Lu Gan

Lu Gan, Connor Lee, Soon-Jo Chung

This work presents a new method for unsupervised thermal image classification and semantic segmentation by transferring knowledge from the RGB domain using a multi-domain attention network. Our method does not require any thermal annotations or co-registered RGB-thermal pairs, enabling robots to perform visual tasks at night and in adverse weather conditions without incurring additional costs of data labeling and registration. Current unsupervised domain adaptation methods look to align global images or features across domains. However, when the domain shift is significantly larger for cross-modal data, not all features can be transferred. We solve this problem by using a shared backbone network that promotes generalization, and domain-specific attention that reduces negative transfer by attending to domain-invariant and easily-transferable features. Our approach outperforms the state-of-the-art RGB-to-thermal adaptation method in classification benchmarks, and is successfully applied to thermal river scene segmentation using only synthetic RGB images. Our code is made publicly available at https://github.com/ganlumomo/thermal-uda-attention.

Connor Lee, Jonathan Gustafsson Frennert, Lu Gan et al.

We present a new method to adapt an RGB-trained water segmentation network to target-domain aerial thermal imagery using online self-supervision by leveraging texture and motion cues as supervisory signals. This new thermal capability enables current autonomous aerial robots operating in near-shore environments to perform tasks such as visual navigation, bathymetry, and flow tracking at night. Our method overcomes the problem of scarce and difficult-to-obtain near-shore thermal data that prevents the application of conventional supervised and unsupervised methods. In this work, we curate the first aerial thermal near-shore dataset, show that our approach outperforms fully-supervised segmentation models trained on limited target-domain thermal data, and demonstrate real-time capabilities onboard an Nvidia Jetson embedded computing platform. Code and datasets used in this work will be available at: https://github.com/connorlee77/uav-thermal-water-segmentation.

Wangyang Xu, Jiancheng An, Chongwen Huang et al.

Reconfigurable intelligent surface (RIS) has recently gained popularity as a promising solution for improving the signal transmission quality of wireless communications with less hardware cost and energy consumption. This letter offers a novel deep reinforcement learning (DRL) algorithm based on a location-aware imitation environment for the joint beamforming design in an RIS-aided mmWave multiple-input multiple-output system. Specifically, we design a neural network to imitate the transmission environment based on the geometric relationship between the user's location and the mmWave channel. Following this, a novel DRL-based method is developed that interacts with the imitation environment using the easily available location information. Finally, simulation results demonstrate that the proposed DRL-based algorithm provides more robust performance without excessive interaction overhead compared to the existing DRL-based approaches.

Sri Aditya Deevi, Connor Lee, Lu Gan et al.

Multimodal deep sensor fusion has the potential to enable autonomous vehicles to visually understand their surrounding environments in all weather conditions. However, existing deep sensor fusion methods usually employ convoluted architectures with intermingled multimodal features, requiring large coregistered multimodal datasets for training. In this work, we present an efficient and modular RGB-X fusion network that can leverage and fuse pretrained single-modal models via scene-specific fusion modules, thereby enabling joint input-adaptive network architectures to be created using small, coregistered multimodal datasets. Our experiments demonstrate the superiority of our method compared to existing works on RGB-thermal and RGB-gated datasets, performing fusion using only a small amount of additional parameters. Our code is available at https://github.com/dsriaditya999/RGBXFusion.

Weidong Wang, Cheng Luo, Jiancheng An et al.

As a promising non-password authentication technology, radio frequency (RF) fingerprinting can greatly improve wireless security. Recent work has shown that RF fingerprinting based on deep learning can significantly outperform conventional approaches. The superiority, however, is mainly attributed to supervised learning using a large amount of labeled data, and it significantly degrades if only limited labeled data is available, making many existing algorithms lack practicability. Considering that it is often easier to obtain enough unlabeled data in practice with minimal resources, we leverage deep semi-supervised learning for RF fingerprinting, which largely relies on a composite data augmentation scheme designed for radio signals, combined with two popular techniques: consistency-based regularization and pseudo-labeling. Experimental results on both simulated and real-world datasets demonstrate that our proposed method for semi-supervised RF fingerprinting is far superior to other competing ones, and it can achieve remarkable performance almost close to that of fully supervised learning with a very limited number of examples.

Jinyue Jiang, Jiancheng An, Lu Gan et al.

Flexible intelligent metasurface (FIM) has recently received considerable interest due to its advantage in realizing a better channel condition by dynamically morphing its surface shape. An FIM consists of multiple elements deposited on a flexible substrate. These elements can not only transmit signals, but also adapt their displacements in a direction perpendicular to the FIM surface via an attached controller. In this paper, we consider the channel estimation problem for the uplink of an FIM-enhanced communication system via customizing the orthogonal matching pursuit (OMP) method. Specifically, we formulate an optimization problem of minimizing the column coherence of the measurement matrix by optimizing the FIM's surface shape, subject to the morphing range constraint. Based on the estimated direction of arrival (DOA) and channel gain, we further investigate the signal-to-noise ratio (SNR) improvement in the FIM-enhanced downlink multiple-input single-output (MISO) system. Numerical results demonstrate that an FIM significantly outperforms a conventional rigid uniform planar array (UPA), thereby showing that FIM can substantially improve channel estimation accuracy and achieve SNR improvement, even when using estimated channel parameters.

Weidong Wang, Jiancheng An, Hongshu Liao et al.

As a revolutionary generative paradigm of deep learning, generative adversarial networks (GANs) have been widely applied in various fields to synthesize realistic data. However, it is challenging for conventional GANs to synthesize raw signal data, especially in some complex cases. In this paper, we develop a novel GAN framework for radio generation called "Radio GAN". Compared to conventional methods, it benefits from three key improvements. The first is learning based on sampling points, which aims to model an underlying sampling distribution of radio signals. The second is an unrolled generator design, combined with an estimated pure signal distribution as a prior, which can greatly reduce learning difficulty and effectively improve learning precision. Finally, we present an energy-constrained optimization algorithm to achieve better training stability and convergence. Experimental results with extensive simulations demonstrate that our proposed GAN framework can effectively learn transmitter characteristics and various channel effects, thus accurately modeling for an underlying sampling distribution to synthesize radio signals of high quality.

Chen Zhan, Xihe Qiu, Xiaoyu Tan et al.

Large language models perform well on static medical examinations, yet clinical diagnosis often requires iterative evidence gathering under uncertainty. Building on prior interactive evaluation efforts, we introduce an OSCE-inspired standardized patient simulator and a controlled, reproducible benchmark for active diagnostic inquiry. Across 468 cases and 15 models in our protocol, we observe that multi-turn evidence seeking reduces diagnostic accuracy by 12.75% and lowers supporting-evidence quality by 24.36% relative to full-context evaluation; error analyses associate these drops with premature diagnostic closure and inefficient questioning. Together, these results suggest that static full-context benchmarks may overestimate performance in interactive evidence-seeking settings, motivating complementary interactive assessment for safer clinical decision support.

Weidong Wang, Hongshu Liao, Lu Gan

Deep learning has been widely used in radio frequency (RF) fingerprinting. Despite its excellent performance, most existing methods only consider a closed-set assumption, which cannot effectively tackle signals emitted from those unknown devices that have never been seen during training. In this letter, we exploit prototype learning for open-set RF fingerprinting and propose two improvements, including consistency-based regularization and online label smoothing, which aim to learn a more robust feature space. Experimental results on a real-world RF dataset demonstrate that our proposed measures can significantly improve prototype learning to achieve promising open-set recognition performance for RF fingerprinting.

Wolfgang Otto, Matthäus Zloch, Lu Gan et al.

Named Entity Recognition (NER) models play a crucial role in various NLP tasks, including information extraction (IE) and text understanding. In academic writing, references to machine learning models and datasets are fundamental components of various computer science publications and necessitate accurate models for identification. Despite the advancements in NER, existing ground truth datasets do not treat fine-grained types like ML model and model architecture as separate entity types, and consequently, baseline models cannot recognize them as such. In this paper, we release a corpus of 100 manually annotated full-text scientific publications and a first baseline model for 10 entity types centered around ML models and datasets. In order to provide a nuanced understanding of how ML models and datasets are mentioned and utilized, our dataset also contains annotations for informal mentions like "our BERT-based model" or "an image CNN". You can find the ground truth dataset and code to replicate model training at https://data.gesis.org/gsap/gsap-ner.

Xihe Qiu, Yang Dai, Xiaoyu Tan et al.

Magnetic Resonance Imaging (MRI) provides detailed tissue information, but its clinical application is limited by long acquisition time, high cost, and restricted resolution. Image translation has recently gained attention as a strategy to address these limitations. Although Pix2Pix has been widely applied in medical image translation, its potential has not been fully explored. In this study, we propose an enhanced Pix2Pix framework that integrates Squeeze-and-Excitation Residual Networks (SEResNet) and U-Net++ to improve image generation quality and structural fidelity. SEResNet strengthens critical feature representation through channel attention, while U-Net++ enhances multi-scale feature fusion. A simplified PatchGAN discriminator further stabilizes training and refines local anatomical realism. Experimental results demonstrate that under few-shot conditions with fewer than 500 images, the proposed method achieves consistent structural fidelity and superior image quality across multiple intra-modality MRI translation tasks, showing strong generalization ability. These results suggest an effective extension of Pix2Pix for medical image translation.

Fengze Xie, Sizhe Wei, Yue Song et al. · gatech

We present a morphological-symmetry-equivariant heterogeneous graph neural network, namely MS-HGNN, for robotic dynamics learning, that integrates robotic kinematic structures and morphological symmetries into a single graph network. These structural priors are embedded into the learning architecture as constraints, ensuring high generalizability, sample and model efficiency. The proposed MS-HGNN is a versatile and general architecture that is applicable to various multi-body dynamic systems and a wide range of dynamics learning problems. We formally prove the morphological-symmetry-equivariant property of our MS-HGNN and validate its effectiveness across multiple quadruped robot learning problems using both real-world and simulated data. Our code is made publicly available at https://github.com/lunarlab-gatech/MorphSym-HGNN/.

Zhenyu Zong, Yuchen Wang, Haohong Lin et al.

Trajectory prediction for traffic agents is critical for safe autonomous driving. However, achieving effective zero-shot generalization in previously unseen domains remains a significant challenge. Motivated by the consistent nature of kinematics across diverse domains, we aim to incorporate domain-invariant knowledge to enhance zero-shot trajectory prediction capabilities. The key challenges include: 1) effectively extracting domain-invariant scene representations, and 2) integrating invariant features with kinematic models to enable generalized predictions. To address these challenges, we propose a novel generalizable Physics-guided Causal Model (PCM), which comprises two core components: a Disentangled Scene Encoder, which adopts intervention-based disentanglement to extract domain-invariant features from scenes, and a CausalODE Decoder, which employs a causal attention mechanism to effectively integrate kinematic models with meaningful contextual information. Extensive experiments on real-world autonomous driving datasets demonstrate our method's superior zero-shot generalization performance in unseen cities, significantly outperforming competitive baselines. The source code is released at https://github.com/ZY-Zong/Physics-guided-Causal-Model.

Chenhao Shuai, Chaohua Shi, Lu Gan et al.

Speech super-resolution (SSR) aims to recover a high resolution (HR) speech from its corresponding low resolution (LR) counterpart. Recent SSR methods focus more on the reconstruction of the magnitude spectrogram, ignoring the importance of phase reconstruction, thereby limiting the recovery quality. To address this issue, we propose mdctGAN, a novel SSR framework based on modified discrete cosine transform (MDCT). By adversarial learning in the MDCT domain, our method reconstructs HR speeches in a phase-aware manner without vocoders or additional post-processing. Furthermore, by learning frequency consistent features with self-attentive mechanism, mdctGAN guarantees a high quality speech reconstruction. For VCTK corpus dataset, the experiment results show that our model produces natural auditory quality with high MOS and PESQ scores. It also achieves the state-of-the-art log-spectral-distance (LSD) performance on 48 kHz target resolution from various input rates. Code is available from https://github.com/neoncloud/mdctGAN

Yuchen Wang, Jiangtao Kong, Sizhe Wei et al.

Trajectory world models play a crucial role in robotic dynamics learning, planning, and control. While recent works have explored trajectory world models for diverse robotic systems, they struggle to scale to a large number of distinct system dynamics and overlook domain knowledge of physical structures. To address these limitations, we introduce WestWorld, a knoWledge-Encoded Scalable Trajectory World model for diverse robotic systems. To tackle the scalability challenge, we propose a novel system-aware Mixture-of-Experts (Sys-MoE) that dynamically combines and routes specialized experts for different robotic systems via a learnable system embedding. To further enhance zero-shot generalization, we incorporate domain knowledge of robot physical structures by introducing a structural embedding that aligns trajectory representations with morphological information. After pretraining on 89 complex environments spanning diverse morphologies across both simulation and real-world settings, WestWorld achieves significant improvements over competitive baselines in zero- and few-shot trajectory prediction. Additionally, it shows strong scalability across a wide range of robotic environments and significantly improves performance on downstream model-based control for different robots. Finally, we deploy our model on a real-world Unitree Go1, where it demonstrates stable locomotion performance (see our demo on the website: https://westworldrobot.github.io/). The code will be available upon publication.

Wolfgang Otto, Lu Gan, Sharmila Upadhyaya et al.

Research in Machine Learning (ML) and AI evolves rapidly. Information Extraction (IE) from scientific publications enables to identify information about research concepts and resources on a large scale and therefore is a pathway to improve understanding and reproducibility of ML-related research. To extract and connect fine-grained information in ML-related research, e.g. method training and data usage, we introduce GSAP-ERE. It is a manually curated fine-grained dataset with 10 entity types and 18 semantically categorized relation types, containing mentions of 63K entities and 35K relations from the full text of 100 ML publications. We show that our dataset enables fine-tuned models to automatically extract information relevant for downstream tasks ranging from knowledge graph (KG) construction, to monitoring the computational reproducibility of AI research at scale. Additionally, we use our dataset as a test suite to explore prompting strategies for IE using Large Language Models (LLM). We observe that the performance of state-of-the-art LLM prompting methods is largely outperformed by our best fine-tuned baseline model (NER: 80.6%, RE: 54.0% for the fine-tuned model vs. NER: 44.4%, RE: 10.1% for the LLM). This disparity of performance between supervised models and unsupervised usage of LLMs suggests datasets like GSAP-ERE are needed to advance research in the domain of scholarly information extraction.

Lu Gan, Xi Li, Xichun Wang

We present TXL-PBC, a curated and re-annotated peripheral blood cell dataset constructed by integrating four publicly available resources: Blood Cell Count and Detection (BCCD), Blood Cell Detection Dataset (BCDD), Peripheral Blood Cells (PBC), and Raabin White Blood Cell (Raabin-WBC). Through rigorous sample selection, semi-automatic annotation using the YOLOv8n model, and comprehensive manual review, we ensured high annotation accuracy and consistency. The final dataset contains 1,260 images and 18,143 bounding box annotations for three major blood cell types: white blood cells (WBC), red blood cells (RBC), and platelets. We provide detailed visual analyses of the data distribution, demonstrating the diversity and balance of the dataset. To further validate the quality and utility of TXL-PBC, we trained several mainstream object detection models, including YOLOv5s, YOLOv8s, YOLOv11s, SSD300, Faster R-CNN, and RetinaNet, and report their baseline performance. The TXL-PBC dataset is openly available on Figshare and GitHub, offering a valuable resource for the development and benchmarking of blood cell detection models and related machine learning research.

Lingjun Zhao, Yandong Luo, James Hay et al.

We introduce ShelfGaussian, an open-vocabulary multi-modal Gaussian-based 3D scene understanding framework supervised by off-the-shelf vision foundation models (VFMs). Gaussian-based methods have demonstrated superior performance and computational efficiency across a wide range of scene understanding tasks. However, existing methods either model objects as closed-set semantic Gaussians supervised by annotated 3D labels, neglecting their rendering ability, or learn open-set Gaussian representations via purely 2D self-supervision, leading to degraded geometry and limited to camera-only settings. To fully exploit the potential of Gaussians, we propose a Multi-Modal Gaussian Transformer that enables Gaussians to query features from diverse sensor modalities, and a Shelf-Supervised Learning Paradigm that efficiently optimizes Gaussians with VFM features jointly at 2D image and 3D scene levels. We evaluate ShelfGaussian on various perception and planning tasks. Experiments on Occ3D-nuScenes demonstrate its state-of-the-art zero-shot semantic occupancy prediction performance. ShelfGaussian is further evaluated on an unmanned ground vehicle (UGV) to assess its in the-wild performance across diverse urban scenarios. Project website: https://lunarlab-gatech.github.io/ShelfGaussian/.

Weidong Wang, Hongshu Liao, Lu Gan

Deep neural networks have been widely used in communication signal recognition and achieved remarkable performance, but this superiority typically depends on using massive examples for supervised learning, whereas training a deep neural network on small datasets with few labels generally falls into overfitting, resulting in degenerated performance. To this end, we develop a semi-supervised learning (SSL) method that effectively utilizes a large collection of more readily available unlabeled signal data to improve generalization. The proposed method relies largely on a novel implementation of consistency-based regularization, termed Swapped Prediction, which leverages strong data augmentation to perturb an unlabeled sample and then encourage its corresponding model prediction to be close to its original, optimized with a scaled cross-entropy loss with swapped symmetry. Extensive experiments indicate that our proposed method can achieve a promising result for deep SSL of communication signal recognition.

Jinbao Li, Jiancheng An, Hao Liu et al.

Semantic communications (SemCom) is a promising paradigm that prioritizes the transmission of task-relevant information, thereby enabling superior communication efficiency over traditional bit-centric systems. However, most existing SemCom systems face critical limitations in computational efficiency and spatial flexibility. To overcome these limitations, we propose a novel unmanned aerial vehicles (UAV)-enabled distributed electromagnetic neural network (EMNN) for a task-oriented SemCom system. Specifically, the proposed distributed EMNN is composed of multiple UAV-mounted stacked intelligent metasurfaces (SIM) and a ground receiving station (GRS), where multiple SIMs collaboratively encode image semantics in the wave domain, and the GRS performs decoding based on the received power distribution. Moreover, we employ a temperature-adaptive gradient optimization algorithm to train the distributed EMNN, which mitigates gradient vanishing and enhances learning stability. Finally, the numerical simulation results demonstrate the effectiveness of distributed EMNN in image recognition task-oriented SemCom, achieving an average $8\%$ accuracy improvement over the single-SIM baseline across multiple datasets.

Lingjun Zhao, Sizhe Wei, James Hays et al. · gatech

3D semantic occupancy prediction is critical for achieving safe and reliable autonomous driving. Compared to camera-only perception systems, multi-modal pipelines, especially LiDAR-camera fusion methods, can produce more accurate and detailed predictions. Although most existing works utilize a dense grid-based representation, in which the entire 3D space is uniformly divided into discrete voxels, the emergence of 3D Gaussians provides a compact and continuous object-centric representation. In this work, we propose a multi-modal Gaussian-based semantic occupancy prediction framework utilizing 3D deformable attention, named as GaussianFormer3D. We introduce a voxel-to-Gaussian initialization strategy to provide 3D Gaussians with geometry priors from LiDAR data, and design a LiDAR-guided 3D deformable attention mechanism for refining 3D Gaussians with LiDAR-camera fusion features in a lifted 3D space. We conducted extensive experiments on both on-road and off-road datasets, demonstrating that our GaussianFormer3D achieves high prediction accuracy that is comparable to state-of-the-art multi-modal fusion-based methods with reduced memory consumption and improved efficiency.

Lu Gan, Martin Blum, Danilo Dessi et al.

Named entity recognition is an important task when constructing knowledge bases from unstructured data sources. Whereas entity detection methods mostly rely on extensive training data, Large Language Models (LLMs) have paved the way towards approaches that rely on zero-shot learning (ZSL) or few-shot learning (FSL) by taking advantage of the capabilities LLMs acquired during pretraining. Specifically, in very specialized scenarios where large-scale training data is not available, ZSL / FSL opens new opportunities. This paper follows this recent trend and investigates the potential of leveraging Large Language Models (LLMs) in such scenarios to automatically detect datasets and software within textual content from GitHub repositories. While existing methods focused solely on named entities, this study aims to broaden the scope by incorporating resources such as repositories and online hubs where entities are also represented by URLs. The study explores different FSL prompt learning approaches to enhance the LLMs' ability to identify dataset and software mentions within repository texts. Through analyses of LLM effectiveness and learning strategies, this paper offers insights into the potential of advanced language models for automated entity detection.

Yirong Shen, Lu Gan, Cong Ling

Generative models that maximize model likelihood have gained traction in many practical settings. Among them, perturbation based approaches underpin many strong likelihood estimation models, yet they often face slow convergence and limited theoretical understanding. In this paper, we derive a tighter likelihood bound for noise driven models to improve both the accuracy and efficiency of maximum likelihood learning. Our key insight extends the classical KL divergence Fisher information relationship to arbitrary noise perturbations, going beyond the Gaussian assumption and enabling structured noise distributions. This formulation allows flexible use of randomized noise distributions that naturally account for sensor artifacts, quantization effects, and data distribution smoothing, while remaining compatible with standard diffusion training. Treating the diffusion process as a Gaussian channel, we further express the mismatched entropy between data and model, showing that the proposed objective upper bounds the negative log-likelihood (NLL). In experiments, our models achieve competitive NLL on CIFAR-10 and SOTA results on ImageNet across multiple resolutions, all without data augmentation, and the framework extends naturally to discrete data.

Abigail R. Cohen, Yuming Sun, Zhihao Qin et al.

Efficient nutrient management is critical for crop growth and sustainable resource consumption (e.g., nitrogen, energy). Current approaches require lengthy analyses, preventing real-time optimization; similarly, imaging facilitates rapid phenotyping but can be computationally intensive, preventing deployment under resource constraints. This study proposes a flexible, tiered pipeline for anomaly detection and status estimation (fresh weight, dry mass, and tissue nutrients), including a comprehensive energy analysis of approaches that span the efficiency-accuracy spectrum. Using a nutrient depletion experiment with three treatments (T1-100%, T2-50%, and T3-25% fertilizer strength) and multispectral imaging (MSI), we developed a hierarchical pipeline using an autoencoder (AE) for early warning. Further, we compared two status estimation modules of different complexity for more detailed analysis: vegetation index (VI) features with machine learning (Random Forest, RF) and raw whole-image deep learning (Vision Transformer, ViT). Results demonstrated high-efficiency anomaly detection (73% net detection of T3 samples 9 days after transplanting) at substantially lower energy than embodied energy in wasted nitrogen. The state estimation modules show trade-offs, with ViT outperforming RF on phosphorus and calcium estimation (R2 0.61 vs. 0.58, 0.48 vs. 0.35) at higher energy cost. With our modular pipeline, this work opens opportunities for edge diagnostics and practical opportunities for agricultural sustainability.

Yeonju Lee, Rui Qi Chen, Joseph Oboamah et al.

Accurate interpretation of soil moisture patterns is critical for irrigation scheduling and crop management, yet existing approaches for soil moisture time-series analysis either rely on threshold-based rules or data-hungry machine learning or deep learning models that are limited in adaptability and interpretability. In this study, we introduce SPADE (Soil moisture Pattern and Anomaly DEtection), an integrated framework that leverages large language models (LLMs) to jointly detect irrigation patterns and anomalies in soil moisture time-series data. SPADE utilizes ChatGPT-4.1 for its advanced reasoning and instruction-following capabilities, enabling zero-shot analysis without requiring task-specific annotation or fine-tuning. By converting time-series data into a textual representation and designing domain-informed prompt templates, SPADE identifies irrigation events, estimates net irrigation gains, detects, classifies anomalies, and produces structured, interpretable reports. Experiments were conducted on real-world soil moisture sensor data from commercial and experimental farms cultivating multiple crops across the United States. Results demonstrate that SPADE outperforms the existing method in anomaly detection, achieving higher recall and F1 scores and accurately classifying anomaly types. Furthermore, SPADE achieved high precision and recall in detecting irrigation events, indicating its strong capability to capture irrigation patterns accurately. SPADE's reports provide interpretability and usability of soil moisture analytics. This study highlights the potential of LLMs as scalable, adaptable tools for precision agriculture, which is capable of integrating qualitative knowledge and data-driven reasoning to produce actionable insights for accurate soil moisture monitoring and improved irrigation scheduling from soil moisture time-series data.

Sunil Kumar Narayanan, Lingjun Zhao, Lu Gan et al.

Hyperspectral imaging (HSI) has been widely used in agricultural applications for non-destructive estimation of plant nutrient composition and precise determination of nutritional elements of samples. Recently, 3D reconstruction methods have been used to create implicit neural representations of HSI scenes, which can help localize the target object's nutrient composition spatially and spectrally. Neural Radiance Field (NeRF) is a cutting-edge implicit representation that can be used to render hyperspectral channel compositions of each spatial location from any viewing direction. However, it faces limitations in training time and rendering speed. In this paper, we propose Diffusion-Denoised Hyperspectral Gaussian Splatting (DD-HGS), which enhances the state-of-the-art 3D Gaussian Splatting (3DGS) method with wavelength-aware spherical harmonics, a Kullback-Leibler divergence-based spectral loss, and a diffusion-based denoiser to enable 3D explicit reconstruction of hyperspectral scenes across the full spectral range. We present extensive evaluations on diverse real-world hyperspectral scenes from the Hyper-NeRF dataset to show the effectiveness of DD-HGS. The results demonstrate that DD-HGS achieves new state-of-the-art performance among previously published methods. Project page: https://dragonpg2000.github.io/DDHGS-website/

Zhenhua Xu, Yuxuan Liu, Lu Gan et al.

Road network graphs provide critical information for autonomous-vehicle applications, such as drivable areas that can be used for motion planning algorithms. To find road network graphs, manually annotation is usually inefficient and labor-intensive. Automatically detecting road network graphs could alleviate this issue, but existing works still have some limitations. For example, segmentation-based approaches could not ensure satisfactory topology correctness, and graph-based approaches could not present precise enough detection results. To provide a solution to these problems, we propose a novel approach based on transformer and imitation learning in this paper. In view of that high-resolution aerial images could be easily accessed all over the world nowadays, we make use of aerial images in our approach. Taken as input an aerial image, our approach iteratively generates road network graphs vertex-by-vertex. Our approach can handle complicated intersection points with various numbers of incident road segments. We evaluate our approach on a publicly available dataset. The superiority of our approach is demonstrated through the comparative experiments. Our work is accompanied with a demonstration video which is available at \url{https://tonyxuqaq.github.io/projects/RNGDet/}.

Wangyang Xu, Lu Gan, Chongwen Huang

Reconfigurable intelligent surface (RIS) has become a promising technology to improve wireless communication in recent years. It steers the incident signals to create a favorable propagation environment by controlling the reconfigurable passive elements with less hardware cost and lower power consumption. In this paper, we consider a RIS-aided multiuser multiple-input single-output downlink communication system. We aim to maximize the weighted sum-rate of all users by joint optimizing the active beamforming at the access point and the passive beamforming vector of the RIS elements. Unlike most existing works, we consider the more practical situation with the discrete phase shifts and imperfect channel state information (CSI). Specifically, for the situation that the discrete phase shifts and perfect CSI are considered, we first develop a deep quantization neural network (DQNN) to simultaneously design the active and passive beamforming while most reported works design them alternatively. Then, we propose an improved structure (I-DQNN) based on DQNN to simplify the parameters decision process when the control bits of each RIS element are greater than 1 bit. Finally, we extend the two proposed DQNN-based algorithms to the case that the discrete phase shifts and imperfect CSI are considered simultaneously. Our simulation results show that the two DQNN-based algorithms have better performance than traditional algorithms in the perfect CSI case, and are also more robust in the imperfect CSI case.

Zhenhua Xu, Yuxuan Liu, Lu Gan et al.

High-Definition (HD) maps can provide precise geometric and semantic information of static traffic environments for autonomous driving. Road-boundary is one of the most important information contained in HD maps since it distinguishes between road areas and off-road areas, which can guide vehicles to drive within road areas. But it is labor-intensive to annotate road boundaries for HD maps at the city scale. To enable automatic HD map annotation, current work uses semantic segmentation or iterative graph growing for road-boundary detection. However, the former could not ensure topological correctness since it works at the pixel level, while the latter suffers from inefficiency and drifting issues. To provide a solution to the aforementioned problems, in this letter, we propose a novel system termed csBoundary to automatically detect road boundaries at the city scale for HD map annotation. Our network takes as input an aerial image patch, and directly infers the continuous road-boundary graph (i.e., vertices and edges) from this image. To generate the city-scale road-boundary graph, we stitch the obtained graphs from all the image patches. Our csBoundary is evaluated and compared on a public benchmark dataset. The results demonstrate our superiority. The accompanied demonstration video is available at our project page \url{https://sites.google.com/view/csboundary/}.

Wangyang Xu, Jiancheng An, Yongjun Xu et al.

To mitigate the effects of shadow fading and obstacle blocking, reconfigurable intelligent surface (RIS) has become a promising technology to improve the signal transmission quality of wireless communications by controlling the reconfigurable passive elements with less hardware cost and lower power consumption. However, accurate, low-latency and low-pilot-overhead channel state information (CSI) acquisition remains a considerable challenge in RIS-assisted systems due to the large number of RIS passive elements. In this paper, we propose a three-stage joint channel decomposition and prediction framework to require CSI. The proposed framework exploits the two-timescale property that the base station (BS)-RIS channel is quasi-static and the RIS-user equipment (UE) channel is fast time-varying. Specifically, in the first stage, we use the full-duplex technique to estimate the channel between a BS's specific antenna and the RIS, addressing the critical scaling ambiguity problem in the channel decomposition. We then design a novel deep neural network, namely, the sparse-connected long short-term memory (SCLSTM), and propose a SCLSTM-based algorithm in the second and third stages, respectively. The algorithm can simultaneously decompose the BS-RIS channel and RIS-UE channel from the cascaded channel and capture the temporal relationship of the RIS-UE channel for prediction. Simulation results show that our proposed framework has lower pilot overhead than the traditional channel estimation algorithms, and the proposed SCLSTM-based algorithm can also achieve more accurate CSI acquisition robustly and effectively.

Aishwarya Unnikrishnan, Joey Wilson, Lu Gan et al.

This paper reports on a dynamic semantic mapping framework that incorporates 3D scene flow measurements into a closed-form Bayesian inference model. Existence of dynamic objects in the environment can cause artifacts and traces in current mapping algorithms, leading to an inconsistent map posterior. We leverage state-of-the-art semantic segmentation and 3D flow estimation using deep learning to provide measurements for map inference. We develop a Bayesian model that propagates the scene with flow and infers a 3D continuous (i.e., can be queried at arbitrary resolution) semantic occupancy map outperforming its static counterpart. Extensive experiments using publicly available data sets show that the proposed framework improves over its predecessors and input measurements from deep neural networks consistently.

Lu Gan, Youngji Kim, Jessy W. Grizzle et al.

This article presents a novel and flexible multitask multilayer Bayesian mapping framework with readily extendable attribute layers. The proposed framework goes beyond modern metric-semantic maps to provide even richer environmental information for robots in a single mapping formalism while exploiting intralayer and interlayer correlations. It removes the need for a robot to access and process information from many separate maps when performing a complex task, advancing the way robots interact with their environments. To this end, we design a multitask deep neural network with attention mechanisms as our front-end to provide heterogeneous observations for multiple map layers simultaneously. Our back-end runs a scalable closed-form Bayesian inference with only logarithmic time complexity. We apply the framework to build a dense robotic map including metric-semantic occupancy and traversability layers. Traversability ground truth labels are automatically generated from exteroceptive sensory data in a self-supervised manner. We present extensive experimental results on publicly available datasets and data collected by a 3D bipedal robot platform and show reliable mapping performance in different environments. Finally, we also discuss how the current framework can be extended to incorporate more information such as friction, signal strength, temperature, and physical quantity concentration using Gaussian map layers. The software for reproducing the presented results or running on customized data is made publicly available.

Tianyu Liu, Qinghai Liao, Lu Gan et al.

Since early 2020, the coronavirus disease 2019 (COVID-19) has spread rapidly across the world. As at the date of writing this article, the disease has been globally reported in 223 countries and regions, infected over 108 million people and caused over 2.4 million deaths (https://covid19.who.int/, accessed on Feb. 17, 2021). Avoiding person-to-person transmission is an effective approach to control and prevent the pandemic. However, many daily activities, such as transporting goods in our daily life, inevitably involve person-to-person contact. Using an autonomous logistic vehicle to achieve contact-less goods transportation could alleviate this issue. For example, it can reduce the risk of virus transmission between the driver and customers. Moreover, many countries have imposed tough lockdown measures to reduce the virus transmission (e.g., retail, catering) during the pandemic, which causes inconveniences for human daily life. Autonomous vehicle can deliver the goods bought by humans, so that humans can get the goods without going out. These demands motivate us to develop an autonomous vehicle, named as Hercules, for contact-less goods transportation during the COVID-19 pandemic. The vehicle is evaluated through real-world delivering tasks under various traffic conditions.

Lu Gan, Ray Zhang, Jessy W. Grizzle et al.

This paper develops a Bayesian continuous 3D semantic occupancy map from noisy point clouds by generalizing the Bayesian kernel inference model for building occupancy maps, a binary problem, to semantic maps, a multi-class problem. The proposed method provides a unified probabilistic model for both occupancy and semantic probabilities and nicely reverts to the original occupancy mapping framework when only one occupied class exists in obtained measurements. The Bayesian spatial kernel inference relaxes the independent grid assumption and brings smoothness and continuity to the map inference, enabling to exploit local correlations present in the environment and increasing the performance. The accompanying software uses multi-threading and vectorization, and runs at about 2 Hz on a laptop CPU. Evaluations using multiple sequences of stereo camera and LiDAR datasets show that the proposed method consistently outperforms current baselines. We also present a qualitative evaluation using data collected with a bipedal robot platform on the University of Michigan - North Campus.

Ross Hartley, Maani Ghaffari Jadidi, Lu Gan et al.

The factor graph framework is a convenient modeling technique for robotic state estimation where states are represented as nodes, and measurements are modeled as factors. When designing a sensor fusion framework for legged robots, one often has access to visual, inertial, joint encoder, and contact sensors. While visual-inertial odometry has been studied extensively in this framework, the addition of a preintegrated contact factor for legged robots has been only recently proposed. This allowed for integration of encoder and contact measurements into existing factor graphs, however, new nodes had to be added to the graph every time contact was made or broken. In this work, to cope with the problem of switching contact frames, we propose a hybrid contact preintegration theory that allows contact information to be integrated through an arbitrary number of contact switches. The proposed hybrid modeling approach reduces the number of required variables in the nonlinear optimization problem by only requiring new states to be added alongside camera or selected keyframes. This method is evaluated using real experimental data collected from a Cassie-series robot where the trajectory of the robot produced by a motion capture system is used as a proxy for ground truth. The evaluation shows that inclusion of the proposed preintegrated hybrid contact factor alongside visual-inertial navigation systems improves estimation accuracy as well as robustness to vision failure, while its generalization makes it more accessible for legged platforms.

Ross Hartley, Josh Mangelson, Lu Gan et al.

State-of-the-art robotic perception systems have achieved sufficiently good performance using Inertial Measurement Units (IMUs), cameras, and nonlinear optimization techniques, that they are now being deployed as technologies. However, many of these methods rely significantly on vision and often fail when visual tracking is lost due to lighting or scarcity of features. This paper presents a state-estimation technique for legged robots that takes into account the robot's kinematic model as well as its contact with the environment. We introduce forward kinematic factors and preintegrated contact factors into a factor graph framework that can be incrementally solved in real-time. The forward kinematic factor relates the robot's base pose to a contact frame through noisy encoder measurements. The preintegrated contact factor provides odometry measurements of this contact frame while accounting for possible foot slippage. Together, the two developed factors constrain the graph optimization problem allowing the robot's trajectory to be estimated. The paper evaluates the method using simulated and real sensory IMU and kinematic data from experiments with a Cassie-series robot designed by Agility Robotics. These preliminary experiments show that using the proposed method in addition to IMU decreases drift and improves localization accuracy, suggesting that its use can enable successful recovery from a loss of visual tracking.

Lu Gan, Maani Ghaffari Jadidi, Steven A. Parkison et al.

Despite impressive advances in simultaneous localization and mapping, dense robotic mapping remains challenging due to its inherent nature of being a high-dimensional inference problem. In this paper, we propose a dense semantic robotic mapping technique that exploits sparse Bayesian models, in particular, the relevance vector machine, for high-dimensional sequential inference. The technique is based on the principle of automatic relevance determination and produces sparse models that use a small subset of the original dense training set as the dominant basis. The resulting map posterior is continuous, and queries can be made efficiently at any resolution. Moreover, the technique has probabilistic outputs per semantic class through Bayesian inference. We evaluate the proposed relevance vector semantic map using publicly available benchmark datasets, NYU Depth V2 and KITTI; and the results show promising improvements over the state-of-the-art techniques.

Maani Ghaffari Jadidi, Lu Gan, Steven A. Parkison et al.

In this paper, we develop a high-dimensional map building technique that incorporates raw pixelated semantic measurements into the map representation. The proposed technique uses Gaussian Processes (GPs) multi-class classification for map inference and is the natural extension of GP occupancy maps from binary to multi-class form. The technique exploits the continuous property of GPs and, as a result, the map can be inferred with any resolution. In addition, the proposed GP Semantic Map (GPSM) learns the structural and semantic correlation from measurements rather than resorting to assumptions, and can flexibly learn the spatial correlation as well as any additional non-spatial correlation between map points. We extend the OctoMap to Semantic OctoMap representation and compare with the GPSM mapping performance using NYU Depth V2 dataset. Evaluations of the proposed technique on multiple partially labeled RGBD scans and labels from noisy image segmentation show that the GP semantic map can handle sparse measurements, missing labels in the point cloud, as well as noise corrupted labels.

Vahid Abolghasemi, Hao Shen, Yaochun Shen et al.

In this paper, the problem of terahertz pulsed imaging and reconstruction is addressed. It is assumed that an incomplete (subsampled) three dimensional THz data set has been acquired and the aim is to recover all missing samples. A sparsity-inducing approach is proposed for this purpose. First, a simple interpolation is applied to incomplete noisy data. Then, we propose a spatio-temporal dictionary learning method to obtain an appropriate sparse representation of data based on a joint sparse recovery algorithm. Then, using the sparse coefficients and the learned dictionary, the 3D data is effectively denoised by minimizing a simple cost function. We consider two types of terahertz data to evaluate the performance of the proposed approach; THz data acquired for a model sample with clear layered structures (e.g., a T-shape plastic sheet buried in a polythene pellet), and pharmaceutical tablet data (with low spatial resolution). The achieved signal-to-noise-ratio for reconstruction of T-shape data, from only 5% observation was 19 dB. Moreover, the accuracies of obtained thickness and depth measurements for pharmaceutical tablet data after reconstruction from 10% observation were 98.8%, and 99.9%, respectively. These results, along with chemical mapping analysis, presented at the end of this paper, confirm the accuracy of the proposed method.

Kezhi Li, Lu Gan, Cong Ling

In this paper, a new class of circulant matrices built from deterministic sequences is proposed for convolution-based compressed sensing (CS). In contrast to random convolution, the coefficients of the underlying filter are given by the discrete Fourier transform of a deterministic sequence with good autocorrelation. Both uniform recovery and non-uniform recovery of sparse signals are investigated, based on the coherence parameter of the proposed sensing matrices. Many examples of the sequences are investigated, particularly the Frank-Zadoff-Chu (FZC) sequence, the \textit{m}-sequence and the Golay sequence. A salient feature of the proposed sensing matrices is that they can not only handle sparse signals in the time domain, but also those in the frequency and/or or discrete-cosine transform (DCT) domain.