arXiv subjects

Chang Wu

Publications and source records attributed to Chang Wu.

32 records · Page 2Linked to original sources

arXiv2024

MSSP : A Versatile Multi-Scenario Adaptable Intelligent Robot Simulation Platform Based on LIDAR-Inertial Fusion

This letter presents a multi-scenario adaptable intelligent robot simulation platform based on LIDAR-inertial fusion, with three main features: (1 The platform includes an versatile robot model that can be freely controlled through manual control or autonomous tracking. This model is equipped with various types of LIDAR and Inertial Measurement Unit (IMU), providing ground truth information with absolute accuracy. (2 The platform provides a collection of simulation environments with diverse characteristic information and supports developers in customizing and modifying environments according to their needs. (3 The platform supports evaluation of localization performance for SLAM frameworks. Ground truth with absolute accuracy eliminates the inherent errors of global positioning sensors present in real experiments, facilitating detailed analysis and evaluation of the algorithms. By utilizing the simulation platform, developers can overcome the limitations of real environments and datasets, enabling fine-grained analysis and evaluation of mainstream SLAM algorithms in various environments. Experiments conducted in different environments and with different LIDARs demonstrate the wide applicability and practicality of our simulation platform. The implementation of the simulation platform is open-sourced on Github.

cs.RO↗

arXiv2024

MolTC: Towards Molecular Relational Modeling In Language Models

Molecular Relational Learning (MRL), aiming to understand interactions between molecular pairs, plays a pivotal role in advancing biochemical research. Recently, the adoption of large language models (LLMs), known for their vast knowledge repositories and advanced logical inference capabilities, has emerged as a promising way for efficient and effective MRL. Despite their potential, these methods predominantly rely on the textual data, thus not fully harnessing the wealth of structural information inherent in molecular graphs. Moreover, the absence of a unified framework exacerbates the issue of information underutilization, as it hinders the sharing of interaction mechanism learned across diverse datasets. To address these challenges, this work proposes a novel LLM-based multi-modal framework for Molecular inTeraction prediction following Chain-of-Thought (CoT) theory, termed MolTC, which effectively integrate graphical information of two molecules in pair. To train MolTC efficiently, we introduce a Multi-hierarchical CoT concept to refine its training paradigm, and conduct a comprehensive Molecular Interactive Instructions dataset for the development of biochemical LLMs involving MRL. Our experiments, conducted across various datasets involving over 4,000,000 molecular pairs, exhibit the superiority of our method over current GNN and LLM-based baselines. Code is available at https://github.com/MangoKiller/MolTC.

q-bio.QM↗

arXiv2024

Performance Optimization in RSMA-assisted Uplink xURLLC IIoT Networks with Statistical QoS Provisioning

Industry 5.0 and beyond networks have driven the emergence of numerous mission-critical applications, prompting contemplation of the neXt-generation ultra-reliable low-latency communication (xURLLC). To guarantee low-latency requirements, xURLLC heavily relies on short-blocklength packets with sporadic arrival traffic. As a disruptive multi-access technique, rate-splitting multiple access (RSMA) has emerged as a promising avenue to enhance quality of service (QoS) and flexibly manage interference for next-generation communication networks. In this paper, we investigate an innovative RSMA-assisted uplink xURLLC industrial internet-of-things (IIoT) (RSMA-xURLLC-IIoT) network. To unveil reliable insights into the statistical QoS provisioning (SQP) for our proposed network with sporadic arrival traffic, we leverage stochastic network calculus (SNC) to develop a dependable theoretical framework. Building upon this theoretical framework, we formulate the SQP-driven short-packet size maximization problem and the SQP-driven transmit power minimization problem, aiming to guarantee the SQP performance to latency, decoding, and reliability while maximizing the short-packet size and minimizing the transmit power, respectively. By exploiting Monte-Carlo methods, we have thoroughly validated the dependability of the developed theoretical framework. Moreover, through extensive comparison analysis with state-of-the-art multi-access techniques, including non-orthogonal multiple access (NOMA) and orthogonal multiple access (OMA), we have demonstrated the superior performance gains achieved by the proposed RSMA-xURLLC-IIoT networks.

cs.NI↗

arXiv2024

Light-SLAM: A Robust Deep-Learning Visual SLAM System Based on LightGlue under Challenging Lighting Conditions

Simultaneous Localization and Mapping (SLAM) has become a critical technology for intelligent transportation systems and autonomous robots and is widely used in autonomous driving. However, traditional manual feature-based methods in challenging lighting environments make it difficult to ensure robustness and accuracy. Some deep learning-based methods show potential but still have significant drawbacks. To address this problem, we propose a novel hybrid system for visual SLAM based on the LightGlue deep learning network. It uses deep local feature descriptors to replace traditional hand-crafted features and a more efficient and accurate deep network to achieve fast and precise feature matching. Thus, we use the robustness of deep learning to improve the whole system. We have combined traditional geometry-based approaches to introduce a complete visual SLAM system for monocular, binocular, and RGB-D sensors. We thoroughly tested the proposed system on four public datasets: KITTI, EuRoC, TUM, and 4Season, as well as on actual campus scenes. The experimental results show that the proposed method exhibits better accuracy and robustness in adapting to low-light and strongly light-varying environments than traditional manual features and deep learning-based methods. It can also run on GPU in real time.

cs.CV↗

arXiv2023

Cross-Layer Optimization for Statistical QoS Provision in C-RAN with Finite-Length Coding

The cloud radio access network (C-RAN) has become the foundational structure for various emerging communication paradigms, leveraging the flexible deployment of distributed access points (APs) and centralized task processing. In this paper, we propose a cross-layer optimization framework based on a practical finite-length coding communication system in C-RAN, aiming at maximizing bandwidth efficiency while providing statistical quality of service (QoS) for individual services. Based on the theoretical results from effective capacity and finite-length coding, we formulate a joint optimization problem involving modulation and coding schemes (MCS), retransmission count, initial bandwidth allocation and AP selection, which reflects the coordinated decision of parameters across the physical layer, data link layer and transport layer. To tackle such a mixed-integer nonlinear programming (MINLP) problem, we firstly decompose it into a transmission parameter decision (TPD) sub-problem and a user association (UA) sub-problem, which can be solved by a binary search-based algorithm and an auction-based algorithm respectively. Simulation results demonstrate that the proposed model can accurately capture the impact of QoS requirements and channel quality on the optimal transmission parameters. Furthermore, compared with fixed transmission parameter setting, the proposed algorithms achieve the bandwidth efficiency gain up to 27.87% under various traffic and channel scenarios.

cs.IT↗

arXiv2023

VoxelMap++: Mergeable Voxel Mapping Method for Online LiDAR(-inertial) Odometry

This paper presents VoxelMap++: a voxel mapping method with plane merging which can effectively improve the accuracy and efficiency of LiDAR(-inertial) based simultaneous localization and mapping (SLAM). This map is a collection of voxels that contains one plane feature with 3DOF representation and corresponding covariance estimation. Considering total map will contain a large number of coplanar features (kid planes), these kid planes' 3DOF estimation can be regarded as the measurements with covariance of a larger plane (father plane). Thus, we design a plane merging module based on union-find which can save resources and further improve the accuracy of plane fitting. This module can distinguish the kid planes in different voxels and merge these kid planes to estimate the father plane. After merging, the father plane 3DOF representation will be more accurate than the kids plane and the uncertainty will decrease significantly which can further improve the performance of LiDAR(-inertial) odometry. Experiments on challenging environments such as corridors and forests demonstrate the high accuracy and efficiency of our method compared to other state-of-the-art methods (see our attached video). By the way, our implementation VoxelMap++ is open-sourced on GitHub which is applicable for both non-repetitive scanning LiDARs and traditional scanning LiDAR.

cs.RO↗

arXiv2023

Streaming 360-degree VR Video with Statistical QoS Provisioning in mmWave Networks from Delay and Rate Perspectives

Millimeter-wave(mmWave) technology has emerged as a promising enabler for unleashing the full potential of 360-degree virtual reality (VR). However, the explosive growth of VR services, coupled with the reliability issues of mmWave communications, poses enormous challenges in terms of wireless resource and quality-of-service (QoS) provisioning for mmWave-enabled 360-degree VR. In this paper, we propose an innovative 360-degree VR streaming architecture that addresses three under-exploited issues: overlapping field-of-views (FoVs), statistical QoS provisioning (SQP), and loss-tolerant active data discarding. Specifically, an overlapping FoV-based optimal joint unicast and multicast (JUM) task assignment scheme is designed to implement the non-redundant task assignments, thereby conserving wireless resources remarkably. Furthermore, leveraging stochastic network calculus, we develop a comprehensive SQP theoretical framework that encompasses two SQP schemes from delay and rate perspectives. Additionally, a corresponding optimal adaptive joint time-slot allocation and active-discarding (ADAPT-JTAAT) transmission scheme is proposed to minimize resource consumption while guaranteeing diverse statistical QoS requirements under loss-intolerant and loss-tolerant scenarios from delay and rate perspectives, respectively. Extensive simulations demonstrate the effectiveness of the designed overlapping FoV-based JUM optimal task assignment scheme. Comparisons with six baseline schemes validate that the proposed optimal ADAPTJTAAT transmission scheme can achieve superior SQP performance in resource utilization, flexible rate control, and robust queue behaviors.

cs.IT↗

arXiv2022

Global Priors Guided Modulation Network for Joint Super-Resolution and Inverse Tone-Mapping

Joint super-resolution and inverse tone-mapping (SR-ITM) aims to enhance the visual quality of videos that have quality deficiencies in resolution and dynamic range. This problem arises when using 4K high dynamic range (HDR) TVs to watch a low-resolution standard dynamic range (LR SDR) video. Previous methods that rely on learning local information typically cannot do well in preserving color conformity and long-range structural similarity, resulting in unnatural color transition and texture artifacts. In order to tackle these challenges, we propose a global priors guided modulation network (GPGMNet) for joint SR-ITM. In particular, we design a global priors extraction module (GPEM) to extract color conformity prior and structural similarity prior that are beneficial for ITM and SR tasks, respectively. To further exploit the global priors and preserve spatial information, we devise multiple global priors guided spatial-wise modulation blocks (GSMBs) with a few parameters for intermediate feature modulation, in which the modulation parameters are generated by the shared global priors and the spatial features map from the spatial pyramid convolution block (SPCB). With these elaborate designs, the GPGMNet can achieve higher visual quality with lower computational complexity. Extensive experiments demonstrate that our proposed GPGMNet is superior to the state-of-the-art methods. Specifically, our proposed model exceeds the state-of-the-art by 0.64 dB in PSNR, with 69$\%$ fewer parameters and 3.1$\times$ speedup. The code will be released soon.

cs.CV↗

arXiv2022

SDRTV-to-HDRTV Conversion via Spatial-Temporal Feature Fusion

HDR(High Dynamic Range) video can reproduce realistic scenes more realistically, with a wider gamut and broader brightness range. HDR video resources are still scarce, and most videos are still stored in SDR (Standard Dynamic Range) format. Therefore, SDRTV-to-HDRTV Conversion (SDR video to HDR video) can significantly enhance the user's video viewing experience. Since the correlation between adjacent video frames is very high, the method utilizing the information of multiple frames can improve the quality of the converted HDRTV. Therefore, we propose a multi-frame fusion neural network \textbf{DSLNet} for SDRTV to HDRTV conversion. We first propose a dynamic spatial-temporal feature alignment module \textbf{DMFA}, which can align and fuse multi-frame. Then a novel spatial-temporal feature modulation module \textbf{STFM}, STFM extracts spatial-temporal information of adjacent frames for more accurate feature modulation. Finally, we design a quality enhancement module \textbf{LKQE} with large kernels, which can enhance the quality of generated HDR videos. To evaluate the performance of the proposed method, we construct a corresponding multi-frame dataset using HDR video of the HDR10 standard to conduct a comprehensive evaluation of different methods. The experimental results show that our method obtains state-of-the-art performance. The dataset and code will be released.

eess.IV↗

arXiv2022

SDRTV-to-HDRTV via Hierarchical Dynamic Context Feature Mapping

In this work, we address the task of SDR videos to HDR videos(SDRTV-to-HDRTV). Previous approaches use global feature modulation for SDRTV-to-HDRTV. Feature modulation scales and shifts the features in the original feature space, which has limited mapping capability. In addition, the global image mapping cannot restore detail in HDR frames due to the luminance differences in different regions of SDR frames. To resolve the appeal, we propose a two-stage solution. The first stage is a hierarchical Dynamic Context feature mapping (HDCFM) model. HDCFM learns the SDR frame to HDR frame mapping function via hierarchical feature modulation (HME and HM ) module and a dynamic context feature transformation (DCT) module. The HME estimates the feature modulation vector, HM is capable of hierarchical feature modulation, consisting of global feature modulation in series with local feature modulation, and is capable of adaptive mapping of local image features. The DCT module constructs a feature transformation module in conjunction with the context, which is capable of adaptively generating a feature transformation matrix for feature mapping. Compared with simple feature scaling and shifting, the DCT module can map features into a new feature space and thus has a more excellent feature mapping capability. In the second stage, we introduce a patch discriminator-based context generation model PDCG to obtain subjective quality enhancement of over-exposed regions. PDCG can solve the problem that the model is challenging to train due to the proportion of overexposed regions of the image. The proposed method can achieve state-of-the-art objective and subjective quality results. Specifically, HDCFM achieves a PSNR gain of 0.81 dB at a parameter of about 100K. The number of parameters is 1/14th of the previous state-of-the-art methods. The test code will be released soon.

cs.MM↗

arXiv2021

Fast Activation of Graphene with Corrugated Surface and its Role in Improved Aqueous Electrochemical Capacitors

In graphene based materials, the energy storage capacity is usually improved by rich porous structures with extremely high surface area. By utilizing surface corrugations, this work shows an alternative strategy to activate graphene materials for large capacitance. We demonstrate how to simply fabricate such activated graphene and how these surface structures helped to realize considerable specific capacitance (e.g., electrode capacitance of ~340 F g-1 at 5 mV s-1 and device capacitance of ~ 343 F g-1 at 1.7 A g-1) and power performance (e.g., power density of 50 and 2500 W kg-1 at the energy density of ~10.7 and 1.53 Wh kg-1, respectively) in aqueous system, which are comparable to and even better than those of highly activated graphene materials with ultra-high surface area. This work demonstrates a new path to enhance the capacity of carbon-based materials, which could be developed and combined with other systems for various improved energy storage applications.

physics.app-ph↗

arXiv2021

Unzipping chemical bond of non-layered bulk structures to form ultrathin nanocrystals

The rich electronic and band structures of monolayered crystals distinguished from their layered bulk counterparts offer versatile physical/chemical properties and applications.1-5 Their fabrications, particularly the top-down "exfoliations", are successful promised by the weak Van der Waals force between monolayers.6-9 Differentially, un-zipping ultra-thin crystals (e.g. with only one layer of crystal plane) from non-layered structures is highly challenging due to the strong chemical bond between planes and atoms. Alterative finely controlled growth of these ultra-thin materials is not really successful. This work demonstrates how a technique can be used to unzip and disintegrate ultra-thin crystal plane (e.g. monolayered nanocrystals and nanosheets) from bulk non-layered structures (ZnO, alpha/belta-MnO2, TiO2, alpha-TiB2), and present how the basic optical properties changed to distinguish from their bulk phases. The work here gives a strong tool kit to various novel 2D non-layered nanomaterials, providing significant contributions to the family of two-dimensional materials, potentially paving the way for various practical applications.

cond-mat.mtrl-sci↗

arXiv2020

Safe Use of Jet Pull

Jet pull is an observable designed to probe colour flow between jets. Thus far, a particular projection of the pull vector, the pull angle, has been employed to distinguish colour flow between jets produced by a colour singlet or an octet decay. This is of particular importance in order to separate the decay of a Higgs boson to a pair of bottom quarks from the QCD background. However, the pull angle is not infra-red and collinear (IRC) safe. In this paper we introduce IRC safe projections of the pull vector that exhibit good sensitivity to colour flow, while maintaining calculability. We calculate these distributions to next-to-leading logarithmic accuracy, in the context of the hadronic decay of a Higgs boson, and compare these results to Monte Carlo simulations. This study allows us to define an IRC safe version of the pull angle in terms of asymmetry distributions. Furthermore, because of their sensitivity to wide-angle soft radiation, we anticipate that these asymmetries can play an important role in assessing subleading colour correlations and their modelling in general-purpose Monte Carlo parton showers.

hep-ph↗

arXiv2019

Theory Predictions for the Pull Angle

Pull is a jet observable that is sensitive to color flow between dipoles. It has seen wide use for discrimination of particles with similar decay topologies but carrying different color representations and has been measured on W bosons from top quark decays by the D$\emptyset$ and ATLAS experiments. In this paper, we present the first theoretical predictions of pull, focusing on a color-singlet decaying in two jets. The pull angle observable is particularly sensitive to color flow, but is not infrared and collinear safe and so cannot be calculated in fixed-order perturbation theory. Nevertheless, all-orders resummation renders its distribution finite, a property referred to as Sudakov safety. In our prediction of the pull angle we also include an estimation of the effects from hadronization, and directly compare our results to simulation and experimental data.

hep-ph↗