Impulse Response Estimation via Laguerre-Fourier Expansion
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Electrical Engineering. The day in electrical engineering: signals, systems and the hardware that carries them.
Rosa: We're looking at the paper: "Impulse Response Estimation via Laguerre-Fourier Expansion".
Dev: Here's the gist of the paper. Impulse Response Estimation via Laguerre-Fourier Expansion presents L-ETFE, a generalization of ETFE based on Laguerre-Fourier expansions that addresses numerical limitations when excitation signals vanish at certain frequencies This method yields accurate IR recovery for spectral-zero and band-limited excitation where standard ETFE fails The empirical transfer function estimate (ETFE) suffers from numerical limitations when the excitation signal is band-limited or vanishes at certain frequencies, which is a common physical constraint of the excitation in practice In such cases, division in the frequency domain becomes heavily ill-conditioned, and small measurement disturbances or numerical inaccuracies can degrade the solution The paper proposes L-ETFE to address these limitations by representing the transfer function by its Laguerre-Fourier coefficients The core of the method involves transforming the problem into one solvable via classical ETFE after a suitable transformation yields a well-conditioned circulant problem The authors propose two algorithmic variants for computing discrete Laguerre-Fourier coefficients from time-domain input-output data by solving a linear system with circulant Toeplitz structure Furthermore, an algorithm is proposed for impulse response recovery from Laguerre-Fourier coefficients, completing the Laguerre ETFE pipeline Key Theoretical Results Several novel results characterize the method's performance and structure The authors derive a condition number bound characterizing the dependence of the Laguerre-Fourier coefficients on the excitation signal and a parameter used in the Laguerre-Fourier expansion Theorem 3.2 shows that if certain conditions are met, the spectral condition number of circulant matrices is bounded by an expression involving input terms The complete method is verified on two discrete-time LTI systems of varying complexity, demonstrating accurate IR recovery for cases where standard ETFE fails Experimental Validation The proposed methodology is validated through experiments on two discrete-time LTI systems with varying complexity Experiment 1 shows that L-ETFE with a = 0 reproduces the ETFE reconstruction, while choosing a non-zero Laguerre parameter leads to substantially larger errors for N = 10,000, 15,000, and 20,000 Experiment 3 demonstrates the method's capability with band-limited input signals where ETFE cannot be used to recover values outside the support of U The results show that L-ETFE can capture system behavior over unexcited frequencies, which is useful for identification scenarios where physical constraints allow only band-limited input signals The study concludes that the proposed methodology can be used to recover the IR of very large LTI systems and can recover information about system behaviour over unexcited frequencies The authors plan to extend their method to multiple-input multiple-output (MIMO) systems, identify large acoustic systems in real applications, and develop a method to automatically obtain the Laguerre parameter The complete pipeline is implemented in Algorithm 5, combining coefficient recovery and IR recovery CRediT Author Statement Tam´as D´ozsa: Conceptualization, Methodology, Software, Formal Analysis, Investigation, Writing - Original Draft, Visualization; Art J.
Rosa: R.
Dev: Pelling: Conceptualization, Methodology, Software, Validation, Investigation, Data Curation and Writing - Original Draft; Matthias Voigt: Writing - Review & Editing and Supervision Funding Acquisition The work of Art J.
Rosa: R.
Title and authors: Rosa: So we're looking at this paper called "Impulse Response Estimation via Laguerre-Fourier Expansion." Rosa, what's the big picture here? It sounds like they’re trying to fix a problem with a common identification tool.
Dev: Exactly. The empirical transfer function estimate, or ETFE, is really used for figuring out how to recover the impulse response of linear time-invariant systems in acoustics and other fields, but this paper points out that ETFE has these numerical headaches when the input signal is band-limited or hits zero at certain frequencies.
Rosa: Right. So it’s like when you’re trying to measure something, and your measurement setup just can't handle certain frequencies because the signal there disappears, which makes the math really unstable >
Taro: It means that for systems where the excitation signal vanishes at specific frequencies, ETFE starts struggling because dividing in the frequency domain becomes really ill-conditioned, and any tiny measurement errors or numerical inaccuracies get magnified and degrade your final result >
Dev: And this is where they bring in L-ETFE. They propose a generalization using Laguerre-Fourier expansions to address these specific limitations that standard ETFE has >
The paper's summary: Rosa: So, what does the actual method involve? What exactly is L-ETFE doing differently compared to the old ETFE approach we talked about? We need to understand the core idea behind this Laguerre expansion stuff.
Dev: The summary says that they transform the problem so it becomes solvable using a more classical ETFE setup after some transformation yields a well-conditioned circulant problem >
Rosa: A circulant problem, which sounds much nicer than whatever ill-conditioned division we were worried about before. But how do they actually get those discrete Laguerre-Fourier coefficients from the time domain data?
Dev: They introduce a novel and numerically efficient method for computing these discrete Laguerre Fourier coefficients using only scaling, shifting, and Fourier transforms >
Rosa: Scaling, shifting, and FFTs. That sounds clever because it suggests we don't need some massive matrix inversion when we are trying to get the representation of the impulse response >
Taro: And they also suggest a way to recover the impulse response from those coefficients using reproducing kernel Hilbert space properties, which is how they complete this whole L-ETFE pipeline >
The paper's improvements: Dev: Beyond just fixing the calculation of the coefficients, there are several specific theoretical results they present that characterize how well this new method performs. They derive a condition number bound that tells you exactly how sensitive those Laguerre Fourier coefficients are to the input signal and the Laguerre parameter they choose >
Rosa: That condition number bound is key because it gives us a concrete way to know when we have enough stability in the transformed problem, which is something we really need for engineering applications >
Dev: Then there's Theorem three point two, which shows that under certain conditions, the spectral condition number of the circulant matrices involved is bounded by an expression that depends on those input terms > <ref:2608.14769#pg1>
Taro: So they’ve mathematically characterized the stability of this whole process, proving it behaves predictably under specific circumstances >
Rosa: And experimentally, they validated this whole thing on two discrete-time systems with varying complexity, showing accurate impulse response recovery even when standard ETFE fails >
Conclusion: Dev: So to wrap up, the main point of the paper "Impulse Response Estimation via Laguerre-Fourier Expansion" is that L-ETFE provides a way to recover the impulse response of very large linear time-invariant systems and even lets you get information about system behavior over frequencies where you didn't actually excite the system >
Rosa: It seems like this is really useful for identification scenarios where physical constraints only allow band-limited input signals, because L-ETFE can capture that behavior over unexcited frequencies >
Taro: I just want to push on what it means practically. If you have a real acoustic system and your measurement setup has these frequency limitations, this method gives you a path to get the true impulse response without getting stuck with unstable division issues >
Dev: Yeah, that’s the implication for us in real-time loops. We can identify these systems more robustly even when we don't have perfect excitation signals available >
Rosa: It seems like they’ve built a complete pipeline now, from recovering those coefficients to getting the final impulse response using RKHS properties, which is pretty comprehensive >
Taro: I just wonder about the next steps. The authors mention planning to extend this method to multiple-input multiple-output systems and develop a way to automatically find that Laguerre parameter itself >
Dev: That makes sense. Moving from single-input systems to MIMO systems is a natural progression, and automating the laguerre parameter finding would make it much easier for real-world deployment >
Rosa: It sounds like they've done a lot of heavy lifting on making this identification method more robust and flexible, especially when dealing with those tricky frequency constraints we talked about >
Faculty of Mathematics and Computer Science, UniDistance Suisse · Department of Engineering Acoustics, Technische Universität Berlin
eess.SP, math.CV, math.DS
Submitted: 2026-08-14
Updated: 2026-11-10
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Importance score: 77/100
The gist: Impulse Response Estimation via Laguerre-Fourier Expansion presents L-ETFE, a generalization of ETFE based on Laguerre-Fourier expansions that addresses numerical limitations when excitation signals…
Key concepts
- ETFE
- Empirical Transfer Function Estimation is a technique used to estimate the system's transfer function from input-output data. It often fails when the input signal is severely limited in frequency range or has specific spectral zeros, leading to numerical instability and inaccurate results.
- Laguerre-Fourier Expansion
- This involves representing the system's transfer function using a specific mathematical expansion based on Laguerre functions. This transformation converts a potentially ill-conditioned problem into one that is solvable via classical ETFE methods, improving numerical stability.
- Circulant Toeplitz Structure
- The transformation yields a problem with a circulant Toeplitz structure, which is mathematically well-behaved and easy to solve compared to the original system. This structure allows for the efficient computation of Laguerre-Fourier coefficients using standard linear algebra techniques.
- Reproducing Kernel Hilbert Space (RKHS)
- RKHS properties are used in the final step to recover the impulse response from its calculated Laguerre-Fourier coefficients. This ensures that the recovered impulse response is mathematically sound and corresponds correctly to a BIBO-stable system.
Terminology
Summary
Impulse Response Estimation via Laguerre-Fourier Expansion presents L-ETFE, a generalization of ETFE based on Laguerre-Fourier expansions that addresses numerical limitations when excitation signals vanish at certain frequencies This method yields accurate IR recovery for spectral-zero and band-limited excitation where standard ETFE fails
Motivation and Problem Statement
The empirical transfer function estimate (ETFE) suffers from numerical limitations when the excitation signal is band-limited or vanishes at certain frequencies, which is a common physical constraint of the excitation in practice In such cases, division in the frequency domain becomes heavily ill-conditioned, and small measurement disturbances or numerical inaccuracies can degrade the solution The paper proposes L-ETFE to address these limitations by representing the transfer function by its Laguerre-Fourier coefficients
Methodology and Novel Contributions
The core of the method involves transforming the problem into one solvable via classical ETFE after a suitable transformation yields a well-conditioned circulant problem The authors propose two algorithmic variants for computing discrete Laguerre-Fourier coefficients from time-domain input-output data by solving a linear system with circulant Toeplitz structure Furthermore, an algorithm is proposed for impulse response recovery from Laguerre-Fourier coefficients, completing the Laguerre ETFE pipeline
Key Theoretical Results
Several novel results characterize the method's performance and structure The authors derive a condition number bound characterizing the dependence of the Laguerre-Fourier coefficients on the excitation signal and a parameter used in the Laguerre-Fourier expansion Theorem 3.2 shows that if certain conditions are met, the spectral condition number of circulant matrices is bounded by an expression involving input terms The complete method is verified on two discrete-time LTI systems of varying complexity, demonstrating accurate IR recovery for cases where standard ETFE fails
Experimental Validation
The proposed methodology is validated through experiments on two discrete-time LTI systems with varying complexity Experiment 1 shows that L-ETFE with a = 0 reproduces the ETFE reconstruction, while choosing a non-zero Laguerre parameter leads to substantially larger errors for N = 10,000, 15,000, and 20,000 Experiment 3 demonstrates the method's capability with band-limited input signals where ETFE cannot be used to recover values outside the support of U The results show that L-ETFE can capture system behavior over unexcited frequencies, which is useful for identification scenarios where physical constraints allow only band-limited input signals
Conclusion and Future Work
The study concludes that the proposed methodology can be used to recover the IR of very large LTI systems and can recover information about system behaviour over unexcited frequencies The authors plan to extend their method to multiple-input multiple-output (MIMO) systems, identify large acoustic systems in real applications, and develop a method to automatically obtain the Laguerre parameter The complete pipeline is implemented in Algorithm 5, combining coefficient recovery and IR recovery
CRediT Author Statement
Tam´as D´ozsa: Conceptualization, Methodology, Software, Formal Analysis, Investigation, Writing - Original Draft, Visualization; Art J. R. Pelling: Conceptualization, Methodology, Software, Validation, Investigation, Data Curation and Writing - Original Draft; Matthias Voigt: Writing - Review & Editing and Supervision Funding Acquisition The work of Art J. p. 0 R. Pelling was funded by the Deutsche Forschungsgemeinschaft (DFG) – project no. p. 0 504367810, and the research of Tam´as D´ozsa has received funding from the Swiss Government Excellence Scholarship No. p. 0 2025.0057, while Matthias Voigt was funded in part by the Swiss National Science Foundation (SNSF) grant No. p. 0 224943 The source code and scripts used to compute the results presented in this paper can be obtained from doi:10.5281/zenodo.21934090 under the MIT licence, authored by Art J. p. 0 R. Pelling and Tam´as D´ozsa The work of Alexandros Soumelidis was thanked for insightful discussions and early feedback over some of the technical details in the manuscript The funding sources are detailed in Appendix D
Acronyms
BIBO bounded-input bounded-output ETFE empirical transfer function estimate FFT fast Fourier transform IR impulse response L-ETFE Laguerre empirical transfer function estimate LTI linear time-invariant MIMO multiple-input multiple-output RKHS reproducing kernel Hilbert space SISO single-input single-output The acronyms are defined in the section dedicated to them
Appendix A. Theoretical Background
The paper develops the mathematical background for the proposed method, including a finite-dimensional deconvolution model and discrete-time LTI systems It introduces Laguerre-Fourier expansions in H2(D) and defines discrete Laguerre-Fourier coefficients The paper discusses the properties of Blaschke factors and their use in discretization sets The paper also details the inner product definitions used for the discrete Laguerre-Fourier coefficients
Appendix B. Considered System Class
The authors discuss important details about the considered class of systems, assuming strict properness of the transfer function and that He(0) = 0 The paper shows that this implies b0 = 0, which is necessary for a strictly proper transfer function This section discusses why (16) holds when using the Z-transform with non-negative exponents as defined in (13)
Appendix C. Proofs
This appendix contains proofs for key theorems, including the Christoffel-Darboux formula and Lemma 2.2 concerning the spectral condition number of circulant matrices The proof of Theorem 3. p. 62 relies on analyzing the eigenvalues of the circulant matrix Γ and relating them to input terms
**Appendix D.
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Generalize system identification from empirical transfer function estimate (ETFE) to Laguerre-ETFE (L-ETFE), which addresses numerical limitations when excitation signals are band-limited or have spectral zeros by using Laguerre-Fourier expansions. This allows for accurate impulse response recovery even in scenarios where standard ETFE fails.
-
Develop a novel and numerically efficient method for computing discrete Laguerre Fourier coefficients of the impulse response using only
scaling, shifting, and Fourier transforms.
This transformation can yield awell-conditioned circulant problem even when the original ETFE system is ill-conditioned.
-
Implement an algorithm to recover the impulse response from these Laguerre-Fourier coefficients via reproducing kernel Hilbert space (RKHS) properties. This completes the L-ETFE pipeline, allowing for direct estimation of IR from data.
-
Design two algorithmic variants for computing discrete Laguerre Fourier coefficients by solving a linear system with circulant Toeplitz structure, one solving it directly and the other using only a single fast Fourier transform (FFT).
-
Derive a condition number bound characterizing the dependence of the Laguerre-Fourier coefficients on the input excitation signal and the Laguerre parameter used in the transformation. This bound explicitly characterizes how well
the transformed problem
is conditioned. -
Create a numerically stable scheme for IR recovery from Laguerre-Fourier coefficients using
reproducing kernel Hilbert space (RKHS) properties.
This ensures that the recovered impulse response h satisfies Z[h] = H, where H is the transfer function of a BIBO-stable system. -
Enable identification of system behavior over unexcited frequencies by exploiting
Laguerre functions
which form an orthonormal basis in the Hardy space H2(D). This is particularly useful for scenarios where physical constraints limit excitation signals to be band-limited.
Related papers
- Open-H-RF: A Large-Scale Ultrasound Raw Channel Dataset Initiative
- Scalable Attention for 5G NR Channel Estimation
- Low-Altitude Wireless Networks: A Comprehensive Survey
- Sensing-Assisted OFDM with Index Modulation: A Waveform for Optical Wireless Integrated Communication and Localisation
- Iteration-Free Candidate Pruning and an Index-Domain VLSI Architecture for Sort-Free MIMO Box Decoding up to 1024-QAM
- A Foundation Model for Large-Scale Wireless Network Planning, Operation and Optimization