SSShunqiao Sun

Shunqiao Sun

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Shunqiao Sun has a total of 29 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • 2D DOA Estimation of Coherent Signals Exploiting Forward-Backward Covariance Tensor

    by Saidur R. Pavel, Yimin D. Zhang, Shunqiao Sun

    Coherent signals commonly arise in applications such as wireless communication and radar sensing, necessitating decorrelation before applying conventional direction-of-arrival (DOA) estimation methods. For two-dimensional (2D) sensor arrays, the covariance entries of the received signals form a tensor, which is rank-deficient when the impinging signals are coherent. This paper develops two novel decorrelation strategies to address the rank deficiency issue in 2D DOA estimation. The first strategy reconstructs a decorrelated tensor by rearranging a slice of the coherent covariance tensor and, to enhance dimensionality and increase the degrees of freedom (DOFs) of the array, integrating a backward slice. The second strategy employs forward backward spatial smoothing (FBSS) with an optimal sensor arrangement. In particular, we optimize four FBSS parameters: the numbers of overlapping subarrays along the X and Y axes, and the numbers of antennas in each subarray along the X and Y axes, in order to maximize the DOFs for a given array size. Furthermore, we derive a closed-form expression for the number of achievable DOFs. Simulation results confirm that the proposed methods outperform existing techniques in terms of DOF enhancement and DOA estimation accuracy.


  • Open Access

    Noninvasive Microwave Imaging and Machine Learning to Assess Ripeness of Avocados

    by M. Wu, Z. Choffin, L. Kong, H. Lee, S. Sun, Nathan Jeong

    The demand for precise noninvasive methods to assess fruit ripeness has driven advancements in microwaveimaging (MWI) and machine learning. Determining avocado ripeness is particularly challenging due to its progressive ripening process and the indistinct external indicators at certain stages. This study presents a novel approach combining noninvasive microwave imaging with machine learning (ML) algorithms to classify avocado ripeness stages. The microwave imaging system employs ten ultrawideband Vivaldi antennas in a circular array with a high-speed automated switching network, enabling a comprehensive 360$^{\circ }$ field of view to image an avocado. Internal changes correlated with ripeness are captured through microwave images and analyzed using a convolutional neural network (CNN), achieving a classification accuracy of 89%. K-fold cross-validation further validates the model's reliability, yielding an average accuracy of 90%. This scalable and nondestructive method shows great potential for agricultural applications, ensuring high-quality and safe avocados.


  • Model-Based Learning for DOA Estimation With One-Bit Single-Snapshot Sparse Arrays

    by Yunqiao Hu, Shunqiao Sun, Yimin D. Zhang

    We address the challenging problem of estimating the directions-of-arrival (DOAs) of multiple off-grid signals using a single snapshot of one-bit quantized measurements. Conventional DOA estimation methods face difficulties in tack ling this problem effectively. This paper introduces a domain knowledge-guided learning framework to achieve high-resolution DOA estimation in such a scenario, thus drastically reducing hardware complexity without compromising performance. We f irst reformulate DOA estimation as a maximum a posteriori (MAP) problem, unifying on-grid and off-grid scenarios under a Laplacian-type sparsity prior to effectively enforce sparsity for both uniform and sparse linear arrays. For off-grid signals, a first-order approximation grid model is embedded into the one-bit signal model. We then reinterpret one-bit sensing as a binary classification task, employing a multivariate Bernoulli likelihood with a logistic link function to enhance stability and estimation accuracy. To resolve the non-convexity inherent in the MAP formulation, we develop augmented algorithmic frameworks based on majorization-minimization principles. Fur ther, we design model-based inference neural networks by deep unrolling these frameworks, significantly reducing computational complexity while preserving the estimation precision. Extensive simulations demonstrate the robustness of the proposed frame work across a wide range of input signal-to-noise ratio values and off-grid deviations. By integrating the unified model-based priors with data-driven learning, this work bridges the gap between theoretical guarantees and practical feasibility in one-bit single snapshot DOA estimation, offering a scalable, hardware-efficient solution for next-generation radar and communication systems.


  • Broad Neural Networks for Sparse Signal Retrieval and Array Interpolation in Automotive Radar

    by Lifan Xu, Shunqiao Sun, Ryan Wu, Binbin Shi, Jun Li

    Sparse antenna arrays are widely used in automotive radar systems to achieve high angular resolution with reduced hardware complexity and mutual coupling. However, the high sidelobes inherent in sparse configurations can degrade angle estimation accuracy, motivating the need for effective interpolation methods. Although deep neural networks (DNNs) have shown strong performance in sparse signal recovery, their deployment is hindered by high computational cost, complex activation functions, and limited interpretability due to deep-layer architectures. To address these challenges, we propose a novel framework based on broad neural networks (BNNs) for efficient and accurate sparse signal retrieval. Unlike conventional DNNs, BNNs utilize parallel network structures with simplified, custom-designed activation functions, eliminating hidden layers to reduce computational overhead and improve transparency. We further enhance this design with an iterative BNN approach that incorporates a scalar mask and a phase-enhanced layer, enabling high-accuracy data recovery with fewer iterations. The experimental results on simulated and real-world radar datasets demonstrate that the proposed BNN framework significantly outperforms baseline methods in angle spectrum estimation and data interpolation while maintaining low computational complexity. These findings highlight the potential of BNNs as a practical, interpretable, and scalable alternative for advanced automotive radar applications.


  • Model-Based Knowledge-Driven Learning Approach for Enhanced High-Resolution Automotive Radar Imaging

    by Ruxin Zheng, Shunqiao Sun, Hongshan Liu, Honglei Chen, Jian Li

    Millimeter-wave (mmWave) radars are indispensable for perception tasks of autonomous vehicles, thanks to their resilience in challenging weather and light conditions. Yet, their deployment is often limited by insufficient spatial resolution for precise semantic scene interpretation. Classical super-resolution techniques adapted from optical imaging inadequately address the distinct characteristics of radar data. In response, our study herein redefines super-resolution radar imaging as a one-dimensional (1D) signal super-resolution spectra estimation problem by harnessing the radar domain knowledge, introducing innovative data normalization, signal-level augmentation, and a domain-informed signal-to-noise ratio (SNR)-guided loss function. Like an image drawn with points and lines, radar imaging can be viewed as generated from points (antenna elements) and lines (frequency spectra). Our tailored deep learning network for automotive radar imaging exhibits remarkable scalability and parameter efficiency, alongside enhanced performance in terms of radar imaging quality and resolution. We further present a novel real-world dataset, pivotal for both advancing radar imaging and refining super-resolution spectra estimation techniques. Extensive testing confirms that our SR-SPECNet sets a new benchmark in producing high-resolution radar range-azimuth images, outperforming existing methods. The source code and radar dataset utilized for evaluation will be made publicly available at https://github.com/ruxinzh/SR-SPECNet.


  • Open Access

    Reconfigurable Beamforming for Automotive Radar Sensing and Communication: A Deep Reinforcement Learning Approach

    by Lifan Xu, Shunqiao Sun, Yimin D. Zhang, Athina P. Petropulu

    In this paper, we present a dual-function radar-communication (DFRC) system for automotive applications, that uses sparse phased arrays equipped with quantized double-phase shifters, thus enabling a low-hardware cost structure. To address the operation in dynamic environments, we propose a deep reinforcement learning (DRL) approach to adaptively select a small subset of transmit antennas and adjust quantized double-phase shifters, such that the transmitted energy is concentrated on the communication user and target of interest, while reducing interference to other radars by creating nulls towards their respective directions. The action space in the DRL approach increases fast with the number of antennas and the number of bits used in quantization, and as a result the complexity of the design problem grows exponentially. To tackle the resulting curse of dimensionality in the action space, we adopt the Wolpertinger strategy, which incorporates the nearest neighborhood component to project the vast action space into a smaller, more manageable space while maintaining the desired performance. Numerical results demonstrate the feasibility of our proposed method.


  • Widely Separated MIMO Radar Using Matrix Completion

    by Shunqiao Sun, Yunqiao Hu, Kumar Vijay Mishra, Athina P. Petropulu

    We present a low-complexity widely separated multiple-input-multiple-output (WS-MIMO) radar that samples the signals at each of its multiple receivers at reduced rates. We process the low-rate samples of all transmit-receive chains at each receiver as data matrices. We demonstrate that each of these matrices is low rank as long as the target moves slowly within a coherent processing interval. We leverage matrix completion (MC) to recover the missing samples of each receiver signal matrix at the common fusion center. Subsequently, we estimate the targets’ positions and Doppler velocities via the maximum likelihood method. Our MC-WS-MIMO approach recovers missing samples and thereafter target parameters at reduced rates without discretization. Our analysis using ambiguity functions shows that antenna geometry affects the performance of MC-WS-MIMO. Numerical experiments demonstrate reasonably accurate target localization at SNR of 20 dB and sampling rate reduction to 20%.


  • Interpretable and Efficient Beamforming-Based Deep Learning for Single-Snapshot DOA Estimation

    by Ruxin Zheng, Shunqiao Sun, Hongshan Liu, Honglei Chen, Jian Li

    We introduce an interpretable deep learning approach for direction of arrival (DOA) estimation with a single snapshot. Classical subspace-based methods like MUSIC and ESPRIT use spatial smoothing on uniform linear arrays for single snapshot DOA estimation but face drawbacks in reduced array aperture and inapplicability to sparse arrays. Single-snapshot methods such as compressive sensing and iterative adaptation approach (IAA) encounter challenges with high computational costs and slow convergence, hampering real-time use. Recent deep learning DOA methods offer promising accuracy and speed. However, the practical deployment of deep networks is hindered by their black-box nature. To address this, we propose a deep-MPDR network translating minimum power distortionless response (MPDR)-type beamformer into deep learning, enhancing generalization and efficiency. Comprehensive experiments conducted using both simulated and real-world datasets substantiate its dominance in terms of inference time and accuracy in comparison to conventional methods. Moreover, it excels in terms of efficiency, generalizability, and interpretability when contrasted with other deep learning DOA estimation networks.


  • Automotive FMCW Radar With Difference Co-Chirps

    by Lifan Xu, Shunqiao Sun, Kumar Vijay Mishra, Yimin D. Zhang

    In an automotive radar scenario, transmission of sparsely spaced chirps along slow-time creates opportunities to significantly reduce or completely avoid mutual interference. However, non-uniform chirps result in high Doppler sidelobes which may introduce ambiguity in Doppler estimation. We present an automotive frequency-modulated continuous-wave radar that exploits difference co-chirps to achieve high-accuracy range-Doppler estimation with low complexity. By exploiting sparsity in slow-time , the proposed method achieves the same Doppler velocity resolution as the regular uniform pulsing but with much fewer chirps. The silent transmission intervals become available to other automotive radars, thereby eliminating the mutual interference. The fast-time samples are treated as snapshots to construct the second-order statistical information for Doppler spectrum estimation. For our proposed non-uniform pulsing, we develop an efficient range-Doppler spectrum estimation method and propose a matching technique that is based on 2-D compressed sensing followed by Doppler de-aliasing. This algorithm achieves ipso facto range-Doppler pairing without grid mismatch errors in parameter estimation and does not require an exhaustive search. Extensive numerical experiments show that accurate range-Doppler estimation is achieved with significantly fewer chirps compared to the conventional consecutive transmission. Field campaigns using Texas Instruments imaging radar support our theoretical investigations.