Bilinear Transformation Technique for IIR FIlter Design

Summary

In this video, the design of an IIR filter using the bilinear transformation is explained. The bilinear transformation is used to convert an analog filter to a digital one, addressing the frequency aliasing problem associated with impulse invariant transformation. The video covers the process of pre-warping analog frequencies, designing the analog filter, and then applying the bilinear transformation. An example is provided, demonstrating how the transformation works step by step. The key focus is on ensuring a linear frequency mapping between the analog and digital domains, avoiding frequency warping effects and ensuring accurate filter design.

Takeaways

• 😀 The main goal is to design a digital IIR filter from an analog filter using bilinear transformation.
• 😀 Impulse invariant transformation can cause many-to-one mapping of poles, leading to aliasing and loss of information.
• 😀 Bilinear transformation provides a one-to-one mapping from the s-plane (analog) to the z-plane (digital), avoiding overlapping issues.
• 😀 The bilinear transformation formula is s = (2/T) * (1 - z^-1) / (1 + z^-1), where T is the sampling period.
• 😀 Frequency warping occurs because the mapping of analog frequency to digital frequency is nonlinear at higher frequencies.
• 😀 Pre-warping is used to maintain a linear relationship between analog and digital frequencies using Ωa = (2/T) * tan(Ωd/2).
• 😀 For small digital frequencies, tan(Ωd/2) ≈ Ωd/2, so pre-warping has minimal effect in low-frequency ranges.
• 😀 The design steps include pre-warping digital frequencies, designing the analog filter, and then applying bilinear transformation to obtain the digital filter.
• 😀 Example: An analog filter H(s) = 2 / ((s+1)(s+2)) is converted to a digital form H(z) = 0.166 * (1 + z^-1)^2 / (1 - 0.33 z^-1) for implementation.
• 😀 Bilinear transformation ensures the digital filter closely matches the analog filter’s frequency response while being realizable in digital systems.

Q & A

What is the main objective of using bilinear transformation in IIR filter design?

The main objective is to transform a designed analog filter into an equivalent digital filter while avoiding issues like frequency overlapping and many-to-one mapping of poles that occur with impulse invariance methods.

What is the major drawback of the impulse invariance method?

Impulse invariance causes many-to-one mapping of poles from the s-plane to the z-plane, resulting in overlapping of frequency components and loss of information.

How does the bilinear transformation map analog to digital frequencies?

Bilinear transformation maps the analog s-plane to the digital z-plane using the formula s = (2/T) * ((1 - z^-1) / (1 + z^-1)), ensuring one-to-one mapping of poles and avoiding aliasing.

What is frequency warping and why does it occur?

Frequency warping is the nonlinear relationship between analog and digital frequencies when using bilinear transformation. It occurs because the transformation inherently compresses higher frequencies nonlinearly.

How can the effect of frequency warping be minimized?

Frequency warping can be minimized by pre-warping the analog frequencies using the formula Omega = (2/T) * tan(omega/2), which linearizes the relationship for design purposes.

What are the steps involved in designing a digital IIR filter using bilinear transformation?

1) Pre-warp digital specifications to analog equivalents; 2) Design the analog filter; 3) Apply bilinear transformation; 4) Simplify the digital filter for implementation.

Why is pre-warping important in the bilinear transformation process?

Pre-warping ensures that the frequency characteristics of the analog filter are accurately reflected in the digital filter, compensating for nonlinear frequency compression caused by the transformation.

In the example provided, what was the digital filter obtained from H(s) = 2 / ((s + 1)(s + 2))?

After applying bilinear transformation and simplification, the digital filter was H(z) = 0.166 * (1 + z^-1)^2 / (1 - 0.33 z^-1), ready for digital implementation.

What is the significance of expressing the digital filter in the form H(z) = (b0 + b1 z^-1 + ...)/(1 + a1 z^-1 + ...)?

This form is standard for digital implementation, making it easier to code and compute in DSP systems while preserving the filter's characteristics.

Does bilinear transformation completely eliminate frequency distortion?

Bilinear transformation reduces aliasing and overlapping issues, but frequency warping still occurs, which must be corrected using pre-warping to maintain accurate frequency response.

How is the sample time T related to the bilinear transformation?

The sample time T is used in the transformation formula s = (2/T) * ((1 - z^-1) / (1 + z^-1)) and affects the scaling of analog frequencies during the conversion to digital form.

Can bilinear transformation be applied directly without pre-warping?

Technically yes, but it may cause inaccuracies in frequency response at higher frequencies due to warping, so pre-warping is recommended for precise design.