arXiv subjects

Yuanqing Wu

Publications and source records attributed to Yuanqing Wu.

8 recordsLinked to original sources

arXiv2026

An Interval Branch-and-Bound-Based Inverse Kinemetics Algorithm Towards Global Optimal Redundancy Resolution

The general inverse kinematics (IK) problem of a manipulator, namely that of acquiring the self-motion manifold (SMM) of all admissible joint angles for a desired end-effector pose, plays a vital role in robotics modeling, planning and control. To efficiently solve the generalized IK, this paper proposes an interval branch-and-bound-based approach, which is augmented with a fast numerical IK-solver-enabled search heuristics. In comparison to independent solutions generated by sampling based methods, our approach generates patches of neighboring solutions to provide richer information of the inherent geometry of the SMM for optimal planning and other applications. It can also be utilized in an anytime fashion to obtain solutions with sub-optimal resolution for applications within a limited period. The performance of our approach is verified by numerical experiments on both non-redundant and redundant manipulators.

cs.RO↗

arXiv2026

HE-VPR: Height Estimation Enabled Aerial Visual Place Recognition Against Scale Variance

In this work, we propose HE-VPR, a visual place recognition (VPR) framework that incorporates height estimation. Our system decouples height inference from place recognition, allowing both modules to share a frozen DINOv2 backbone. Two lightweight bypass adapter branches are integrated into our system. The first estimates the height partition of the query image via retrieval from a compact height database, and the second performs VPR within the corresponding height-specific sub-database. The adaptation design reduces training cost and significantly decreases the search space of the database. We also adopt a center-weighted masking strategy to further enhance the robustness against scale differences. Experiments on two self-collected challenging multi-altitude datasets demonstrate that HE-VPR achieves up to 6.1\% Recall@1 improvement over state-of-the-art ViT-based baselines and reduces memory usage by up to 90\%. These results indicate that HE-VPR offers a scalable and efficient solution for height-aware aerial VPR, enabling practical deployment in GNSS-denied environments. All the code and datasets for this work have been released on https://github.com/hmf21/HE-VPR.

cs.RO↗

arXiv2026

Adaptive Dual-Weighting Framework for Federated Learning via Out-of-Distribution Detection

Federated Learning (FL) enables collaborative model training across large-scale distributed service nodes while preserving data privacy, making it a cornerstone of intelligent service systems in edge-cloud environments. However, in real-world service-oriented deployments, data generated by heterogeneous users, devices, and application scenarios are inherently non-IID. This severe data heterogeneity critically undermines the convergence stability, generalization ability, and ultimately the quality of service delivered by the global model. To address this challenge, we propose FLood, a novel FL framework inspired by out-of-distribution (OOD) detection. FLood dynamically counteracts the adverse effects of heterogeneity through a dual-weighting mechanism that jointly governs local training and global aggregation. At the client level, it adaptively reweights the supervised loss by upweighting pseudo-OOD samples, thereby encouraging more robust learning from distributionally misaligned or challenging data. At the server level, it refines model aggregation by weighting client contributions according to their OOD confidence scores, prioritizing updates from clients with higher in-distribution consistency and enhancing the global model's robustness and convergence stability. Extensive experiments across multiple benchmarks under diverse non-IID settings demonstrate that FLood consistently outperforms state-of-the-art FL methods in both accuracy and generalization. Furthermore, FLood functions as an orthogonal plug-in module: it seamlessly integrates with existing FL algorithms to boost their performance under heterogeneity without modifying their core optimization logic. These properties make FLood a practical and scalable solution for deploying reliable intelligent services in real-world federated environments.

cs.LG↗

arXiv2025

Primal-dual policy learning for mean-field stochastic LQR problem

Integrating data-driven techniques with mechanism-driven insights has recently gained popularity as a powerful learning approach to solving traditional LQR problems for designing intelligent controllers in complex dynamic systems. However, the theoretical understanding of various reinforcement learning algorithms needs further exploration to enhance their efficiency and safety. In this article, by means of primal-dual optimization tools, we study the partially model-free design of the mean-field stochastic LQR (MF-SLQR) controller using a policy learning approach. Firstly, by designing appropriate optimizing variables, the considered MF-SLQR problem is transformed into a new static nonconvex constrained optimization problem with equivalence preserved in certain senses. After that, the equivalent formulation of the duality results is constructed via finding the solution of the generalized Lyapunov equation. Then, the strong duality is analyzed, based on which we establish a primal-dual algorithm by Karush-Kuhn-Tucker conditions. More importantly, a partially model-free implementation is also presented, which has a direct connection with the classical policy iteration algorithm. Finally, we use a high-dimensional example to validate our methods.

math.OC↗

arXiv2020

A Decoupled Scheme to Solve the Mass and Momentum Conservation Equations of the Improved Darcy-Brinkman-Forchheimer Framework in Matrix Acidization

Matrix acidization simulation is a challenging task in the study of flows in porous media, due to the changing porosity in the procedure. The improved DBF framework is one model to do this simulation, and its numerical scheme discretises the mass and momentum conservation equations together to form a pressure-velocity linear system. However, this linear system can only be solved by direct solvers to solve for pressure and velocity simultaneously, since zeros appear in the diagonal of the coefficient matrix. Considering the large-scale attribute of matrix acidization simulation, the solving time of direct solvers is not intolerant. Thus, a decoupled scheme is proposed in this work to decouple the coupled pressure-velocity linear system into two independent linear systems: one is to solve for pressure, and the other one is to solve for velocity. Both of the new linear systems can be solved by parallel and iterative solvers, which guarantees the large-scale simulation can be finished in a reasonable time period. A numerical experiment is carried out to demonstrate the correctness of the decoupled scheme and its higher computing efficiency.

physics.comp-ph↗

arXiv2020

Thermodynamically Consistent Darcy-Brinkman-Forchheimer Framework in Matrix Acidization

Matrix acidization is an important technique to enhance oil production at the tertiary recovery stage, and its numerical simulation is never concluded. From one of the earliest models, i.e. the two-scale model (Darcy framework), the Darcy-Brinkman-Forchheimer (DBF) framework is developed by adding Brinkman term and Forchheimer term to the momentum conservation equation. However, in the momentum conservation equation of the DBF framework, porosity is put outside of the time derivation term, which cannot describe the change of porosity well. Thus, this work changes the expression so that the modified momentum conservation equation can satisfy Newton's second law. The modified framework is called improved DBF framework. Furthermore, based on the improved DBF framework, the thermal DBF framework is given by introducing the energy balance equation to the improved DBF framework. Both of the frameworks are verified by the former works through numerical experiments and chemical experiments in labs. Parallelization to the codes of the complicated frameworks is also realized, and good scalability can be achieved.

physics.flu-dyn↗

arXiv2016

A Novel GPU-based Parallel Implementation Scheme and Performance Analysis of Robot Forward Dynamics Algorithms

We propose a novel unifying scheme for parallel implementation of articulated robot dynamics algorithms. It is based on a unified Lie group notation for deriving the equations of motion of articulated robots, where various well-known forward algorithms differ only by their joint inertia matrix inversion strategies. This new scheme leads to a unified abstraction of state-of-the-art forward dynamics algorithms into combinations of block bi-diagonal and/or block tri-diagonal systems, which may be efficiently solved by parallel all-prefix-sum operations (scan) and parallel odd-even elimination (OEE) respectively. We implement the proposed scheme on a Nvidia CUDA GPU platform for the comparative study of three algorithms, namely the hybrid articulated-body inertia algorithm (ABIA), the parallel joint space inertia inversion algorithm (JSIIA) and the constrained force algorithm (CFA), and the performances are analyzed.

cs.RO↗

arXiv2016

Parallel Dynamics Computation using Prefix Sum Operations

We propose a new parallel framework for fast computation of inverse and forward dynamics of articulated robots based on prefix sums (scans). We re-investigate the well-known recursive Newton-Euler formulation of robot dynamics and show that the forward-backward propagation process for robot inverse dynamics is equivalent to two scan operations on certain semigroups. We show that the state-of-the-art forward dynamics algorithms may almost completely be cast into a sequence of scan operations, with unscannable parts clearly identified. This suggests a serial-parallel hybrid approach for systems with a moderate number of links. We implement our scan based algorithms on Nvidia CUDA platform with performance compared with multithreading CPU-based recursive algorithms; a significant level of acceleration is demonstrated.

cs.RO↗