On sub-ideal causal smoothing filters

Smoothing causal linear time-invariant filters are studied for continuous time processes. The paper suggests a family of causal filters with almost exponential damping of the energy on the higher frequencies. These filters are sub-ideal meaning that a faster decay of the frequency response would lead to the loss of causality.

💡 Research Summary

The paper addresses a fundamental problem in continuous‑time signal processing: how to design causal linear time‑invariant (LTI) smoothing filters that strongly attenuate high‑frequency components while preserving causality. Classical ideal low‑pass or exponential‑decay filters (with frequency response |H(iω)|≈e^{‑α|ω|}) cannot be realized causally because the Paley‑Wiener theorem dictates that such rapid decay forces poles or zeros into the right half of the complex s‑plane, violating the causality condition. The authors therefore introduce the notion of a “sub‑ideal” filter – a filter that achieves the fastest possible decay compatible with causality.

The main theoretical contribution is a set of rigorous bounds that delineate the achievable decay rate. The authors prove that for any positive attenuation constant a, a decay exponent γ≥1 leads inevitably to non‑causal behavior. Consequently, any causal smoothing filter must satisfy a decay of the form |H(iω)|≈e^{‑a|ω|^{γ}} with 0<γ<1. This establishes a precise trade‑off: the closer γ is to 1, the steeper the attenuation, but the more precarious the causality condition.

To construct filters that meet this bound, the paper proposes a parametric family of transfer functions:

 H_{a,β,γ}(s)=exp