Conmaiiy Ab Transpose sn for linear equations fl solution A.13.5 Properties of the ge 5g Gaussian elimination A211 Definite quadratic forms References 228 206 220 29 230 232 232 230 239 230 240 247 251 253 255 259 261 264 265 270 210 216 279 I Basic concepts, terms, and definitions 1d Since this book is concemed with the Principles of Linear Systems, natural to first give a definition of a system. roduction, nents that performa specific function. These componenus ma mechanical, hydraulic, thermal and so fort, ‘every system is a variety of physical quantities such as electric wrces and displacements, low rates, and wid npuis or excitations; the latier ae called oufputs or responses Clearly, the set of inputs and the set of outputs of a system are not | uniquely defined, and some of them may have to be chosen. at been made, a defini forms them into ou «4 —— | __.. , Inputs < Pee et Sytem oe eee excitations 5 s Tesponses [—> Yncontinuous-time and discrete-time signals exist or. frequently, one in the analog signals cun be uniquely specified via discrete signals. 1.4 The techniques of linear system analysis be concemed with physical, dynamic, ing, lincer, analog, and discrete systems. Such systems are thescequations in orderto obtain the input. ing systems. There are two approaches to this problem: direct and indirect, ‘The direct approach is to obtain the solution as the sum of the homogeneous solution and the particular integral. This approach may yield quick answers for certain forms of inputs, and it has advantages whenever one is interested in the jput for whieh the particular integral can be determined ne function, but its relation to the tance in the theory of linear systems, The indirect ving steps: functions, jon of the response to a typical elementary function. 2 Composition or recombination of a Tinear combination of elementary re iin the response to the arbitrary input 2 two important applications of the indirect approach, we will make two assertions. The proof of these assertions is not presented here, First, we can easily prove by induction that for a linear system sf Y 1} = Sas (0h for finite n Before out Assertion 1 Itis tue that for a bounded or continuous linear system {Soa oi} ~ Sas Assertion 2 Its tue that fora bounded or continuous linear syste ar ) provided the Riemann negra f° @ ofasim, lu(t, $)d exists uniformly in 2s the limit Application 1: Time-demain analysis / If the elementary functions are unit impulses, then «(¢) can be expressed by the well-known formula u= [ u(x)6(t— vee ‘Then, using linearity and Assertion 2, we obtain vin =swinn=s{ fuera mar} = J" wrstee-nier Dedining the impulse response by h(t, 1) 2 ${8¢¢ — x)} we obtain vo= [ma nucnar i . : which is the input-output relation for a system pulse response h( & # Application 2: Frequency-domain analysis: a Here the input is expressed in terms of exponential time functions. Such an expre sion is given by the well-known inverse Laplace transform formula ye [io veetas. © Bij dene hls POE was raed uy ‘Then, using linearity and Assertion 2, we obiain vO= ore s{e u(se" as} = -al. Uls)Ste"}as. ‘We will prove later that the response of a linear and fixed analog system to an exponential input eis of the form Sfet"] = H(s)e", where H(s) is the systemand definitions Hee System ewer: s Fig, 12, ye 1.1 the inputs 1y,.11g,... 1, andthe outputs, Y2,.... Ye are in general mnctions, and if we use the vector notati ult) yt) Ya. wwe can represent the system as in Figure 1.2. 1.2 Input-output relations ‘tis an essential part of the system concept that clianges of the input eause changes che output. An important problem in the study of systems is to determine what sponse of a due to a given inpat. One way to solve this by applying the given inpu system and record fowever, experimentation of this kind is usually not a satisfactory approach to the problem. An alternate approach is to ‘express the signals as time functions and use the mathematical relations describing the components of the system combined in accordance with their interconnections. ‘Thus a mathematical relation between the input and the cutput can be obtained, which may be written as wt) = Siut)) ‘This is called the inpur-ourput relation for the given system. Equation (1.1) rep- notation (+) * the value of a vector function at the spe lio. 4] © ete) fe = t < 1 (i.e. the entire time function defined over 4 "' en eee eee er eee ~ interval of definition ime). This is also written ) may be written as 4 © the entire time function wi Ge.,u(¢) forall: on an unspecified as {u(1)}. Now. using the above notation, y= Stab (1.2) or simply ya Su. 3) “The three expressions needs ll be used interchangeably and according to notational 1.3 Classification of systems 43.1 Physical realizability Definition 1.2. system S with an input-output relation y = Su} iscalled Definition 1.3 A system is c real. ‘An example of a nonreal system is he Sone POF Cento. onde cas a z = g does not depe' g 2 z 3 5 2 ime # may depend only Bf input values oceurting for ’ < £, Obviously a noncausal or nonreal system is ie realizable, q Examples of an anticipative or noneausal system: ) because the value of y at any time 1 depen. tHlzt because for times + < O ort > 1, y(1) depends on A system described by