Consider the continuous-time system shown in the figure above, which is characterized by the following input-output relationships :
[Figure, described in words: x(t) enters System 1, whose output w(t) enters System 2, whose output is y(t); the two blocks in cascade form the Overall system.]
System 1 : w(t) = ∫ from t−3 to t of x(T) dT
System 2 : y(t) = ∫ from t−6 to t of w(T) dT
(i) Determine whether the overall system is linear and time-invariant.
(ii) Determine and plot impulse response of the overall system.
10 marks · directive: Determine
The core of a good answer
Both blocks are running (moving window) integrators, so each is linear and time-invariant and so is the cascade; h(t) = [u(t) − u(t−3)] * [u(t) − u(t−6)], a trapezoid that rises linearly from 0 at t = 0 to 3 at t = 3, stays at 3 up to t = 6, and falls linearly to 0 at t = 9.
An outline written by upsc.help, not an official UPSC answer; every fact checked by a separate review. How we work
What the question asks
Part (i) tests the formal tests of linearity (superposition) and time invariance (shift of input gives the same shift of output). Part (ii) asks for the impulse response of a cascade, which is the convolution of the two individual impulse responses, and a labelled sketch.
Answer outline
Introduction
Each system integrates its input over a fixed window ending at the present instant: 3 s for System 1 and 6 s for System 2. Write each in convolution form and then convolve.
Step 1: Linearity
- For x = a x₁ + b x₂: w(t) = ∫ from t−3 to t of [a x₁(T) + b x₂(T)] dT = a w₁(t) + b w₂(t), so System 1 is linear.
- The same argument holds for System 2, and a cascade of linear systems is linear.
Step 2: Time invariance
- Input x(t − t₀) gives ∫ from t−3 to t of x(T − t₀) dT; put σ = T − t₀: ∫ from t−t₀−3 to t−t₀ of x(σ) dσ = w(t − t₀).
- So System 1 is time-invariant; System 2 likewise, hence the overall system is LTI (it is also causal and BIBO stable, since each h is absolutely integrable).
Step 3: Individual impulse responses
- Put x(t) = δ(t): h₁(t) = ∫ from t−3 to t of δ(T) dT = 1 for 0 ≤ t < 3, else 0, so h₁(t) = u(t) − u(t − 3).
- Similarly h₂(t) = u(t) − u(t − 6).
Step 4: Convolution h = h₁ * h₂
- h(t) = ∫ h₁(τ) h₂(t − τ) dτ = length of overlap of [0, 3] and [t − 6, t]
- t < 0: h = 0
- 0 ≤ t ≤ 3: h = t
- 3 ≤ t ≤ 6: h = 3
- 6 ≤ t ≤ 9: h = 9 − t
- t > 9: h = 0
- Compact form: h(t) = r(t) − r(t − 3) − r(t − 6) + r(t − 9), with r(t) = t u(t) the unit ramp.
Step 5: Plot (describe on the answer sheet)
- A trapezoid: straight line from (0, 0) to (3, 3), flat top from (3, 3) to (6, 3), straight line down from (6, 3) to (9, 0).
- Area under h = 3 × 6 = 18, which equals (area of h₁) × (area of h₂) = 3 × 6, a quick check.
Conclusion
The overall system is LTI with h(t) = r(t) − r(t−3) − r(t−6) + r(t−9), a trapezoid of height 3 between t = 0 and t = 9; the area check 18 = 3 × 6 confirms it.
Examples and data to cite
- Moving average filters in instrumentation and DSP are scaled versions of System 1 (window integrator divided by window length).
- Convolving two rectangles of unequal width always gives a trapezoid; equal widths give a triangle.
Facts to remember
- Impulse response of a cascade of LTI systems: h = h₁ * h₂ (order does not matter).
- A system is linear if it obeys superposition and time-invariant if a shifted input gives an identically shifted output.
- ∫ h(t) dt of a convolution equals the product of the individual areas.
- u(t) − u(t − a) is the rectangular pulse of width a starting at t = 0.