Error Rate Analysis and Low-Complexity Receiver Design for Zero-Padded AFDM

Basic Information

• Paper ID: 2510.14507
• Title: Error Rate Analysis and Low-Complexity Receiver Design for Zero-Padded AFDM
• Authors: Qin Yi, Zeping Sui, Zilong Liu
• Classification: eess.SP cs.IT math.IT
• Publication Date: October 16, 2025
• Paper Link: https://arxiv.org/abs/2510.14507

Abstract

This paper investigates the bit error rate (BER) performance and low-complexity receiver design for zero-padded affine frequency division multiplexing (ZP-AFDM) systems. By exploiting the unique zero-padded lower triangular structure of the time-domain (TD) channel matrix, a novel low-complexity minimum mean square error (MMSE) detector and a maximum ratio combining time-domain (MRC-TD) detector are proposed. Furthermore, the theoretical BER performance of MMSE and maximum likelihood detectors is analyzed. Simulation results demonstrate that the proposed detectors achieve the same BER performance as the conventional matrix inversion-based MMSE detector while significantly reducing computational complexity.

Research Background and Motivation

Problem Definition

Next-generation wireless communication systems must support reliable data transmission in high-mobility scenarios, including high-speed trains, vehicular communication systems, and low-earth orbit satellite networks. In these scenarios, conventional orthogonal frequency division multiplexing (OFDM) becomes infeasible due to the destruction of subcarrier orthogonality and increased inter-carrier interference.

Importance Analysis

Affine frequency division multiplexing (AFDM) has emerged as a Doppler-resilient waveform technology. In AFDM, each information symbol is modulated via inverse discrete affine Fourier transform (IDAFT) onto chirp subcarriers occupying the entire bandwidth, thereby enhancing robustness against Doppler spread.

Limitations of Existing Approaches

Existing research primarily focuses on AFDM systems with cyclic prefix padding (CPP). Although low-complexity linear detection schemes (such as MRC detectors) based on approximate sparsity of the DAFT-domain channel matrix have been developed, they exhibit significant performance degradation compared to conventional MMSE detection.

Research Motivation

Zero-padding (ZP) technology has been widely adopted in OFDM as an alternative to cyclic prefix, capable of eliminating inter-symbol interference (ISI) while maintaining favorable sparsity and lower triangular structure of the time-domain channel matrix. ZP systems offer higher power transmission efficiency, reliable symbol recovery under deep channel fading, and enhanced channel estimation and tracking capabilities.

Core Contributions

1. System Architecture Design: Derives a general transceiver architecture for ZP-AFDM systems and designs two low-complexity detectors:
   • MMSE detector based on Cholesky decomposition (without matrix inversion)
   • Maximum ratio combining time-domain (MRC-TD) detector
2. Theoretical Performance Analysis: Derives closed-form BER expressions for ZP-AFDM under maximum likelihood (ML) and MMSE detection:
   • ML detector: Obtains conditional pairwise error probability (PEP) first, then derives tight BER upper bounds based on union bound techniques
   • MMSE detector: Approximates BER by analyzing signal-to-interference-plus-noise ratio (SINR) for each chirp subcarrier
3. Performance Verification: Numerical results demonstrate:
   • Derived BER curves closely match Monte Carlo simulation results
   • ZP-AFDM outperforms CPP-AFDM (the former allocates higher power to data symbols)
   • Proposed detectors achieve nearly identical performance to conventional MMSE detectors while significantly reducing complexity

Methodology Details

Task Definition

Design low-complexity detection algorithms for ZP-AFDM systems under doubly-selective channels, with input as received signal vector and output as estimated information symbol vector, constrained to maintain the same BER performance as conventional MMSE detectors.

System Model

AFDM Modulation

Information symbol vector $\mathbf{x} = [x(0), x(1), \ldots, x(N-1)]^T \in \mathbb{C}^{N \times 1}$ is transformed to time-domain via IDAFT:

$s(n) = \frac{1}{\sqrt{N}} \sum_{m=0}^{N-1} x(m) e^{j2\pi(c_1n^2 + c_2m^2 + nm/N)}$

where $N$ is the number of chirp subcarriers, and $c_1$ and $c_2$ are chirp parameters.

Channel Model

A doubly-selective channel with $P$ propagation paths has impulse response at time $n$ and delay $l$:

$g_n(l) = \sum_{i=1}^P h_i e^{-j\frac{2\pi}{N}\nu_i n} \delta(l - l_i)$

where $h_i$, $\nu_i$, and $l_i$ denote the complex path gain, normalized Doppler frequency shift, and integer delay of the $i$-th path, respectively.

Detection Algorithms

1. Conventional MMSE Detection

The linear MMSE detector is expressed as: $\hat{\mathbf{x}}_{MMSE} = \mathbf{G}_{AF} \mathbf{y}$

where the equalization matrix is: $\mathbf{G}_{AF} = \left(\mathbf{H}_{eff}^H \mathbf{H}_{eff} + \frac{1}{\gamma_s} \mathbf{I}_N\right)^{-1} \mathbf{H}_{eff}^H$

2. Proposed Low-Complexity MMSE Detection

Exploiting the lower triangular sparse structure of the time-domain channel matrix $\mathbf{H}$, the matrix $\boldsymbol{\Psi} = \mathbf{H}^H\mathbf{H} + \frac{1}{\gamma_s}\mathbf{I}_N$ exhibits a banded structure with bandwidth $(2Q+1)$. Using Cholesky decomposition $\boldsymbol{\Psi} = \mathbf{L}\mathbf{L}^H$, the estimated symbol vector is:

$\hat{\mathbf{x}}_{MMSE} = \mathbf{A}(\mathbf{L}^H)^{-1}\mathbf{L}^{-1}\mathbf{H}^H\mathbf{r}$

Efficient computation via forward and backward substitution avoids matrix inversion operations.

3. Proposed MRC-TD Detection

Directly exploits the sparse lower triangular structure of the time-domain channel matrix. At the $k$-th iteration, each symbol is sequentially updated as:

$\hat{s}^{(k)}(n) = \frac{g_n^{(k)}}{d_n + \frac{1}{\gamma_s}}$

where: $g_n^{(k)} = \sum_{p \in \mathcal{P}_n} H(p,n)^* \Delta r^{(k-1)}(p) + d_n \hat{s}^{(k-1)}(n)$$d_n = \sum_{p \in \mathcal{P}_n} |H(p,n)|^2$

Complexity Analysis

• Conventional MMSE detector: $O(N^3)$
• Proposed low-complexity MMSE detector: $O(NQ^2 + N\log N)$
• Proposed MRC-TD detector: $O(KNQ + N\log N)$

where $Q$ is the maximum delay spread and $K$ is the number of iterations.

Performance Analysis

BER Analysis for ML Detection

The conditional pairwise error probability is: $\Pr(\mathbf{x} \to \hat{\mathbf{x}} | \mathbf{h}) = Q\left(\sqrt{\|\boldsymbol{\Phi}(\boldsymbol{\Delta})\mathbf{h}\|^2 \cdot \gamma_s/2}\right)$

where $\boldsymbol{\Delta} = \mathbf{x} - \hat{\mathbf{x}}$, $\boldsymbol{\Phi}(\mathbf{x}) = [\mathbf{H}_1\mathbf{x}, \mathbf{H}_2\mathbf{x}, \cdots, \mathbf{H}_P\mathbf{x}]$.

The unconditional PEP can be expressed as: $\Pr(\mathbf{x} \to \hat{\mathbf{x}}) \approx \frac{1}{12}\prod_{i=1}^r \frac{1}{1 + \frac{\lambda_i \gamma_s}{4P}} + \frac{1}{4}\prod_{i=1}^r \frac{1}{1 + \frac{\lambda_i \gamma_s}{3P}}$

BER Analysis for MMSE Detection

The SINR for the $i$-th chirp subcarrier is: $\beta_i = \frac{T(i,i)^2}{\text{Var}\left(\sum_{j \neq i} T(i,j)x(j) + \bar{w}(i)\right)} = \frac{T(i,i)}{1 - T(i,i)}$

The theoretical average BER can be approximated as: $P_{MMSE} = \frac{1}{N} \sum_{i=0}^{N-1} a_M \text{erfc}\left(\sqrt{b_M \frac{T(i,i)}{1-T(i,i)}}\right)$

Experimental Setup

Simulation Parameters

• Channel coefficients modeled as $h_i \sim \mathcal{CN}(0, 1/P)$
• Maximum delay spread set to $Q = P - 1$
• Doppler frequency shifts generated according to Jake's model: $\nu_i = \nu_{max}\cos(\theta_i)$, where $\theta_i \sim \mathcal{U}[-\pi, \pi]$
• Primary simulation parameters: $N = 256$, $N_{ZP} = 64$

Evaluation Metrics

• Bit error rate (BER) performance
• Computational complexity (number of multiplication operations)
• Matching degree between theoretical analysis and simulation results

Experimental Results

Main Results

1. Theoretical Analysis Verification

• ML detector BER curves closely match theoretical bounds at high SNR, validating the effectiveness of error performance analysis
• MMSE detector simulation results perfectly overlap with analytical curves, confirming the accuracy of derived BER expressions

2. Performance Comparison

• ZP-AFDM consistently outperforms CPP-AFDM because the former allocates all transmission energy to data symbols
• Under identical configurations, both ZP-AFDM and CPP-AFDM significantly outperform corresponding OFDM systems

3. Detector Performance

• Proposed low-complexity MMSE detector achieves identical BER performance to conventional MMSE detector
• Proposed MRC-TD detector achieves nearly identical BER performance to conventional MMSE detector

Complexity Analysis Results

Complexity comparisons demonstrate that the proposed low-complexity MMSE and MRC-TD detectors provide significant complexity reduction while maintaining comparable BER performance to conventional MMSE detectors.

Experimental Findings

1. Increasing path count $P$ improves BER performance due to enhanced multipath diversity
2. ZP-AFDM exhibits power efficiency advantages over CPP-AFDM
3. Proposed detectors successfully exploit the structural characteristics of ZP-assisted channel matrices

Related Work

Current State of AFDM Research

Existing AFDM research primarily includes:

• Pilot-assisted channel estimation
• Non-orthogonal AFDM
• AFDM-aided integrated sensing and communication
• Integration schemes with sparse code multiple access, index modulation, and generalized spatial modulation

Development of ZP Technology

Zero-padding has been widely adopted in OFDM as an alternative to cyclic prefix, offering:

• Higher power transmission efficiency
• Reliable symbol recovery under deep channel fading
• Enhanced channel estimation and tracking capabilities

Novelty of This Work

This paper is the first to introduce ZP into AFDM systems, filling a gap in the field.

Conclusions and Discussion

Main Conclusions

1. Successfully designed a general transceiver architecture for ZP-AFDM systems
2. Developed two low-complexity detectors exploiting sparse lower triangular structure of time-domain channel matrices
3. Derived closed-form BER expressions under ML and MMSE detection
4. Verified the effectiveness and superiority of proposed methods

Limitations

1. Assumes perfect channel state information
2. Simulations primarily based on specific channel models and parameter settings
3. Does not consider practical hardware implementation constraints

Future Directions

1. Joint design of channel estimation and detection
2. Robust detection under imperfect channel information
3. Optimization for hardware implementation

In-Depth Evaluation

Strengths

1. Technical Innovation: First to introduce ZP technology into AFDM systems, with pioneering significance
2. Theoretical Completeness: Provides comprehensive theoretical analysis framework including BER derivation and complexity analysis
3. Practical Value: Proposed low-complexity detectors significantly reduce computational complexity while maintaining performance
4. Sufficient Verification: Extensive simulations validate the correctness of theoretical analysis

Weaknesses

1. Application Scope: Primarily focuses on specific channel models; adaptability to practical channels requires further verification
2. Comparison Baselines: Lacks comprehensive comparison with other advanced detection algorithms
3. Hardware Considerations: Does not address complexity and constraints of practical hardware implementation

Impact

1. Academic Contribution: Provides new insights and theoretical foundation for AFDM system design
2. Practical Value: Offers important reference for next-generation high-mobility communication systems
3. Reproducibility: Detailed algorithm descriptions facilitate reproduction and further research

Applicable Scenarios

• High-speed mobile communication systems (high-speed trains, satellite communications)
• Vehicular communication systems
• Low-earth orbit satellite networks
• Other wireless communication scenarios requiring Doppler resilience

References

The paper cites 16 relevant references covering fundamental AFDM theory, ZP technology, detection algorithm design, and other key areas, providing a solid theoretical foundation for the research.