Digital filter implementation in Simulink

Introduction

A digital filter helps clean or change signals by removing unwanted parts or enhancing useful parts.

Removing noise from audio recordings to hear voices clearly.

Smoothing sensor data from a wearable device to get steady readings.

Separating different frequency parts in a music track for effects.

Improving image quality by filtering pixel values.

Extracting heartbeats from noisy ECG signals.

Syntax

Simulink

1. Open Simulink.
2. Add a 'Digital Filter' block from the DSP System Toolbox.
3. Set filter parameters (type, order, cutoff frequencies).
4. Connect input signal to the filter block.
5. Connect filter output to display or sink block.
6. Run the simulation.

You can choose filter types like low-pass, high-pass, band-pass, or band-stop.

Filter order controls how sharp the filter works; higher order means sharper filtering but more computation.

Examples

This keeps signals below 1000 Hz and removes higher frequencies.

Simulink

Use a low-pass filter block with cutoff frequency 1000 Hz to remove high-frequency noise.

This allows only signals in this range to pass through, blocking others.

Simulink

Set a band-pass filter block to pass frequencies between 300 Hz and 3000 Hz.

This removes slow changes or trends in the signal below 500 Hz.

Simulink

Use a high-pass filter block with cutoff frequency 500 Hz to remove low-frequency drift.

Sample Program

This example shows how to filter a sine wave signal to remove frequencies above 50 Hz using a low-pass FIR filter in Simulink.

Simulink

1. Open Simulink and create a new model.
2. Add a 'Sine Wave' block to generate a signal.
3. Add a 'Digital Filter' block (from DSP System Toolbox).
4. Double-click the filter block and set it to 'Lowpass FIR' with cutoff frequency 50 Hz.
5. Connect the Sine Wave output to the filter input.
6. Add a 'Scope' block to see the output.
7. Connect the filter output to the Scope input.
8. Run the simulation.

% This setup filters the sine wave signal to keep frequencies below 50 Hz.

Important Notes

Always check the filter type and parameters to match your signal needs.

Simulation speed may slow down with very high filter orders.

Use the Scope block to visually check how the filter changes the signal.

Summary

Digital filters clean or change signals by removing or keeping certain frequencies.

Simulink uses blocks to build and run digital filters visually.

Choose filter type and parameters carefully to get the desired signal effect.

Practice

1. What is the main purpose of a digital filter in signal processing within Simulink?

easy

A. To increase the amplitude of all frequencies equally

B. To convert analog signals to digital signals

C. To remove or keep certain frequencies in a signal

D. To generate random noise in the signal

Solution

1. Step 1: Understand digital filter function

   Digital filters are designed to modify signals by removing or preserving specific frequency components.

2. Step 2: Identify correct purpose in Simulink context

   Simulink uses digital filters to clean or change signals by targeting frequencies, not to convert or add noise.

3. Final Answer:

   To remove or keep certain frequencies in a signal -> Option C

4. Quick Check:

   Digital filter = frequency selection [OK]

Hint: Filters change signals by frequency, not by adding noise [OK]

Common Mistakes:

• Confusing filtering with signal conversion
• Thinking filters amplify all frequencies equally
• Assuming filters add noise intentionally

2. Which Simulink block is commonly used to implement a digital FIR filter?

easy

A. Discrete FIR Filter

B. Analog Filter

C. Integrator

D. Gain

Solution

1. Step 1: Identify FIR filter block in Simulink

   The 'Discrete FIR Filter' block is specifically designed for FIR digital filtering.

2. Step 2: Exclude unrelated blocks

   'Analog Filter' is not digital, 'Integrator' and 'Gain' do not perform filtering.

3. Final Answer:

   Discrete FIR Filter -> Option A

4. Quick Check:

   FIR filter block = Discrete FIR Filter [OK]

Hint: FIR filters use 'Discrete FIR Filter' block in Simulink [OK]

Common Mistakes:

• Choosing 'Analog Filter' for digital filtering
• Confusing 'Integrator' with filter blocks
• Selecting 'Gain' which only scales signals

3. Consider this Simulink filter setup: a Discrete FIR Filter with coefficients [0.2, 0.5, 0.3]. What is the output for input signal [1, 2, 3, 4] at the third sample?

medium

A. 3.0

B. 1.9

C. 2.3

D. 1.5

Solution

1. Step 1: Recall FIR filter computation

   The output at time n is y(n) = 0.2*x(n) + 0.5*x(n-1) + 0.3*x(n-2).

2. Step 2: Identify inputs for third sample (n=3)

   x(3)=3, x(2)=2, x(1)=1 (zero padding before n=1).

3. Step 3: Compute y(3)

   0.2*3 + 0.5*2 + 0.3*1 = 0.6 + 1.0 + 0.3 = 1.9.

4. Final Answer:

   1.9 -> Option B

5. Quick Check:

   y(3) = sum(coeffs * inputs) = 1.9 [OK]

Hint: Multiply coefficients by recent inputs, sum for output [OK]

Common Mistakes:

• Mixing order of coefficients and inputs
• Ignoring zero padding at start
• Calculating output for wrong sample index

4. You implemented a digital filter in Simulink but the output signal is delayed unexpectedly. What is the most likely cause?

medium

A. Filter introduces group delay due to its order

B. Sampling time is set to zero

C. Input signal is constant zero

D. Filter coefficients are all zeros

Solution

1. Step 1: Understand filter delay concept

   Digital filters, especially FIR, introduce delay proportional to their order (number of coefficients).

2. Step 2: Analyze options for delay cause

   All zero coefficients produce zero output, constant zero input produces zero output, zero sampling time is invalid. Only filter order causes delay.

3. Final Answer:

   Filter introduces group delay due to its order -> Option A

4. Quick Check:

   Filter order causes delay = Filter introduces group delay due to its order [OK]

Hint: Filter order causes delay, not zero coefficients or input [OK]

Common Mistakes:

• Assuming zero coefficients cause delay
• Confusing input signal with filter behavior
• Setting sampling time to zero without error

5. You want to design a low-pass digital filter in Simulink to remove noise above 1 kHz from a signal sampled at 10 kHz. Which approach is best to implement this filter?

hard

A. Use an Analog Filter block with cutoff at 1 kHz

B. Use a Gain block to amplify frequencies below 1 kHz

C. Use a Discrete FIR Filter block with cutoff frequency set to 10 (normalized)

D. Use a Discrete FIR Filter block with cutoff frequency set to 0.2 (normalized)

Solution

1. Step 1: Understand normalized cutoff frequency

   Normalized frequency = cutoff frequency / (sampling frequency / 2). Here, 1 kHz / (10 kHz / 2) = 0.2.

2. Step 2: Choose correct cutoff frequency for filter block

   Cutoff frequency should be less than 1 (Nyquist normalized). Use a Discrete FIR Filter block with cutoff frequency set to 0.2 (normalized).

3. Step 3: Exclude incorrect options

   Gain block does not filter frequencies, Analog Filter block is not digital, cutoff 10 is invalid (above Nyquist).

4. Final Answer:

   Use a Discrete FIR Filter block with cutoff frequency set to 0.2 (normalized) -> Option D

5. Quick Check:

   Normalized cutoff < 1 for digital low-pass filter [OK]

Hint: Cutoff freq normalized = cutoff/(Fs/2), use FIR block [OK]

Common Mistakes:

• Using analog filter instead of digital
• Setting cutoff frequency above Nyquist (1)
• Using Gain block for filtering