The zero-input response is found by solving the homogeneous part of the system’s differential equation using the given initial conditions. With y(0-) = 2 and dy(0-)/dt = 1, substitute these into the homogeneous solution form Yh(t) = C1e^(r1t) + C2e^(r2t), derived from characteristic equation roots. These conditions determine constants C1 and C2. The solution represents the natural response of the system, independent of any external input, capturing how energy initially supplied to the system dissipates over time .

The block diagram for an LTI system described by H(s) can be constructed using basic elements like integrators, summers, and gain blocks. For H(s) = (s^2 - s - 6)/(s^2 + 2s + 1), express the numerator and denominator as a cascade of simpler transfer functions. Represent this using blocks where gains and feedback paths correspond to system poles and zeros. This visualizes system dynamics through feedback loops and gain stages, where forward paths relate to numerator coefficients and feedback paths relate to denominator terms .

The output response is computed by adding both the zero-input and zero-state responses. Use Laplace transforms to translate the differential equation and initial conditions into the s-domain. For input x(t) = e^(-t)u(t), calculate the zero-state response. Separately, using initial conditions y(0-) = 2 and dy(0-)/dt = 1, solve for constants in the general solution, applying the final value and initial value theorems. Combined, these responses form the complete output using inverse transform methods .

Partial-fraction expansion decomposes a complex rational function into simpler fractions easier to inverse-transform. For X(s) = (5s+13) / (s(s^2 + 4s + 13)), the partial-fraction decomposes into A/s + (Bs+C)/(s^2 + 4s + 13). Solving gives A, B, and C as constants. Each term can be individually inverse-transformed using standard tables, reconstructing the complete time-domain signal as the sum of simpler inverted terms .

To determine the zero-state response, apply the Laplace transform to both sides of the differential equation. The Laplace transform of the input x(t) = e^(-t)u(t) is 1/(s+1). The equivalent Laplace domain equation is Y(s)(s² + 4s + 4) = (s + 1)X(s). Solve for Y(s) = 1/(s+1)(s² + 4s + 4). The poles determine system behavior and partial fraction expansion allows finding the inverse Laplace transform. This results in a response involving exponential decays and oscillatory terms based on the pole locations .

The roots of the numerator, or zeros of the system function, impact specific frequency components' attenuation in the output response. For H(s) = (s^2 - s - 6), factorize it to find zeros at s = 3 and s = -2. Zeros cause specific frequencies to be completely attenuated from the output, thereby shaping the overall frequency characteristics without affecting stability directly, which is dominated by the poles .

The inverse unilateral Laplace transform is determined using partial fraction expansion. First, factor the denominator as (s+2)(s+3). Express X(s) as A/(s+2) + B/(s+3). Solving for constants A and B gives A = -1 and B = 3. Thus, X(s) translates to -1/(s+2) + 3/(s+3), corresponding to the time functions -e^(-2t) + 3e^(-3t)u(t) using standard table transforms .

The initial conditions given are y(0-) = 2 and dy(0-)/dt = 1. These initial conditions define the particular solution for the homogeneous part of the differential equation. Given these, transform the initial conditions into the Laplace domain using final and initial value theorems, facilitating the calculation of specific constants in the inverse Laplace transform process. These conditions modify the response by adding additional terms to account for initial energy stored in the system, affecting how the system evolves over time .

The stability of a causal LTI system is determined by the location of the poles of its system function in the s-plane. The denominator s^2 + 2s + 1 can be expressed as (s+1)^2, indicating a repeated pole at s = -1. For stability, all poles must have negative real parts, and since this pole is negative, the system is stable for all bounded inputs .

To determine the Laplace transform of x(t) = u(t) - u(t-1), we first recognize that it represents a rectangular pulse of unit amplitude from t = 0 to t = 1. The Laplace transform of u(t) is 1/s and for u(t-a) is e^(-as)/s. Thus, the transform for x(t) is L{u(t)} - L{u(t-1)} = 1/s - e^(-s)/s. For the region of convergence, since both terms converge for Re(s) > 0, the final ROC is Re(s) > 0 .