arXiv subjects

Hanxin Zhang

Publications and source records attributed to Hanxin Zhang.

7 recordsLinked to original sources

arXiv2026

R2HandoverSim: A Simulation Framework and Benchmark for Robot-to-Human Object Handovers

We present R2HandoverSim, a simulation benchmark for robot-to-human (R2H) object handovers. Although R2H handover methods have advanced rapidly, the lack of standardized evaluation protocols impedes objective comparison. Our benchmark enables reproducible evaluation by systematically comparing four baselines on their predicted shared grasp poses. We conduct a user study with 30 participants, analyze baseline performance, and show that simulation results correlate with real-world evaluation outcomes. Crucially, five complementary metrics (planning feasibility, reachability, grasp stability, grasp affordance, and safety) better reflect user-perceived handover quality than overall success rate alone. Website and code: https://robot-future.github.io/r2handoversim/.

cs.RO↗

arXiv2026

Embodied Interpretability: Linking Causal Understanding to Generalization in Vision-Language-Action Models

Vision-Language-Action (VLA) policies often fail under distribution shift, suggesting that decisions may depend on spurious visual correlations rather than task-relevant causes. We formulate visual-action attribution as an interventional estimation problem. Accordingly, we introduce the Interventional Significance Score (ISS), an interventional masking procedure for estimating the causal influence of visual regions on action predictions, and the Nuisance Mass Ratio (NMR), a scalar measure of attribution to task-irrelevant features. We analyze the statistical properties of ISS and show that it admits unbiased estimation, and we characterize conditions under which action prediction error provides a valid proxy for causal influence. Experiments across diverse manipulation tasks indicate that NMR predicts generalization behavior and that ISS yields more faithful explanations than existing interpretability methods. These results suggest that interventional attribution provides a simple diagnostic approach for identifying causal misalignment in embodied policies.

cs.RO↗

arXiv2025

Intent-Handover: Grounding Language in Human-Usage Regions for Trustworthy Robot-to-Human Handovers

Spoken instructions in robot-to-human handovers may specify either an object ("the cup") or an intended use ("pour water"); in both cases, successful handover requires the robot to infer the target object and the region remaining available for the human to hold. If the robot grasps that hold region, the object could become awkward to receive and immediately use, potentially reducing perceived competence and trust; if the gripper approaches too close to the receiving hand during delivery, perceived safety may also suffer. We present Intent-Handover, which grounds unconstrained speech and visual scene context into explicit grasp and delivery constraints. Given a spoken instruction and a scene observation, a vision-language model identifies the target object and the intended human-usage region. A grasp optimization module then selects a feasible grasp keeping this region accessible while enforcing clearance from the predicted receiving hand. During execution, the robot tracks upper-body key points to estimate the user's receiving pose and places the handover at an ergonomically feasible location. In a within-subjects ablation study (n=30), human-usage region awareness increases perceived trust, hand-gripper collision avoidance increases perceived safety, and interaction comfort is highest when both are enabled. Website and code: https://robot-future.github.io/intent-handover/.

cs.RO↗

arXiv2021

Lightweight Mobile Automated Assistant-to-physician for Global Lower-resource Areas

Importance: Lower-resource areas in Africa and Asia face a unique set of healthcare challenges: the dual high burden of communicable and non-communicable diseases; a paucity of highly trained primary healthcare providers in both rural and densely populated urban areas; and a lack of reliable, inexpensive internet connections. Objective: To address these challenges, we designed an artificial intelligence assistant to help primary healthcare providers in lower-resource areas document demographic and medical sign/symptom data and to record and share diagnostic data in real-time with a centralized database. Design: We trained our system using multiple data sets, including US-based electronic medical records (EMRs) and open-source medical literature and developed an adaptive, general medical assistant system based on machine learning algorithms. Main outcomes and Measure: The application collects basic information from patients and provides primary care providers with diagnoses and prescriptions suggestions. The application is unique from existing systems in that it covers a wide range of common diseases, signs, and medication typical in lower-resource countries; the application works with or without an active internet connection. Results: We have built and implemented an adaptive learning system that assists trained primary care professionals by means of an Android smartphone application, which interacts with a central database and collects real-time data. The application has been tested by dozens of primary care providers. Conclusions and Relevance: Our application would provide primary healthcare providers in lower-resource areas with a tool that enables faster and more accurate documentation of medical encounters. This application could be leveraged to automatically populate local or national EMR systems.

cs.LG↗

arXiv2018

Interpretation of wake instability at slip line in rotating detonation

In studies on instabilities of flowfield in rotating detonation, one of the most common concerns is the instability at the slip line originating from the conjunction of the detonation wave and oblique shock. Using Euler equations associated with 7-species-and-8-reaction finite-rate chemical reaction model of hydrogen/air mixtures, further studies are performed to simulate the 2-D rotating detonation, and the flow mechanism of instability at the slip line is investigated in depth. The results show that the distinct wake profile exists at the slip line, which is different from the typical mixing layer. Analysis indicates that the generation of wake is caused by the transition shock between the detonation wave and oblique shock. Because of the wake profile, the vorticity distribution therein appears in a double-layer layout, and different evolution exist in different vorticity layers. Based on the velocity profile across the slip line, the analysis by the linear stability theory is made, and two unstable modes which have different shape profiles and phase velocities are found. Discrete Fourier transformation is utilized to analyze the numerical results, and similar shape profiles are obtained. A general coincidence in velocity of vortex movement is also attained between the theoretical predictions and simulations. Investigations show that the wake instability is responsible for the unstable mechanism, and corresponding unstable structures differs from the canonical ones in typical mixing layers.

physics.flu-dyn↗

arXiv2017

Further investigations on the interface instability between fresh injections and burnt products in 2-D rotating detonation

Instabilities in rotating detonation are concerned because of their potential influence on the stability of operation. Previous studies on instability of 2-D rotating detonation mainly cared about the one of the contact discontinuity originated from the conjunction of the detonation and oblique shock. Hishida et al. first found the rippled structure existed in the interface between fresh injections and burnt product from the previous cycle (Shock Waves 19, 2009), and a mechanism of Kelvin-Helmholtz instability was suggested as well. Similar structures were observed as well in simulations by current authors, where a fifth-order WENO-type scheme with improved resolution and 7-species-and-8-reaction chemical model were used. In order to achieve a deep understanding on the flow mechanism, more careful simulations are carried out by using three grids with increasing resolution. The results show that besides the previously-mentioned Kelvin-Helmholtz instability, there are two other mechanisms which take effect in the interface instability, i.e., the effect of baroclinic torque and Rayleigh-Taylor instability. Occurrence conditions for two instabilities are checked and testified. Especially, the spike- and bubble-like structures are observed at the interface, which show appearances different from canonical structures by Kelvin-Helmholtz instability.

physics.flu-dyn↗

arXiv2016

On free-stream preservation in stationary grids for arbitrary linear upwind schemes

In order to improve the application maturity of high-order difference schemes, the free-stream preservation property, whose importance has been widely recognized in recent years, has been developed into a focus of study.. In past literatures, only central schemes are considered to be suitable for free-stream preservation. In this study, the methodology for arbitrary linear schemes to achieve the property is investigated. First, derivations of grid metric by Thomas, Lombard and Neier are reviewed, through which linear schemes for the metric and unsplit flux could attain the property by the proof of Vinokur and Yee firstly. In practical applications, flux splittings are usually at presence and therefore the direct use of upwind schemes seems difficult to fulfill free-stream preservation. To overcome the difficulty, two attempts are made: firstly, a central-scheme-decomposition is worked out, through which a central difference scheme is derived to approximate the first-order partial derivative in metric evaluations; secondly, treatments are proposed for flux splitting and a concrete example is presented. Through these two attempts, the linear upwind node schemes can achieve the property. For half-node or mixed type schemes, interpolations should be used to derive variables at half nodes. As a result, a directionally consistent interpolation is proposed, which is shown to be necessary in order to avoid violation of the metric identity and free-stream preservation. Two numerical problems are also tested, i.e., the free-stream and vortex preservation on wavy, largely randomized and triangular grids. Numerical results validate aforementioned theoretical outcomes; especially, simulations on the triangular grids indicate the methods discussed in this study, which are typical algorithms on structured grids, has the application potential for problems on unstructured meshes.

physics.flu-dyn↗