting x(t) w.r.t. the y-axis (t = 0) x(-t) x(t) 2 2 1 1 t t -2 -1 1 2 -3 3 -2 -1 1 -1 -1 Reflection OPERATIONS: REFLECTION • Example: t + 1 − 1 t 0 x(t ) = 1 0 t 2 0 o.w. – Find x(3-t) • The operations are always performed w.r.t. the time variable t directly! 32 33 OPERATIONS: TIME-SCALING • Time-scaling operation – x(at ) is obtained by scaling the signal x(t) in time. • a 1 , signal shrinks in time domain • a 1 , signal expands in time domain x(2t) x(t/2) x(t) 2 2 1 1 2 1 t t -1 1 -1.5 -1 -0.5 0.5 Time scaling 1 1.5 t -2 -1 1 2 OPERATIONS: TIME-SCALING • Example: – Find t +1 −1 t 0 1 0t 2 x(t ) = − t + 3 2 t 3 0 o.w. x(3t − 6) x(at + b) 1. scale the signal by a: y(t) = x(at) 2. left shift the signal by b/a: z(t) = y(t+b/a) = x(a(t+b/a))=x(at+b) • The operations are always performed w.r.t. the time variable t directly (be careful about –t or at)! 34 OUTLINE • Signals • Classifications • Basic Signal Operations • Elementary Signals 35 36 ELEMENTARY SIGNALS: UNIT STEP FUNCTION • Unit step function u(t) 1, t 0 u (t ) = 0, t 0 1 t 1 Unit step function • Example: rectangular pulse 1 , − t p ( t ) = 2 2 otherwise 0, Express p (t ) as a function of u(t) u(t) 1/ à t - à /2 à /2 Rectangular pulse 37 ELEMENTARY SIGNALS: RAMP FUNCTION • The Ramp function r (t ) r (t ) = t u(t ) 0 t Unit ramp function – The Ramp function is obtained by integrating the unit step function u(t) t − u (t )dt = 38 ELEMENTARY SIGNALS: UNIT IMPULSE FUNCTION • Unit impulse function (Dirac delta function) (0) = (t ) = 0, t 0 t − (t ) 1, t 0 0, t 0 (t )dt = t 0 Unit impulse function – delta function can be viewed as the limit of the rectangular pulse (t ) = lim pΔ (t ) →0 u(t) 1/ à – Relationship between (t ) and u(t) t - à /2 t − (t )dt = u(t ) (t ) = du (t ) dt à /2 Rectangular pulse ELEMENTARY SIGNALS: UNIT IMPULSE FUNCTION • Sampling property x(t ) (t − t0 ) = x(t0 ) (t − t0 ) • Shifting property + − – Proof: x(t ) (t − t0 )dt = x(t0 ) 39 ELEMENTARY SIGNALS: UNIT IMPULSE FUNCTION • Scaling property 1 b (at + b) = t + |a| a – Proof: 40 ELEMENTARY SIGNALS: UNIT IMPULSE FUNCTION • Examples 4 (t + t 2 ) (t − 3)dt = −2 1 −2 3 −2 (t + t 2 ) (t − 3)dt = exp(t − 1) (2t − 4)dt = 41 42 ELEMENTARY SIGNALS: SAMPLING FUNCTION sa(t) • Sampling function Sa ( x ) = sin x x t Sampling function – Sampling function can be viewed as scaled version of sinc(x) Sinc ( x) = sin x = sa (x) x sinc(t) 1 t -4 -3 -2 -1 1 2 3 4 Sinc function ELEMENTARY SIGNALS: COMPLEX EXPONENTIAL • Complex exponential x(t ) = e( r+ j0 )t – Is it periodic? • Example: ( −1+ j 2 ) t – Use Matlab to plot the real part of x(t ) = e [u(t + 2) − u(t − 4)] 43 44 SUMMARY • Signals and Classifications – – – – – – Mathematical representation s (t ), Continuous-time v.s. discrete-time Analog v.s. digital Odd v.s. even Periodic v.s. aperiodic Power v.s. energy t1 t t2 • Basic Signal Operations – Time shifting – reflection – Time scaling • Elementary Signals – Unit step, unit impulse, ramp, sampling function, complex exponential Department of Electrical Engineering University of Arkansas ELEG 3124 SYSTEMS AND SIGNALS Ch. 2 Continuous-Time Systems Dr. Jingxian Wu wuj@uark.edu 46 OUTLINE • Classifications of continuous-time system • Linear time-invariant system (LTI) • Properties of LTI system • System described by differential equations 47 CLASSIFICATIONS: SYSTEM DEFINITION • What is system? – A system is a process that transforms input signals into output signals • Accept an input • Process the input • Send an output (also called: the response of the system to input) – System examples: • Radio: input: electrical signals from air, output: music • Robot: input: electrical control signals, output: motion or action • Continuous-time system – A system in which continuous-time input signals are transformed to continuous-time output signals • Discrete-time system – A system in which discrete-time input signals are transformed to discrete-time output signals. x(t ) continuous-time System Continuous-time system y(t ) x(n) Discrete-time System discrete-time system y (n) 48 CLASSIFICATIONS: SYSTEM DEFINITION • Classifications – – – – – – Linear v.s. non-linear Time-invariant v.s. time-varying Dynamic v.s. static (memory v.s. memoryless) Causal v.s. non-causal Invertible v.s. non-invertible Stable v.s. non-stable 49 CLASSIFICATIONS: LINEAR AND NON-LINEAR • Linear system – Let y1 (t )be the response of a system to an input x1 (t ) – Let y (t )be the response of a system to an input x2 (t ) 2 – The system is linear if the superposition principle is satisfied: • 1. the response to is x1 (t ) + x2 (t ) y1 (t ) + y2 (t ) • 2. the response to is x1 (t ) x1 (t ) + x2 (t ) y1 (t ) Linear System y1 (t ) + y2 (t ) Linear system • Non-linear system – If the superposition principle is not satisfied, then the system is a non-linear system 50 CLASSIFICATIONS: LINEAR AND NON-LINEAR • Example: check if the following systems are linear – System 1: y(t ) = exp[ x(t )] – System 2: charge a capacitor. Input: i(t), output v(t) v (t ) = 1 t i ( )d C − – System 3: inductor. Input: i(t), output v(t) v (t ) = L di (t ) dt 51 CLASSIFICATIONS: LINEAR AND NON-LINEAR • Example – System 4: – System 5: y(t ) =| x(t ) | – System 6: y(t ) = x2 (t ) CLASSIFICATIONS: LINEAR V.S. NON-LINEAR • Example: – Amplitude Modulation: • Is it linear? Amplitude modulation 52 CLASSIFICATIONS: TIME-VARYING V.S. TIME-INVARIANT • Time-invariant – A system is time-invariant if a time shift in the input signal causes an identical time shift in the output signal x(t ) Time-invariant System y(t ) x(t − t0 ) Time-invariant system • Examples – y(t) = cos(x(t)) t – y(t ) = 0 x(v)dv Time-invariant System y(t − t0 ) 53 54 CLASSIFICATIONS: MEMORY V.S. MEMORYLESS • Memoryless system – If the present value of the output depends only on the present value of input, then the system is said to be memoryless (or instantaneous). – Example: input x(t): the current passing through a resistor output y(t): the voltage across the resistor y(t ) = Rx(t ) – The output value at time t depends only on input value at time t. • System with memory – If the present value of the output depends on not only present value of input, but also previous input values, then the system has memory. – Example: capacitor, current: x(t), output voltage: y(t) 1 t y (t ) = x ( ) d C 0 – the output value at t depends on all input values before t 55 CLASSIFICATIONS: MEMORY V.S. MEMORYLESS • Examples: determine if the systems has memory or not – N y (t ) = ai x(t − Ti ) i =0 – y(t ) = sin( 2x2 (t ) + ) x(t ) 56 CLASSIFICATIONS: CAUSAL V.S. NON-CAUSAL • Causal system – A system is causal if the output only on values of input y (tdepends 0) for t t0 • The output depends on only input from the past and present – Example y (t ) = 1 C t x( )d 0 • Non-causal system – A system is non-causal if the output depends on the input from the future (prediction). – Examples: 1 T /2 y ( t ) = x( )d y(t ) = x(t + a) a0 − T / 2 T – The output value at t depends on the input value at t + a (from future) – All practical systems are causal. 57 CLASSIFICATION: INVERTIBILITY • Invertible – A system is invertible if • by observing the output, we can determine its input. x(t ) y(t ) System Inverse System x(t ) invertible system – Question: for a system, if two different inputs result in the same output, is this system invertible? • Example y(t ) = 2 x(t ) y (t ) = cosx(t ) – If two different inputs result in the same output, the system is noninvertible 58 CLASSIFICATION: STABILITY • Bounded signal – Definition: a signal x(t) is said to be bounded if | x(t ) | B t • Bounded-input bounded-output (BIBO) stable system – Definition: a system is BIBO stable if, for any bounded input x(t), the response y(t) is also bounded. t | x(t ) | B1 | y(t) | B2 • Example: determine if the systems are BIBO stable y (t ) = expx(t ) t y (t ) = x( )d − 59 OUTLINE • Classifications of continuous-time system • Linear time-invariant system (LTI) • Properties of LTI system • System described by differential equations 60 LTI: DEFINTION • Linear time-invariant (LTI) system – Definition: a system is said to be LTI if it’s linear and time-invariant xi (t ) System yi (t ) system – Linear Input: N x(t ) = a1 x1 (t ) + a2 x2 (t ) + + a N x N (t ) = ai xi (t ) i =1 N Output: y (t ) = a1 y1 (t ) + a2 y2 (t ) + + a N y N (t ) = ai yi (t ) i =1 – Time-invariant x(t ) = xi (t − t0 ) Input: Output: y(t ) = yi (t − t0 ) 61 LTI: IMPULSE RESPONSE • Impulse response of LTI system – Def: the output (response) of a system when the input is a unit impulse function (delta function). • Usually denoted as h(t) x(t ) = (t ) y(t ) = h(t ) System LTI system • For system with an arbitrary input x(t), we want to find out the output y(t). – Method 1: differential equations – Methods 2: convolution integral – Methods 3: Laplace transform, Fourier transform, 62 LTI: CONVOLUTION • Derivation – Any signal can be approximated by the sum of a sequence of delta functions + + − z ( )d = lim →0 z ( n ) n = − + x(t ) = x( ) (t − )d = lim − →0 + x(n) (t − n) n = − x(t) t integration 63 LTI: CONVOLUTION • Derivation – Any signal can be approximated by the sum of a sequence of delta functions + x(t ) = x( ) (t − )d = lim − →0 (t ) + x(n) (t − n) n = − h(t ) System – Time invariant (t − n) h(t − n) System – Linear + x(n) (t − n) n = − + x(n)h(t − n) System LTI system n = − 64 LTI: CONVOLUTION • Convolution + x(t ) System y(t ) = x( )h(t − )d − LTI system – Definition: the convolution of two signals x(t) and h(t) is defined as + y(t ) = x( )h(t − )d − – The operation of convolution is usually denoted with the symbol + y(t ) = x(t ) h(t ) = x( )h(t − )d − x(t ) h(t ) x(t ) h(t) LTI system For LTI system, if we know input x(t) and impulse response h(t), Then the output is x(t ) h(t ) 65 LTI: CONVOLUTION • Examples x(t ) (t ) x(t ) (t − t0 ) x(t ) u(t ) 66 LTI: CONVOLUTION • Examples exp( −bt )u(t ) exp( −at )u(t ) LTI system y(t ) = ? 67 LTI: CONVOLUTION • Example – Obtain the impulse response of a capacitor and use it to find the unit-step response by using convolution. Assume the input is the current, and the output is the voltage. Let C = 1F. v (t ) = 1 t i ( )d C − 68 LTI: CONVOLUTION PROPERTIES • Commutativity x(t ) y(t ) = y(t ) x(t ) – Proof: + x(t ) y(t ) = x( ) y(t − )d − x(t ) h(t ) x(t ) h(t) ➔ commutativity h(t ) x(t ) h(t ) x(t) 69 LTI: CONVOLUTION PROPERTIES • Associativity x(t ) h1 (t ) h2 (t ) = x(t ) h1 (t ) h2 (t ) = x(t ) h1 (t ) h2 (t ) – proof h(t ) x(t ) h1 (t ) y1 (t ) h2 (t ) y(t ) Associativity ➔ x(t ) h1 (t ) h2 (t ) y(t ) 70 LTI: CONVOLUTION PROPERTIES • Distributivity x(t ) h1 (t ) + h2 (t ) = x(t ) h1 (t ) + x(t ) h1 (t ) – proof h1 (t ) x(t ) y(t ) + ➔ h2 (t ) Distributivity x(t ) h1 (t ) + h2 (t ) y(t ) 71 LTI: CONVOLUTION PROPERTIES • Example h1 (t ) h2 (t ) x(t ) + h3 (t ) h1 (t ) = exp( −2t )u (t ) h3 (t ) = exp( −3t )u(t ) h(t ) = ? y(t ) h4 (t ) h2 (t ) = 2 exp( −t )u (t ) h4 (t ) = 4 (t ) 72 LTI: GRAPHICAL CONVOLUTION • Graphical interpretation of convolution x(t) x(t) + y(t ) = x( )h(t − )d − t t x( ) h(-t) t x(t) h( ) t – 1. Reflection g ( ) = h(− ) t – 3. Multiplication x( )h(t0 − ) – 4. Integration y(t0 ) = x( )h(t0 − )d + − g ( − t0 ) = h(−( − t0 )) = h(t0 − ) 73 LTI: GRAPHICAL CONVOLUTION • Example y(t ) = [2a p2a (t )] [2a p2a (t − a)] 74 OUTLINE • Classifications of continuous-time system • Linear time-invariant system (LTI) • Properties of LTI system • System described by differential equations 75 LTI PROPERTIES • Memoryless LTI system – Review: present output only depends on present input y(t ) = Kx(t ) – The impulse response of Memoryless LTI system is h(t ) = K (t ) • Causal LTI system – Review: output depends on only current input and past input. – The impulse response of causal LTI system must satisfy: h(t ) = 0 – Why? for t 0 76 LTI PROPERTIES • Invertible LTI Systems – Review: a system is invertible iff (if and only if) there is an inverse system that, when connected in cascade with the original system, yields an output equal to original system input x(t ) x(t ) y(t ) h(t) g(t) x(t ) h(t ) g (t ) = x(t ) – For invertible LTI systems with IR (impulse response) h(t ) , there exists inverse system g (t ) such that g (t ) h(t ) = (t ) – Example: find the inverse system of LTI system h(t ) = (t − t0 ) 77 LTI PROPERTIES • BIBO Stable LTI state – Review: a system is BIBO stable iff every bounded input produces a bounded output. – LTI system: an LTI system is BIBO stable iff + − • Proof: h(t ) dt 78 LTI PROPERTIES • Examples – Determine: causal or non-causal, memory or memoryless, stable or unstable – 1. h1 (t ) = t exp( −2t )u (t ) + exp(3t )u (−t ) + (t − 1) – 2. h2 (t ) = −3 exp( 2t )u (t ) – 3. h3 (t ) = 5 (t + 5) 79 OUTLINE • Classifications of continuous-time system • Linear time-invariant system (LTI) • Properties of LTI system • System described by differential equations 80 DIFFERENTIAL EQUATIONS • LTI system can be represented by differential equations (N) (M ) a y ( t ) + a y ' ( t ) + + a y ( t ) = b x ( t ) + b x ' ( t ) + + b x (t ) – Initial conditions: 0 1 N 0 1 M d k y (t ) dt k t =0 – Notation: n-th derivative: d n y (t ) y (t ) = dt n (n) k = 0,, N − 1 81 DIFFERENTIAL EQUATION • Example: – Consider a circuit with a resistor R = 1 Ohm and an inductor L = 1H, with a voltage source v(t) = Bu(t), and is the initialIcurrent in the inductor. The output of the system is the current across theoinductor. • Represent the system with a differential equation. • Find the output of the system with and Io = 0 Io = 1 82 DIFFERENTIAL EQUATION a0 y(t ) + a1 y' (t ) + + aN y ( N ) (t ) = b0 x(t ) + b1 x' (t ) + + bM x ( M ) (t ) d k y (t ) dt k t =0 k = 0,, N − 1 • Zero-state response – The output of the system when the initial conditions are zero – Denoted as • Zero-input response yzs (t ) – The output of the system when the input is zero – Denoted as • The actual output of the system yzi (t ) y (t ) = y zs (t ) + y zi (t ) 83 DIFFERENTIAL EQUATION • Example – Find the zero-state output and zero-input response of the RL circuit in the previous example. Department of Electrical Engineering University of Arkansas ELEG 3124 SYSTEMS AND SIGNALS Ch. 3 Fourier Series Dr. Jingxian Wu wuj@uark.edu 85 OUTLINE • Introduction • Fourier series • Properties of Fourier series • Systems with periodic inputs 86 INTRODUCTION: MOTIVATION • Motivation of Fourier series – Convolution is derived by decomposing the signal into the sum of a series of delta functions • Each delta function has its unique delay in time domain. • Time domain decomposition + x(t ) = x( ) (t − )d = lim − →0 + x(n) (t − n) n = − x(t) t Illustration of integration 87 INTRODUCTION: MOTIVATION • Can we decompose the signal into the sum of other functions – Such that the calculation can be simplified? – Yes. We can decompose periodic signal as the sum of a sequence of complex exponential signals ➔ Fourier series. e j0t =e j 2f 0t f0 = 0 2 – Why complex exponential signal? (what makes complex exponential signal so special?) • 1. Each complex exponential signal has a unique frequency ➔ frequency decomposition • 2. Complex exponential signals are periodic 88 INTRODUCTION: REVIEW • Complex exponential signal e j 2ft = cos(2ft) + j sin( 2ft) – Complex exponential function has a one-to-one relationship with sinusoidal functions. – Each sinusoidal function has a unique frequency: f • What is frequency? – Frequency is a measure of how fast the signal can change within a unit time. • Higher frequency ➔ signal changes faster f = 0 Hz f = 1 Hz Department of Engineering Science Sonoma State University Sinusoidal at different frequencies f = 3 Hz 89 INTRODUCTION: ORTHONORMAL SIGNAL SET • Definition: orthogonal signal set , are 0 (t ),1 (t ), 2 (t ), said to be orthogonal over an – A set of signals, interval (a, b) if C , l = k a l (t ) (t )dt = 0, l k b • Example: * k k (t ) = e jk t k = 0,1,2, are – the signal set: orthogonal over the interval [0, T0 ] , where 2 0 = T0 0 90 OUTLINE • Introduction • Fourier series • Properties of Fourier series • Systems with periodic inputs 91 FOURIER SERIES • Definition: – For any periodic signal with fundamental period , it Tcan be decomposed as the sum of a set of complex exponential signals as 0 + c e x (t ) = n = − jn 0t n 0 = • cn , n = 0,1,2, , Fourier series coefficients cn = 1 T0 • derivation of cn : T0 x (t )e − jn0t dt 2 T0 92 FOURIER SERIES • Fourier series x (t ) = + c e n = − jn 0t n – The periodic signal is decomposed into the weighted summation of a set of orthogonal complex exponential functions. – The frequency of the n-th complex exponential function: n 0 • The periods of the n-th complex exponential function: Tn = T0 n – The values of coefficients, cn , n = 0,1,2, , depend on x(t) • Different x(t) will result in different c n • There is a one-to-one relationship between x(t) and cn s(t ) ➔ [, c−2 , c−1,c0 , c1 , c2 ,] For a periodic signal, it can be either represented as s(t), or represented as cn 93 FOURIER SERIES • Example x(t) − K , − 1 t 0 x(t ) = K, 0 t 1 t -3 -2 -1 1 Rectangle pulses 2 94 FOURIER SERIES • Amplitude and phase – The Fourier series coefficients are usually complex numbers j cn = an + jbn = cn e n – Amplitude line spectrum: amplitude as a function of cn = an2 + bn2 – Phase line spectrum: phase as a function of bn n = a tan an n 0 n 0 95 FOURIER SERIES: FREQUENCY DOMAIN • Signal represented in frequency domain: line spectrum – – – – phase amplitude Each c n has its own frequency n 0 The signal is decomposed in frequency domain. c n is called the harmonic of signal s(t) at frequency n 0 Each signal has many frequency components. • The power of the harmonics at different frequencies determines how fast the signal can change. 96 FOURIER SERIES: FREQUENCY DOMAIN • Example: Piano Note piano notes One piano note E5 E6 B6 E7 E5: 659.25 Hz E6: 1318.51 Hz B6: 1975.53 Hz E7: 2637.02 Hz spectrum All graphs in this page are created by using the open-source software Audacity. 97 FOURIER SERIES • Example – Find the Fourier series of s(t ) = exp( j 0t ) 98 FOURIER SERIES • Example – Find the Fourier series of s(t ) = B + Acos(0t + ) y(t ) = 1 + sin( 100t ) Time domain Amplitude spectrum Phase spectrum 99 FOURIER SERIES • Example – Find the Fourier series of − T / 2 t − / 2 0, s (t ) = K , 0, − / 2 t / 2 /2t T /2 = 1, T = 5 x(t) = 1, T = 10 t Time domain cn = K n sin c( ) T T = 1, T = 15 100 FOURIER SERIES: DIRICHLET CONDITIONS • Can any periodic signal be decomposed into Fourier series? – Only signals satisfy Dirichlet conditions have Fourier series • Dirichlet conditions – 1. x(t) is absolutely integrable within one period T | x(t ) | dt – 2. x(t) has only a finite number of maxima and minima. – 3. The number of discontinuities in x(t) must be finite. 101 OUTLINE • Introduction • Fourier series • Properties of Fourier series • Systems with periodic inputs 102 PROPERTIES: LINEARITY • Linearity – Two periodic signals with the same period x (t ) = + e n = − jn 0t y (t ) = n T0 = + 2 0 jn t e n 0 n = − – The Fourier series of the superposition of two signals is k1 x (t ) + k2 y (t ) = – If x(t ) = n + jn0t ( k + k ) e 1 n 2 n n = − y(t ) = n • then k1x(t ) + k2 y(t ) = (k1 n + k2 n ) 103 PROPERTIES: EFFECTS OF SYMMETRY • Symmetric signals – A signal is even symmetry if: x(t ) = x(−t ) – A signal is odd symmetry if: x(t ) = − x(−t ) – The existence of symmetries simplifies the computation of Fourier series coefficients. x(t) x(t) t -5 -4 -3 -2 -1 1 2 3 t -4 -3 -2 -1 1 2 Even symmetric 3 4 Odd symmetric 4 5 104 PROPERTIES: EFFECTS OF SYMMETRY • Fourier series of even symmetry signals – If a signal is even symmetry, then x (t ) = + a n = − n cos(n 0t ) 2 an = T0 T0 / 2 0 x (t ) cos(n0t )dt • Fourier series of odd symmetry signals – If a signal is odd symmetry, then + x (t ) = bn sin (n0t ) n =1 2 bn = T0 T0 / 2 0 x (t ) sin (n 0t )dt 105 PROPERTIES: EFFECTS OF SYMMETRY • Example x(t) 4A A − t, 0 t T / 2 T x(t ) = 4A t − 3 A, T / 2 t T T t Graph of x(t) 106 PROPERTIES: SHIFT IN TIME • Shift in time – If x(t ) has Fourier series ,cthen n x(t −has t0 ) Fourier series cne− jn0t0 if x(t ) ➔ cn , then x(t − t0 ) ➔ – Proof: cne− jn0t0 107 PROPERTIES: PARSEVAL’S THEOREM • Review: power of periodic signal 1 T P = | x(t ) |2 dt T 0 • Parseval’s theorem if x(t ) ➔ n then 1 T | x(t ) |2 dt = T 0 + 2 | | m m = − – Proof: The power of signal can be computed in frequency domain! 108 PROPERTIES: PARSEVAL’S THEOREM • Example – Use Parseval’s theorem find the power of x(t ) = Asin( 0t ) 109 OUTLINE • Introduction • Fourier series • Properties of Fourier series • Systems with periodic inputs 110 PERIODIC INPUTS: COMPLEX EXPONENTIAL INPUT • LTI system with complex exponential input x(t ) = e jt h(t ) y(t ) y(t ) = x(t ) h(t ) = h(t ) x(t ) + = h( ) x(t − )d − + = exp( jt ) h( ) exp( − j )d − • Transfer function + H () = h( ) exp( − j )d − – For LTI system with complex exponential input, the output is y(t ) = H () exp( jt ) – It tells us the system response at different frequencies 111 PERIODIC INPUT • Example: – For a system with impulse response find the transfer function h(t ) = (t − t0 ) 112 PERIODIC INPUT: • Example – Find the transfer function of the system shown in figure. RL circuit 113 PERIODIC INPUTS • Example – Find the transfer function of the system shown in figure RC circuit 114 PERIODIC INPUTS: TRANSFER FUNCTION • Transfer function – For system described by differential equations n py i =0 i m (i ) (t ) = qi x ( i ) (t ) i =0 m H () = q ( j) i i i =0 n p ( j) i =0 i i 115 PERIODIC INPUTS • LTI system with periodic inputs – Periodic inputs: x (t ) = + c n = − e linear: + jn0t c e n = − n h(t ) e jn0t H (n0 ) + jn0t n exp( jn0t ) 2 0 = T h(t ) jn0t c e n H ( n 0 ) n = − + x(t ) h(t ) jn0t c e n H ( n 0 ) n = − For system with periodic inputs, the output is the weighted sum of the transfer function. 116 PERIODIC INPUTS • Procedures: – To find the output of LTI system with periodic input • 1. Find the Fourier series coefficients of periodic input x(t). 1 n = T T 0 x (t )e − jn0t dt 0 = 2f 0 = • 2. Find the transfer function of LTI system H () • 3. The output of the system is y (t ) = + jn 0t c e n H (n0 ) n = − 2 T period of x(t) 117 PERIODIC INPUTS • Example – Find the response of the system when the input is x(t ) = 4 cos(t ) − 2 cos(2t ) RL Circuit 118 PERIODIC INPUTS • Example – Find the response of the system when the input is shown in figure. x(t) t -3 RC circuit -2 -1 1 Square pulses 2 119 PERIODIC INPUTS: GIBBS PHENOMENON • The Gibbs Phenomenon – Most Fourier series has infinite number of elements→ unlimited bandwidth x (t ) = + jn 0t c e n n = − • What if we truncate the infinite series to finite number of elements? x N (t ) = +N jn 0t c e n – The truncated signal,n = − N signal x(t) , is an approximation of the original xN (t ) 120 PERIODIC INPUTS: GIBBS PHENOMENON x(t) t -3 -2 -1 1 2 2K 1 , n odd, cn = j n 0, n even. x N (t ) = +N jn 0t c e n n=− N Square pulses x3 (t ) x5 (t ) x19 (t ) 121 FOURIER SERIES • Analogy: Optical Prism – Each color is an Electromagnetic wave with a different frequency Optical prism Department of Electrical Engineering University of Arkansas ELEG 3124 SYSTEMS AND SIGNALS Ch. 4 Fourier Transform Dr. Jingxian Wu wuj@uark.edu 123 OUTLINE • Introduction • Fourier Transform • Properties of Fourier Transform • Applications of Fourier Transform 124 INTRODUCTION: MOTIVATION • Motivation: – Fourier series: periodic signals can be decomposed as the summation of orthogonal complex exponential signals + cn exp jn0t 1 T cn = x (t ) exp jn0t dt n = − T 0 • each harmonic contains a unique frequency: n/T x (t ) = x(t) t Time domain • time domain ➔ frequency domain Frequency domain (T = ) How about aperiodic signals ? 125 INTRODUCTION: TRANSFER FUNCTION • System transfer function e e jt H () jt h(t ) H ( ) = h(t ) exp jt dt + − • System with periodic inputs e + jn0t c e n = − h(t ) + jn0t n e jn0t H (n0 ) h(t ) jn0t c e n H ( n 0 ) n = − + x(t ) h(t ) jn0t c e n H ( n 0 ) n = − 126 OUTLINE • Introduction • Fourier Transform • Properties of Fourier Transform • Applications of Fourier Transform 127 FOURIER TRANSFORM • Fourier Transform + X ( ) = x(t )e − jt dt − – given x(t), we can find its Fourier transform X ( ) • Inverse Fourier Transform x (t ) = 1 2 + − X ( )e jt d – given X ( ) , we can find the time domain signal x(t) – signal is decomposed into the “weighted summation” of complex exponential functions. (integration is the extreme case of summation) x(t ) ➔ X ( ) 128 FOURIER TRANSFORM • Example – Find the Fourier transform of x(t ) = rect (t / ) x(t) x(t) t t 129 FOURIER TRANSFORM • Example – Find the Fourier transform of x(t ) = exp( −a | t |) a0 130 FOURIER TRANSFORM • Example – Find the Fourier transform of x(t ) = exp( −at )u(t ) a0 131 FOURIER TRANSFORM • Example – Find the Fourier transform of x(t ) = (t − a) 132 FOURIER TRANSFORM: TABLE 133 FOURIER TRANSFORM • The existence of Fourier transform – Not all signals have Fourier transform – If a signal have Fourier transform, it must satisfy the following two conditions • 1. x(t) is absolutely integrable + | x(t ) | dt − • 2. x(t) is well behaved – The signal has finite number of discontinuities, minima, and maxima within any finite interval of time. • Example – x(t ) = exp(t )u(t ) 134 OUTLINE • Introduction • Fourier Transform • Properties of Fourier Transform • Applications of Fourier Transform 135 PROPERTIES: LINEARITY • Linearity – If x1 (t ) X 1 ( ) – then • Example x2 (t ) X 2 ( ) ax1 (t ) + bx2 (t ) aX 1 ( ) + bX 2 ( ) – Find the Fourier transform of x(t ) = 2rect (t / ) + 3 exp( −2t )u(t ) + 4 (t ) 136 PROPERTY: TIME-SHIFT • Time shift – If – Then x(t ) X ( ) x(t − t0 ) X () exp[ − jt0 ] phase shift • Review: complex number c =| c | e j =| c | cos( ) + j | c | sin( ) = a + jb a =| c | cos b =| c | sin | c |= a 2 + b 2 = a tan( b / a) – Phase shift of a complex number c by 0 : c exp( j0 ) =| c | exp j ( + 0 ) time shift in time domain ➔ frequency shift in frequency domain 137 PROPERTY: TIME SHIFT • Example: – Find the Fourier transform of x(t ) = rect t − 2 138 PROPERTY: TIME SCALING • Time scaling – If x(t ) X ( ) – Then 1 x(at ) X |a| a • Example – Let X ( ) = rect( − 1) / 2 , find the Fourier transform of x(−2t + 4) 139 PROPERTY: SYMMETRY • Symmetry – If x(t ) X ( ) , and x(t ) is a real-valued time signal – Then X (−) = X * () 140 PROPERTY: DIFFERENTIATION • Differentiation – If x(t ) X ( ) – Then dx(t ) jX ( ) dt • Example – Let d n x(t ) n ( ) j X ( ) n dt X ( ) = rect( − 1) / 2 , find the Fourier transform of dx(t ) dt 141 PROPERTY: DIFFERENTIATION • Example – Find the Fourier transform of (Hint: d 1 ) sgn( t ) = ( t ) dt 2 x(t ) = sgn( t ) 142 PROPERTY: CONVOLUTION • Convolution – If x(t ) X ( ), – Then h(t ) H ( ) x(t ) h(t ) X ( ) H ( ) x(t ) h(t ) x(t ) h(t ) time domain X ( ) H ( ) X ( ) H ( ) frequency domain 143 PROPERTY: CONVOLUTION • Example – An LTI system has impulse response h(t ) = exp (− at )u (t ) If the input is x(t ) = (a − b) exp (− bt )u (t ) + (c − a) exp( −ct )u (t ) Find the output (a 0, b 0, c 0) 144 PROPERTY: MULTIPLICATION • Multiplication – If x(t ) X ( ) , – Then x(t )m(t ) m(t ) M () 1 X ( ) M ( ) 2 145 PROPERTY: DUALITY • Duality – If – Then g (t ) G( ) G(t ) 2g (− ) 146 PROPERTY: DUALITY • Example – Find the Fourier transform of (recall: ) rect (t / ) sinc 2 t h(t ) = Sa 2 147 PROPERTY: DUALITY • Example – Find the Fourier transform of – Find the Fourier transform of x(t ) = 1 x(t ) = e j0t 148 PROPERTY: SUMMARY 149 PROPERTY: EXAMPLES • Examples – 1. Find the Fourier transform of x(t ) = cos(0t ) – 2. Find the Fourier transform of x(t ) = u(t ) 1 sgn( t ) + 1 2 2 j u (t ) = sgn( t ) 150 PROPERTY: EXAMPLES • Examples – 3. A LTI system with impulse response Find the output when input is – 4. If (Hint: h(t ) = exp− at u (t ) x(t ) = u(t ) x(t ) X ( ) , find the Fourier transform of t − x( )d = x(t ) u (t ) ) t − x( )d 151 PROPERTY: EXAMPLES • Example – 5. (Modulation) If , ) m(t ) = cos( t ) x(t ) X ( 0 Find the Fourier transform of x(t )m(t ) – 6. If X ( ) = 1 , find x(t) a + j 152 PROPERTY: DIFFERENTIATION IN FREQ. DOMAIN • Differentiation in frequency domain – If: – Then: x(t ) X ( ) d n X ( ) (− jt) x(t ) = d n n PROPERTY: DIFFERENTIATION IN FREQ. DOMAIN • Example – Find the Fourier transform of t exp( −at )u(t ), a0 153 154 PROPERTY: FREQUENCY SHIFT • Frequency shift – If: – Then: • Example – If x(t ) X ( ) x(t ) exp( j0t ) X ( − 0 ) X ( ) = rect( − 1) / 2 , find the Fourier transform x(t ) exp( − j 2t ) 155 PROPERTY: PARSAVAL’S THEOREM • Review: signal energy + E = | x(t ) |2 dt − • Parsaval’s theorem + 1 | x ( t ) | dt = − 2 2 + − | X ( ) |2 d 156 PROPERTY: PARSAVAL’S THEOREM • Example: – Find the energy of the signal x(t ) = exp( −2t )u(t ) 157 PROPERTY: PERIODIC SIGNAL • Fourier transform of periodic signal – Periodic signal can be written as Fourier series x(t ) = + c n = − n exp jn0t – Perform Fourier transform on both sides X ( ) = 2 + c ( − n ) n = − n 0 158 OUTLINE • Introduction • Fourier Transform • Properties of Fourier Transform • Applications of Fourier Transform 159 APPLICATIONS: FILTERING • Filtering – Filtering is the process by which the essential and useful part of a signal is separated from undesirable components. • Passing a signal through a filter (system). • At the output of the filter, some undesired part of the signal (e.g. noise) is removed. – Based on the convolution property, we can design filter that only allow signal within a certain frequency range to pass through. x(t ) h(t ) x(t ) X ( ) h(t ) H ( ) filter filter time domain X ( ) H ( ) frequency domain 160 APPLICATIONS: FILTERING • Classifications of filters Passband Stop band Low pass filter Stop Passband Stop band band Band pass filter Stop band Passband High pass filter Passband Stop Passband band Band stop (Notch) filter 161 APPLICATION: FILTERING • A filtering example – A demo of a notch filter X ( ) H ( ) Corrupted sound Filter X ( ) H ( ) Filtered sound 162 APPLICATIONS: FILTERING • Example – Find out the frequency response of the RC circuit – What kind of filters it is? RC circuit 163 APPLICATION: SAMPLING THEOREM • Sampling theorem: time domain – Sampling: convert the continuous-time signal to discrete-time signal. x(t ) p (t ) = + (t − nT ) n = − sampling period xs (t ) = x(t ) p(t ) Sampled signal 164 APPLICATION: SAMPLING THEOREM • Sampling theorem: frequency domain – Fourier transform of the impulse train • impulse train is periodic + 1 p(t ) = (t − nTs ) = Ts n = − Fourier series + jns t 1 e n = − 2 s = Ts • Find Fourier transform on both sides 2 P( ) = Ts + ( − n ) n = − s • Time domain multiplication ➔ Frequency domain convolution x (t ) p (t ) 1 X ( ) P( ) 2 1 x(t ) p(t ) Ts + X ( − n ) n = − s 165 APPLICATION: SAMPLING THEOREM • Sampling theorem: frequency domain – Sampling in time domain ➔ Repetition in frequency domain Time domain Frequency domain 166 APPLICATION: SAMPLING THEOREM • Sampling theorem – If the sampling rate is twice of the bandwidth, then the original signal can be perfectly reconstructed from the samples. s 2B s 2B s = 2B s 2B Frequency domain 167 APPLICATION: AMPLITUDE MODULATION • What is modulation? – The process by which some characteristic of a carrier wave is varied in accordance with an information-bearing signal Information bearing signal modulation Modulated signal Carrier wave • Three signals: – Information bearing signal (modulating signal) • Usually at low frequency (baseband) • E.g. speech signal: 20Hz – 20KHz – Carrier wave • Usually a high frequency sinusoidal (passband) • E.g. AM radio station (1050KHz) FM radio station (100.1MHz), 2.4GHz, etc. – Modulated signal: passband signal 168 APPLICATION: AMPLITUDE MODULATION • Amplitude Modulation (AM) s(t ) = Ac m(t ) cos(2f ct ) – A direct product between message signal and carrier signal m(t ) s(t ) Mixer Ac cos(2f ct ) Local Oscillator Amplitude modulation 169 APPLICATION: AMPLITUDE MODULATION • Amplitude Modulation (AM) S( f ) = Ac M ( f − f c ) + M ( f + f c ) 2 Amplitude modulation Department of Electrical Engineering University of Arkansas ELEG 3124 SYSTEMS AND SIGNALS Ch. 5 Laplace Transform Dr. Jingxian Wu wuj@uark.edu 171 OUTLINE • Introduction • Laplace Transform • Properties of Laplace Transform • Inverse Laplace Transform • Applications of Laplace Transform 172 INTRODUCTION • Why Laplace transform? – Frequency domain analysis with Fourier transform is extremely useful for the studies of signals and LTI system. • Convolution in time domain ➔ Multiplication in frequency domain. – Problem: many signals do not have Fourier transform x(t ) = exp( at )u(t ), a 0 x(t ) = tu(t ) – Laplace transform can solve this problem • It exists for most common signals. • Follow similar property to Fourier transform • It doesn’t have any physical meaning; just a mathematical tool to facilitate analysis. – Fourier transform gives us the frequency domain representation of signal. 173 OUTLINE • Introduction • Laplace Transform • Properties of Laplace Transform • Inverse Lapalace Transform • Applications of Fourier Transform LAPLACE TRANSFORM: BILATERAL LAPLACE TRANSFORM • Bilateral Laplace transform (two-sided Laplace transform) + X B ( s) = x(t ) exp( − st )dt , − s = + j – s = + j is a complex variable – s is often called the complex frequency – Notations: X B ( s ) = L[ x(t )] x (t ) X B ( s ) • Time domain v.s. S-domain – x(t ) : a function of time t → x(t) is called the time domain signal – X B (s ) : a function of s → X B (s ) is called the s-domain signal – S-domain is also called as the complex frequency domain 174 175 LAPLACE TRANSFORM • Time domain v.s. s-domain – x(t ) : a function of time t → x(t) is called the time domain signal – X B (s ) : a function of s → X B (s ) is called the s-domain signal • S-domain is also called the complex frequency domain – By converting the time domain signal into the s-domain, we can usually greatly simplify the analysis of the LTI system. – S-domain system analysis: • 1. Convert the time domain signals to the s-domain with the Laplace transform • 2. Perform system analysis in the s-domain • 3. Convert the s-domain results back to the time-domain LAPLACE TRANSFORM: BILATERAL LAPLACE TRANSFORM • Example – Find the Bilateral Laplace transform of x(t ) = exp( −at )u(t ) • Region of Convergence (ROC) – The range of s that the Laplace transform of a signal converges. – The Laplace transform always contains two components • The mathematical expression of Laplace transform • ROC. 176 177 LAPLACE TRANSFORM: BILATERAL LAPLACE TRANSFORM • Example – Find the Laplace transform of x(t ) = − exp( −at )u(−t ) LAPLACE TRANSFORM: BILATERAL LAPLACE TRANSFORM • Example – Find the Laplace transform of x(t ) = 3 exp( −2t )u(t ) + 4 exp(t )u(−t ) 178 179 LAPLACE TRANSFORM: UNILATERAL LAPLACE TRANSFORM • Unilateral Laplace transform (one-sided Laplace transform) + X ( s) = − x(t ) exp( − st )dt 0 – 0− :The value of x(t) at t = 0 is considered. – Useful when we dealing with causal signals or causal systems. • Causal signal: x(t) = 0, t < 0. • Causal system: h(t) = 0, t < 0. – We are going to simply call unilateral Laplace transform as Laplace transform. 180 LAPLACE TRANSFORM: UNILATERAL LAPLACE TRANSFORM • Example: find the unilateral Laplace transform of the following signals. – 1. x(t ) = A – 2. x(t ) = (t ) 181 LAPLACE TRANSFORM: UNILATERAL LAPLACE TRANSFORM • Example – 3. x(t ) = exp( j 2t ) – 4. x(t ) = cos(2t ) – 5. x(t ) = sin( 2t ) LAPLACE TRANSFORM: UNILATERAL LAPLACE TRANSFORM 182 183 OUTLINE • Introduction • Laplace Transform • Properties of Laplace Transform • Inverse Lapalace Transform • Applications of Fourier Transform 184 PROPERTIES: LINEARITY • Linearity x (t ) X ( s) x1 (t ) X 1 ( s) 2 2 – If ax1 (t ) + bx2 (t ) aX 1 ( s) + bX 2 ( s) – Then The ROC is the intersection between the two original signals • Example – Find the Laplace transfrom of A + B exp( −bt )u(t ) 185 PROPERTIES: TIME SHIFTING • Time shifting t0 0 x(t ) X (s)and – If x(t − t0 )u(t − t0 ) X (s) exp( −st0 ) – Then The ROC remain unchanged 186 PROPERTIES: SHIFTING IN THE s DOMAIN • Shifting in the s domain – If – Then Re(s) x(t ) X (s) y(t ) = x(t ) exp( s0t ) X (s − s0 ) Re( s) + Re( s0 ) • Example – Find the Laplace transform of x(t ) = A exp( −at ) cos(0t )u(t ) 187 PROPERTIES: TIME SCALING • Time scaling – If – Then x(t ) X (s) 1 s x(at ) X a a • Example – Find the Laplace transform of Re{s} 1 Re{s} a 1 x(t ) = u(at ) 188 PROPERTIES: DIFFERENTIATION IN TIME DOMAIN • Differentiation in time domain – If – Then g (t ) G(s) dg (t ) sG ( s ) − g (0 − ) dt d 2 g (t ) 2 − − s G ( s ) − sg ( 0 ) − g ' ( 0 ) 2 dt d n g (t ) n n −1 − ( n−2) − ( n −1) − s G ( s ) − s g ( 0 ) − − sg ( 0 ) − g ( 0 ) n dt • Example 2 – Find the Laplace transform of g (t ) = sin t u(t ), g (0 − ) = 0 189 PROPERTIES: DIFFERENTIATION IN TIME DOMAIN • Example – Use Laplace transform to solve the differential equation y' ' (t ) + 3 y' (t ) + 2 y(t ) = 0, y(0− ) = 3 y ' (0 − ) = 1 190 PROPERTIES: DIFFERENTIATION IN S DOMAIN • Differentiation in s domain – If – Then x(t ) X (s) d n X (s) (−t ) x(t ) ds n n • Example – Find the Laplace transform of t n u (t ) 191 PROPERTIES: CONVOLUTION • Convolution x(t ) X (s) h(t ) H ( s) – If – Then x(t ) h(t ) X (s) H (s) The ROC of X ( s) H ( s) is the intersection of the ROCs of X(s) and H(s) 192 PROPERTIES: INTEGRATION IN TIME DOMAIN • Integration in time domain – If – Then x(t ) X (s) t 1 0 x( )d s X (s) • Example – Find the Laplace transform of r(t ) = tu(t ) 193 PROPERTIES: CONVOLUTION • Example – Find the convolution t −a t −a rect rect 2a 2a 194 PROPERTIES: CONVOLUTION • Example – For a LTI system, the input is system is x(t ) = exp( −2t ), uand (t )the output of the y (t ) = exp( −t ) + exp( −2t ) − exp( −3t )u(t ) Find the impulse response of the system 195 PROPERTIES: CONVOLUTION • Example – Find the Laplace transform of the impulse response of the LTI system described by the following differential equation 2 y' ' (t ) − 3 y' (t ) + y(t ) = 3x' (t ) + x(t ) (n) (n) assume the system was initially relaxed ( y (0) = x (0) = 0 ) 196 PROPERTIES: MODULATION • Modulation – If – Then x(t ) X (s) 1 x(t ) cos(0t ) X ( s + j0 ) + X ( s − j0 ) 2 j x(t ) sin( 0t ) X ( s + j0 ) − X ( s − j0 ) 2 197 PROPERTIES: MODULATION • Example – Find the Laplace transform of x(t ) = exp( −at ) sin( 0t )u(t ) 198 PROPERTIES: INITIAL VALUE THEOREM • Initial value theorem – If the signal x(t ) is infinitely differentiable on an interval around then x(0 + ) = lim sX ( s) s → x(0+ ) s = must be in ROC – The behavior of x(t) for small t is determined by the behavior of X(s) for large s. 199 PROPERTIES: INITIAL VALUE THEOREM • Example – The Laplace transform of x(t) is Find the value of + x(0 ) cs + d X ( s) = ( s − a)( s − b) 200 PROPERTIES: FINAL VALUE THEOREM • Final value theorem – If – Then: x(t ) X (s) lim x(t ) lim sX (s) t → s →0 s = 0 must be in ROC • Example – The input x(t ) = Au(t ) is applied to a system with transfer c function , find the value of lim y(t ) H ( s) = s ( s + b) + c t → 201 PROPERTIES 202 OUTLINE • Introduction • Laplace Transform • Properties of Laplace Transform • Inverse Lapalace Transform • Applications of Fourier Transform 203 INVERSE LAPLACE TRANSFORM • Inverse Laplace transform 1 + j x(t ) = X ( s) exp( st )ds − j 2j – Evaluation of the above integral requires the use of contour integration in the complex plan ➔ difficult. • Inverse Laplace transform: special case – In many cases, the Laplace transform can be expressed as a rational function of s bm s m + bm−1s m−1 + + b1s + b0 X (s) = an s n + an −1s n −1 + + a1s + a0 – Procedure of Inverse Laplace Transform • 1. Partial fraction expansion of X(s) • 2. Find the inverse Laplace transform through Laplace transform table. 204 INVERSE LAPLACE TRANSFORM • Review: Partial Fraction Expansion with non-repeated linear factors X (s) = A = (s − a1 ) X (s) s =a 1 A B C + + s − a1 s − a2 s − a3 B = (s − a2 ) X (s) s =a 2 C = (s − a3 ) X (s) s =a • Example – Find the inverse Laplace transform of X (s) = 3 2s + 1 s 3 + 3s 2 − 4 s 205 INVERSE LAPLACE TRANSFORM • Example – Find the Inverse Laplace transform of 2s 2 X ( s) = 2 s + 3s + 2 • If the numerator polynomial has order higher than or equal to the order of denominator polynomial, we need to rearrange it such that the denominator polynomial has a higher order. 206 INVERSE LAPLACE TRANSFORM • Partial Fraction Expansion with repeated linear factors X (s) = 1 A2 A1 B = + + ( s − a ) 2 ( s − b) (s − a )2 s − a s − b A2 = (s − a ) X (s) 2 s =a A1 = d (s − a )2 X (s) ds B = (s − b )X ( s) s =b s =a 207 INVERSE LAPLACE TRANSFORM • High-order repeated linear factors AN A1 A2 B X ( s) = = + ++ + N 2 N ( s − a ) ( s − b) s − a ( s − a ) ( s − a) s −b 1 1 d N −k N ( ) Ak = s − a X ( s) N −k ( N − k )! ds B = (s − b )X ( s) s =b k = 1,, N s =a 208 OUTLINE • Introduction • Laplace Transform • Properties of Laplace Transform • Inverse Lapalace Transform • Applications of Laplace Transform 209 APPLICATION: LTI SYSTEM REPRESENTATION • LTI system – System equation: a differential equation describes the input output relationship of the system. y ( N ) (t ) + aN −1 y ( N −1) (t ) + + a1 y (1) (t ) + a0 y(t ) = bM x( M ) (t ) + + b1 x (1) (t ) + b0 x(t ) N −1 y (N) M (t ) + an y (t ) = bm x ( m ) (t ) (n) n =0 m =0 – S-domain representation N N −1 n M m s + an s Y ( s) = bm s X ( s) n =0 m =0 – Transfer function M Y (s) H (s) = = X (s) b m =0 m sm N −1 s + an s n N n =0 210 APPLICATION: LTI SYSTEM REPRESENTATION • Simulation diagram (first canonical form) Simulation diagram 211 APPLICATION: LTI SYSTEM REPRESENTATION • Example – Show the first canonical realization of the system with transfer function s 2 − 3s + 2 H (S ) = 3 s + 6s 2 + 11s + 6 212 APPLICATION: COMBINATIONS OF SYSTEMS • Combination of systems – Cascade of systems H ( S ) = H1 ( s ) H 2 ( s ) – Parallel systems H ( S ) = H1 ( s ) + H 2 ( s ) 213 APPLICATION: LTI SYSTEM REPRESENTATION • Example – Represent the system to the cascade of subsystems. s 2 − 3s + 2 H (S ) = 3 s + 6s 2 + 11s + 6 214 APPLICATION: LTI SYSTEM REPRESENTATION • Example: – Find the transfer function of the system LTI system 215 APPLICATION: LTI SYSTEM REPRESENTATION • Poles and zeros H (s) = – Zeros: – Poles: ( s − z M )( s − z M −1 ) ( s − z1 ) ( s − p N )( s − p N −1 ) ( s − p1 ) z1 , z 2 ,, z M p1 , p2 ,, pN 216 APPLICATION: STABILITY • Review: BIBO Stable – Bounded input always leads to bounded output + − | h(t ) | dt • The positions of poles of H(s) in the s-domain determine if a system is BIBO stable. H (s) = AN A1 A2 + + + s − s1 ( s − s2 ) m s − sN – Simple poles: the order of the pole is 1, e.g. s1 sN – Multiple-order poles: the poles with higher order. E.g. s 2 217 APPLICATION: STABILITY • Case 1: simple poles in the left half plane 1 1 = (s − k )2 + k2 ( s − k + jk )( s − k − jk ) p1 = k − j k hk (t ) = + − 1 k k 0 p2 = k + jk exp( k t ) sin( k t )u (t ) hk (t ) dt = Impulse response • If all the poles of the system are on the left half plane, then the system is stable. 218 APPLICATION: STABILITY • Case 2: Simple poles on the right half plane 1 1 = (s − k )2 + k2 ( s − k + jk )( s − k − jk ) p1 = k + jk hk (t ) = 1 k k 0 p2 = k − jk exp( k t ) sin( k t )u (t ) Impulse response • If at least one pole of the system is on the right half plane, then the system is unstable. 219 APPLICATION: STABILITY • Case 3: Simple poles on the imaginary axis 1 1 = (s − k )2 + k2 ( s − k + jk )( s − k − jk ) hk (t ) = 1 k k = 0 sin( k t )u (t ) • If the pole of the system is on the imaginary axis, it’s unstable. 220 APPLICATION: STABILITY • Case 4: multiple-order poles in the left half plane 1 m k 0 stable hk (t ) = t exp( k t ) sin( k t )u (t ) k • Case 5: multiple-order poles in the right half plane 1 m hk (t ) = t exp( k t ) sin( k t )u (t ) unstable k 0 k • Case 6: multiple-order poles on the imaginary axis hk (t ) = 1 k t m sin( k t )u (t ) k 0 unstable k 0 221 APPLICATION: STABILITY • Example: – Check the stability of the following system. H ( s) = 3s + 2 s 2 + 6 s + 13 Department of Electrical Engineering University of Arkansas ELEG 3124 Signals & Systems Ch. 6 Discrete-Time System Dr. Jingxian Wu wuj@uark.edu 223 OUTLINE • Discrete-time signals • Discrete-time systems • Z-transform 224 SIGNAL • Discrete-time signal – The time takes discrete values n x(n) = cos 4 1 n x ( n) = exp 2 4 225 SIGNAL: CLASSIFICATION • Energy signal v.s. Power signal – Energy: E = lim N → N x ( n) 2 n=− N – Power: N 1 2 P = lim x ( n) N → 2 N + 1 n=− N – Energy signal: E – Power signal: P 226 SIGNAL: CLASSIFICATION • Periodic signal v.s. aperiodic signal – Periodic signal x(n) = x(n + N ) • The smallest value of N that satisfies this relation is the fundamental periods. – Is periodic? cos(n) cos(n) is periodic if – Example: cos(3n) cos(n) 3 cos( n) 4 2 k is integer for integer k. 227 SIGNAL: ELEMENTARY SIGNAL • Unit impulse function 1, n = 0, ( n) = 0, n 0. • Unit step function 0, n 0, u ( n) = 1, n 0. • Relation between unit impulse function and unit step function (n) = u(n) − u(n − 1) u ( n) = n (k ) k = − 228 SIGNAL: ELEMENTARY SIGNAL • Exponential function x(n) = exp(n) • Complex exponential function x(n) = exp( j0n) = cos(0n) + j sin( 0n) 229 OUTLINE • Discrete-time signals • Discrete-time systems • Z-transform 230 SYSTEM: IMPULSE RESPONSE • Impulse response of LTI system (n) – The response of the system when the input is x(n) = (n) y(n) = h(n) System LTI system • System response for arbitrary input – Any signal can be decomposed as the sum of time-shifted impulses x ( n) = – Time invariant (n − k ) + x(k ) (n − k ) k = − h( n − k ) System LTI system – Linear + + x(k ) (n − k ) k = − x(k )h(n − k ) System k = − LTI system 231 SYSTEM: CONVOLUTION SUM • Convolution sum – The convolution sum of two signals x(n) and x ( n) h( n) = h(n) is + x(k )h(n − k ) k = − • Response of LTI system – The output of a LTI system is the convolution sum of the input and the impulse response of the system. x(n) h(n) x(n) h(n) LTI system 232 SYSTEM: CONVOLUTION SUM • Example – 1. x(n) (n − m) – 2. x(n) = nu(n), x(n) h(n) = h(n) = nu(n) 233 SYSTEM: CONVOLUTION SUM • Example: – Let x(n) = [1,3,−1,−2] sequences, find two h(n) = [1,2,0,−1be,1], x(n) h(n) 234 STSTEM: COMBINATION OF SYSTEMS • Combination of systems ➔ Two systems in series + ➔ Two systems in parallel 235 SYSTEM: DIFFERENCE EQUATION REPRESENTATION • Difference equation representation of system N a k =0 M k y (n − k ) = bk x(n − k ) k =0 236 OUTLINE • Discrete-time signals • Discrete-time systems • Z-transform 237 Z-TRANSFORM • Bilateral Z-transform X ( z) = + −n x ( n ) z n = − • Unilateral Z-transform + X ( z ) = x(n)z − n n =0 • Z-transform: – Ease of analysis – Doesn’t have any physical meaning (the frequency domain representation of discrete-time signal can be obtained through discrete-time Fourier transform) – Counterpart for continuous-time systems: Laplace transform. 238 Z-TRANSFORM • Example: find Z-transforms – 1. x(n) = (n) n – 2. x(n) = 1 u (n) 2 239 Z-TRANSFORM • Example – 3. n 1 x(n) = − u (− n − 1) 2 • Region of convergence (ROC) Region of convergence 240 Z-TRANSFORM: CONVERGENCE • Convergence of causal signal x(n) = nu(n) • Convergence of anti-causal signal x(n) = nu(−n −1) 241 Z-TRANSFORM: TIME SHIFTING PROPERTY • Time Shifting – Let x (n ) be a causal sequence with the Z-transform – Then Z x(n + n0 ) = z X ( z ) − z n0 Z x(n − n0 ) = z − n0 X ( z) + z n0 − n0 n0 −1 x ( m) z −m m =0 −1 x(m) z m = − n0 −m X (z ) 242 Z-TRANSFORM: LTI SYSTEM • LTI System – Difference equation representation N a k =0 M k y (n − k ) = bk x(n − k ) k =0 – Z-domain representation N M −k −k a z Y ( z ) = b z k k X ( z ) k =0 k =0 – Transfer function M −k b z k Y ( z ) k =0 H ( z) = = N X ( z) −k a z k k =0 243 Z-TRANSFORM: LTI SYSTEM • Example – Find the transfer function of the system described by the following difference equation 1 y (n) − 2 y (n − 1) + 2 y (n − 2) = x(n) + x(n − 1) 2 244 Z-TRANSFORM: STABILITY • Stability z H ( z) = z−a h( n) = a n u ( n) – A LTI system is BIBO stable is all the poles are within the unit circle (|a| < 1) – A LTI system is unstable is at least one pole is on or outside of the unit circult ( | a | 1 )