Why discrete-time matters for digital control in Simulink - Performance Analysis

Time Complexity: Why discrete-time matters for digital control

O(N^2)

Understanding Time Complexity

When working with digital control systems, we use discrete-time signals instead of continuous ones.

We want to understand how the time it takes to run a digital control model changes as we increase the number of time steps.

Scenario Under Consideration

Analyze the time complexity of the following Simulink model snippet.

% Discrete-time digital controller update loop
for k = 2 : N
    u(k) = Kp * e(k) + Ki * sum(e(1:k)) + Kd * (e(k) - e(k-1));
    y(k) = plantModel(u(k));
end
    

This code updates the control signal and system output at each discrete time step.

Identify Repeating Operations

Identify the loops, recursion, array traversals that repeat.

• Primary operation: The for-loop runs from 2 to N, updating control and output.
• How many times: It runs exactly N-1 times, once per time step.
• Inside the loop: The sum of errors from 1 to k is calculated each time, which grows with k.

How Execution Grows With Input

Each time step, the sum of all previous errors is recalculated, so work grows as more steps happen.

Input Size (N)	Approx. Operations
10	~55 (sum of 1 to 10)
100	~5050 (sum of 1 to 100)
1000	~500500 (sum of 1 to 1000)

Pattern observation: The total work grows roughly with the square of the number of time steps.

Final Time Complexity

Time Complexity: O(N^2)

This means if you double the number of time steps, the work roughly quadruples.

Common Mistake

[X] Wrong: "The loop runs N times, so the time is just O(N)."

[OK] Correct: Inside the loop, summing errors from 1 to k takes more work each step, adding up to much more than just N.

Interview Connect

Understanding how discrete-time steps affect computation helps you design efficient digital controllers and explain your reasoning clearly.

Self-Check

"What if we stored the running sum of errors instead of summing each time? How would the time complexity change?"

Practice

1. Why is discrete-time important in digital control systems modeled in Simulink?

easy

A. Because discrete-time eliminates the need for sampling

B. Because continuous signals are easier to process digitally

C. Because it allows infinite precision in signal values

D. Because digital controllers update signals at fixed time intervals

Solution

1. Step 1: Understand digital controller operation

   Digital controllers work by updating signals at fixed intervals, not continuously.

2. Step 2: Connect discrete-time to Simulink modeling

   Simulink models this by using discrete blocks with sample times to represent these fixed updates.

3. Final Answer:

   Because digital controllers update signals at fixed time intervals -> Option D

4. Quick Check:

   Digital control = discrete updates [OK]

Hint: Digital control means fixed update times, not continuous [OK]

Common Mistakes:

• Thinking digital control is continuous
• Confusing sampling with elimination
• Assuming infinite precision in digital signals

2. Which Simulink block parameter correctly sets the discrete sample time for a digital controller?

easy

A. Continuous sample time set to 0

B. Sample time set to a positive fixed value

C. Sample time set to -1 for inherited

D. Sample time set to infinity

Solution

1. Step 1: Identify how discrete sample time is set

   Discrete blocks require a positive fixed sample time to update at intervals.

2. Step 2: Evaluate options

   Setting sample time to 0 means continuous, -1 means inherited, infinity is invalid.

3. Final Answer:

   Sample time set to a positive fixed value -> Option B

4. Quick Check:

   Discrete sample time = positive value [OK]

Hint: Discrete sample time must be positive fixed number [OK]

Common Mistakes:

• Using 0 which means continuous time
• Using -1 which inherits sample time
• Setting sample time to infinity

3. Given this Simulink discrete-time update code snippet:

y[k] = 0.5 * y[k-1] + u[k]

and input sequence u = [2, 4, 6], with initial y[0] = 0, what is y[3]?

medium

A. 8.5

B. 7.0

C. 9.5

D. 10.0

Solution

1. Step 1: Calculate y[1], y[2], y[3] step-by-step

   y[1] = 0.5* y[0] + u[1] = 0.5*0 + 2 = 2
   y[2] = 0.5* y[1] + u[2] = 0.5*2 + 4 = 1 + 4 = 5
   y[3] = 0.5* y[2] + u[3] = 0.5*5 + 6 = 2.5 + 6 = 8.5

2. Step 2: Verify calculations

   Check each step carefully to avoid off-by-one errors in indexing.

3. Final Answer:

   8.5 -> Option A

4. Quick Check:

   Discrete update sums to 8.5 [OK]

Hint: Calculate each step using previous output and input [OK]

Common Mistakes:

• Mixing up indices k and k-1
• Forgetting initial condition y[0]
• Adding instead of multiplying coefficients

4. Identify the error in this Simulink discrete-time block setup:

Sample time = 0
Initial condition = 5

What is wrong?

medium

A. No error, setup is correct

B. Initial condition cannot be 5

C. Sample time 0 means continuous, not discrete

D. Sample time must be negative

Solution

1. Step 1: Understand sample time meaning

   Sample time 0 means continuous time, not discrete updates.

2. Step 2: Check initial condition validity

   Initial condition 5 is valid; no issue there.

3. Final Answer:

   Sample time 0 means continuous, not discrete -> Option C

4. Quick Check:

   Discrete sample time ≠ 0 [OK]

Hint: Sample time 0 means continuous, use positive for discrete [OK]

Common Mistakes:

• Thinking initial condition causes error
• Believing sample time can be zero for discrete
• Ignoring sample time meaning

5. You want to design a digital controller in Simulink that samples input every 0.1 seconds and holds output constant between samples. Which setup correctly models this behavior?

hard

A. Use a discrete block with sample time 0.1 and a Zero-Order Hold block

B. Use a continuous block with sample time 0 and a Unit Delay block

C. Use a discrete block with sample time -1 and no hold block

D. Use a continuous block with sample time 0.1 and a Zero-Order Hold block

Solution

1. Step 1: Identify correct sample time setting

   Discrete block with sample time 0.1 means updates every 0.1 seconds.

2. Step 2: Understand output holding

   Zero-Order Hold block keeps output constant between samples, matching requirement.

3. Final Answer:

   Use a discrete block with sample time 0.1 and a Zero-Order Hold block -> Option A

4. Quick Check:

   Discrete sample + hold = correct model [OK]

Hint: Discrete block + Zero-Order Hold models sampled and held output [OK]

Common Mistakes:

• Using continuous block for discrete sampling
• Omitting Zero-Order Hold block
• Using inherited sample time (-1) incorrectly