Design of Optimized Coded LFM Waveform for Spectrum Shared Radar System

Dong-Hoon Kim, Hyung-Jung Kim, Jae-Han LimSensors 2021 · DOI

TL;DR

This paper introduces a novel optimization framework for designing coded Linear Frequency Modulation (LFM) waveforms that significantly improve orthogonality in Spectrum Shared Radar Systems (SSRS). By directly optimizing the coded LFM waveforms themselves, rather than just the underlying codes, the authors achieve superior performance in terms of autocorrelation sidelobe peaks (ASP) and cross-correlation peaks (CP), crucial for minimizing interference in multi-radar environments.

What Problem Does This Paper Address?

Modern radar systems are increasingly demanding, leading to a scarcity of available frequency resources. Spectrum Shared Radar Systems (SSRS) offer a solution by allowing multiple radars to operate within the same frequency band. However, this proximity introduces the challenge of signal interference between radars. To mitigate this, SSRS typically relies on orthogonal waveforms. While Linear Frequency Modulation (LFM) is a popular choice for its hardware simplicity and good range resolution, incorporating orthogonal codes into LFM waveforms often degrades their orthogonality, especially when the codes are multiplied with rapidly changing LFM signals. Existing methods that first optimize orthogonal codes and then apply them to LFM waveforms struggle to maintain this orthogonality. This paper addresses the need for a method that generates highly orthogonal coded LFM waveforms directly, maximizing their suitability for SSRS.

Methods

The core of this paper is a new optimization framework designed to generate a set of optimized coded LFM waveforms. Unlike previous approaches that optimize orthogonal codes first and then combine them with LFM signals, this method directly optimizes the coded LFM waveforms. The optimization process aims to minimize a weighted sum of Autocorrelation Sidelobe Peaks (ASP) and Cross-Correlation Peaks (CP), effectively maximizing the orthogonality of the waveforms.

The process involves:

  1. Generating an LFM waveform set (X): This can include LFM waveforms with varying chirp rates.
  2. Generating a random polyphase code set (S): Polyphase codes are chosen for their flexibility and higher degree of freedom compared to binary codes.
  3. Generating coded LFM waveforms (Y): Each coded LFM waveform is created by multiplying an LFM waveform with a corresponding code element.
  4. Optimization using Simulated Annealing (SA): The framework employs SA to iteratively adjust the phase elements within the polyphase code set (S) to find an optimal set (Sopt) that minimizes the objective function (weighted sum of ASP and CP). This iterative process allows for the direct optimization of the coded LFM waveforms (Yopt).

The objective function is defined as $Obj(Y) = (1 - \lambda)ASP(Y) + \lambda CP(Y)$, where $\lambda$ is a weighting factor that allows for prioritizing the reduction of CP over ASP.

Datasets

Not discussed in the paper. The evaluation relies on computer simulations using MATLAB.

Key Insights and Ideas

  • Direct Optimization of Coded Waveforms: The most significant insight is the shift from optimizing codes to optimizing the final coded LFM waveforms. This direct approach proves more effective in preserving orthogonality.
  • Polyphase Codes for Flexibility: The use of polyphase codes, with their multiple phase options, provides a richer search space for optimization compared to simpler binary codes, leading to better orthogonality.
  • Trade-off between ASP and CP: The weighting factor $\lambda$ allows for a tunable trade-off between minimizing autocorrelation sidelobe peaks (ASP) and cross-correlation peaks (CP), enabling customization for specific radar requirements.
  • Code Length is Crucial: The paper highlights that code length (N) is a dominant factor in achieving orthogonality. Longer code lengths exponentially increase the degrees of freedom, allowing for the selection of more orthogonal waveform sets.
  • Number of Radars Impacts Orthogonality: As the number of radars (L) increases, maintaining high orthogonality becomes more challenging, as it's harder to find a set of waveforms that are mutually orthogonal for all pairs.

Key Results

The proposed optimized coded LFM waveforms demonstrated superior performance compared to existing methods like SSCL-Hadamard and SSCL-Polyphase. Specifically:

  • Improved Orthogonality: The proposed scheme achieved lower ASP and CP values. For instance, the paper reports improvements of up to 11.652 dB in ASP and 17.560 dB in CP compared to previous proposals.
  • Performance Gains with Parameters:
    • Increasing the number of distinct phases (M) generally improved orthogonality.
  • Increasing the code length (N) significantly decreased CP and improved ASP, indicating its critical role.
  • The weighting factor ($\lambda$) allowed for tuning the trade-off: higher $\lambda$ reduced CP at the expense of ASP.
  • Frequency Efficiency: The proposed scheme is noted for its better frequency efficiency, supporting more radars with a narrower bandwidth compared to previous methods.
  • Range-Doppler Performance: While the ambiguity function of the proposed waveform showed slightly higher non-zero delay returns (potentially reducing range accuracy for closely spaced targets), the Doppler ambiguity function was comparable to a single LFM waveform, indicating minimal degradation in Doppler measurement accuracy.

Limitations and Future Work

  • Computational Complexity: The optimization process, especially with large numbers of distinct phases (M) and code lengths (N), can lead to high computational complexity.
  • Vulnerability to Noise: Coded LFM waveforms with a large number of distinct phases (M) can be more susceptible to noise.
  • Future Work: The authors suggest that careful selection of design parameters (L, M, N, $\lambda$) is crucial for generating optimal waveform sets. Further analysis of the range and Doppler measurement performance of the optimized waveforms compared to single LFM waveforms is also suggested, along with exploring post-processing algorithms (like CLEAN) to mitigate potential range accuracy issues.

Summary updated

BibTeX
@article{kim2021design,
  author = {Dong-Hoon Kim and Hyung-Jung Kim and Jae-Han Lim},
  title = {{Design of Optimized Coded LFM Waveform for Spectrum Shared Radar System}},
  journal = {Sensors},
  year = {2021},
  doi = {10.3390/s21175796},
}

Loading related content...

Figure