arXiv subjects

Zhenhua Yu

Publications and source records attributed to Zhenhua Yu.

At least 19 recordsLinked to original sources

arXiv2026

Underwater MMG-Based Muscle State Monitoring with Integrated Emergency Buoyancy Assistance

This paper presents an underwater MMG-driven wearable emergency assistance system for lower-leg muscle-state monitoring and automatic buoyancy deployment. A compact microphone-based MMG sensor was waterproofed using a flexible 5 mil PE membrane, preserving identifiable muscle-vibration responses under immersion, depth variation, and stirring disturbances. Two lower-leg sensors captured stroke-dependent MMG patterns across four swimming styles, and a MiniRocket classifier achieved 91.91% window-level and 97.56% file-level accuracy. For cramp-related monitoring, a pattern-based risk score was used to identify representative pre-cramp abnormal muscle-state transitions during rhythmic motion. A controlled underwater test demonstrated the closed sensing--decision--actuation chain, triggering CO_2 release, airbag inflation, and flotation in less than 5~s. These results support underwater MMG as a sensing basis for wearable robotic emergency assistance in aquatic environments.

cs.HC↗

arXiv2025

Bioinspired Tapered-Spring Turbulence Sensor for Underwater Flow Detection

This paper presents a bio-inspired underwater whisker sensor for robust hydrodynamic disturbance detection and efficient signal analysis based on Physical Reservoir Computing (PRC). The design uses a tapered nylon spring with embedded accelerometers to achieve spatially distributed vibration sensing and frequency separation along the whisker. Towing-tank experiments and computational fluid dynamics simulations confirmed that the whisker effectively distinguishes vortex regimes across different fin angles and maintains Strouhal scaling with flow velocity, where higher speeds increase vibration intensity without affecting the dominant frequencies. Frequency-domain analysis, Shannon entropy, and machine learning further validated the sensing performance: vortex shedding frequencies were identified with less than 10\% error, entropy captured the transition from coherent vortex streets to turbulence, and logistic regression achieved 86.0\% classification accuracy with millisecond-level inference. These results demonstrate that structurally encoded whisker sensing provides a scalable and real-time solution for underwater perception, wake tracking, and turbulence-aware navigation in autonomous marine robots.

physics.ins-det↗

arXiv2025

FlashRecovery: Fast and Low-Cost Recovery from Failures for Large-Scale Training of LLMs

Large language models (LLMs) have made a profound impact across various fields due to their advanced capabilities. However, training these models at unprecedented scales requires extensive AI accelerator clusters and sophisticated parallelism strategies, which pose significant challenges in maintaining system reliability over prolonged training periods. A major concern is the substantial loss of training time caused by inevitable hardware and software failures. To address these challenges, we present FlashRecovery, a fast and low-cost failure recovery system comprising three core modules: (1) Active and real-time failure detection. This module performs continuous training state monitoring, enabling immediate identification of hardware and software failures within seconds, thus ensuring rapid incident response; (2) Scale-independent task restart. By employing different recovery strategies for normal and faulty nodes, combined with an optimized communication group reconstruction protocol, our approach ensures that the recovery time remains nearly constant, regardless of cluster scale; (3) Checkpoint-free recovery within one step. Our novel recovery mechanism enables single-step restoration, completely eliminating dependence on traditional checkpointing methods and their associated overhead. Collectively, these innovations enable FlashRecovery to achieve optimal Recovery Time Objective (RTO) and Recovery Point Objective (RPO), substantially improving the reliability and efficiency of long-duration LLM training. Experimental results demonstrate that FlashRecovery system can achieve training restoration on training cluster with 4, 800 devices in 150 seconds. We also verify that the time required for failure recovery is nearly consistent for different scales of training tasks.

cs.DC↗

arXiv2025

Dimer-projection contact and the clock shift of a unitary Fermi gas

Understanding the dynamics of short-range correlations is a central challenge in strongly interacting Fermi gases. In ultracold gases, these correlations are quantified by the contact parameter, yet measurements to date have been limited to equilibrium systems or relatively slow, global dynamics. Here, we introduce a rapid spectroscopic technique based on projection of the interacting state onto an alternate scattering channel with a low-lying dimer state. We demonstrate contact measurements on the microsecond timescale -- faster than the inverse Fermi energy. Using $^{40}$K near a broad $s$-wave Feshbach resonance, we show that the strength of the dimer-projection feature scales proportionally with the contact parameter extracted from the high-frequency tail of radio-frequency spectroscopy, in agreement with coupled-channels calculations. Analysis of the spectra further reveals that the dimer feature provides the dominant contribution to the clock shift of the unitary Fermi gas, allowing the first experimental bound on this quantity. The observed deviations from universal predictions highlight the importance of multichannel effects. Our results open new avenues for studying contact correlators, hydrodynamic attractors, and quantum critical behavior.

cond-mat.quant-gas↗

arXiv2025

An LLM-enabled Multi-Agent Autonomous Mechatronics Design Framework

Existing LLM-enabled multi-agent frameworks are predominantly limited to digital or simulated environments and confined to narrowly focused knowledge domain, constraining their applicability to complex engineering tasks that require the design of physical embodiment, cross-disciplinary integration, and constraint-aware reasoning. This work proposes a multi-agent autonomous mechatronics design framework, integrating expertise across mechanical design, optimization, electronics, and software engineering to autonomously generate functional prototypes with minimal direct human design input. Operating primarily through a language-driven workflow, the framework incorporates structured human feedback to ensure robust performance under real-world constraints. To validate its capabilities, the framework is applied to a real-world challenge involving autonomous water-quality monitoring and sampling, where traditional methods are labor-intensive and ecologically disruptive. Leveraging the proposed system, a fully functional autonomous vessel was developed with optimized propulsion, cost-effective electronics, and advanced control. The design process was carried out by specialized agents, including a high-level planning agent responsible for problem abstraction and dedicated agents for structural, electronics, control, and software development. This approach demonstrates the potential of LLM-based multi-agent systems to automate real-world engineering workflows and reduce reliance on extensive domain expertise.

cs.RO↗

arXiv2025

Physical Reservoir Computing in Hook-Shaped Rover Wheel Spokes for Real-Time Terrain Identification

Effective terrain detection in unknown environments is crucial for safe and efficient robotic navigation. Traditional methods often rely on computationally intensive data processing, requiring extensive onboard computational capacity and limiting real-time performance for rovers. This study presents a novel approach that combines physical reservoir computing with piezoelectric sensors embedded in rover wheel spokes for real-time terrain identification. By leveraging wheel dynamics, terrain-induced vibrations are transformed into high-dimensional features for machine learning-based classification. Experimental results show that strategically placing three sensors on the wheel spokes achieves 90$\%$ classification accuracy, which demonstrates the accuracy and feasibility of the proposed method. The experiment results also showed that the system can effectively distinguish known terrains and identify unknown terrains by analyzing their similarity to learned categories. This method provides a robust, low-power framework for real-time terrain classification and roughness estimation in unstructured environments, enhancing rover autonomy and adaptability.

cs.RO↗

arXiv2025

Multi-Sensor Fusion-Based Mobile Manipulator Remote Control for Intelligent Smart Home Assistance

This paper proposes a wearable-controlled mobile manipulator system for intelligent smart home assistance, integrating MEMS capacitive microphones, IMU sensors, vibration motors, and pressure feedback to enhance human-robot interaction. The wearable device captures forearm muscle activity and converts it into real-time control signals for mobile manipulation. The wearable device achieves an offline classification accuracy of 88.33\%\ across six distinct movement-force classes for hand gestures by using a CNN-LSTM model, while real-world experiments involving five participants yield a practical accuracy of 83.33\%\ with an average system response time of 1.2 seconds. In Human-Robot synergy in navigation and grasping tasks, the robot achieved a 98\%\ task success rate with an average trajectory deviation of only 3.6 cm. Finally, the wearable-controlled mobile manipulator system achieved a 93.3\%\ gripping success rate, a transfer success of 95.6\%\, and a full-task success rate of 91.1\%\ during object grasping and transfer tests, in which a total of 9 object-texture combinations were evaluated. These three experiments' results validate the effectiveness of MEMS-based wearable sensing combined with multi-sensor fusion for reliable and intuitive control of assistive robots in smart home scenarios.

cs.RO↗

arXiv2025

PARF-Net: integrating pixel-wise adaptive receptive fields into hybrid Transformer-CNN network for medical image segmentation

Convolutional neural networks (CNNs) excel in local feature extraction while Transformers are superior in processing global semantic information. By leveraging the strengths of both, hybrid Transformer-CNN networks have become the major architectures in medical image segmentation tasks. However, existing hybrid methods still suffer deficient learning of local semantic features due to the fixed receptive fields of convolutions, and also fall short in effectively integrating local and long-range dependencies. To address these issues, we develop a new method PARF-Net to integrate convolutions of Pixel-wise Adaptive Receptive Fields (Conv-PARF) into hybrid Network for medical image segmentation. The Conv-PARF is introduced to cope with inter-pixel semantic differences and dynamically adjust convolutional receptive fields for each pixel, thus providing distinguishable features to disentangle the lesions with varying shapes and scales from the background. The features derived from the Conv-PARF layers are further processed using hybrid Transformer-CNN blocks under a lightweight manner, to effectively capture local and long-range dependencies, thus boosting the segmentation performance. By assessing PARF-Net on four widely used medical image datasets including MoNuSeg, GlaS, DSB2018 and multi-organ Synapse, we showcase the advantages of our method over the state-of-the-arts. For instance, PARF-Net achieves 84.27% mean Dice on the Synapse dataset, surpassing existing methods by a large margin.

cs.CV↗

arXiv2024

Dynamic transition of the generalized Jaynes-Cummings model: multi-particles and inter-particle interaction effects

How environments affect dynamics of quantum systems remains a central question in understanding transitions between quantum and classical phenomena and optimizing quantum technologies. A paradigm model to address the above question is the generalized Jaynes-Cummings model, in which a two-level particle is coupled to its environment modeled by a continuum boson modes. Previous analytic solution shows that, starting from the initial state that the particle is in its excited state and the boson modes in their vacuum state, the time evolution of the probability that the particle occupies the excited state exhibits a dynamic transition as the system-environment coupling varies; when the coupling is weak, the probability decays to zero monotonically, while a finite weight of the particle is localized in the excited state when the coupling is sufficiently strong. Here, we study the dynamic transition for the case that $N$ particles are initially excited with the boson modes in their vacuum state. In particular, we access the effects of an all to all Ising type interaction we introduce between the particles. Our calculation is carried out by the non-perturbative time-dependent numerical renormalization group method. We find that the critical coupling for the transition decreases with $N$, and is suppressed (enlarged) by the anti-ferromagnetic (ferromagnetic) Ising interaction. Our results enrich understanding on environmental effects on interacting quantum systems.

cond-mat.quant-gas↗

arXiv2024

Bifurcations of nonlinear dynamics in coupled twin spin masers

Spin masers are a prototype nonlinear dynamic system. They undergo a bifurcation at a critical amplification factor, transiting into a limit cycle phase characterized by a Larmor precession around the external bias magnetic field, thereby serving as a key frequency reference for precision measurements. Recently, a system of coupled twin spin masers placed in dual bias magnetic fields, involving simultaneously two intrinsic Larmor frequencies, has been studied. Compared with previous spin masers, this setup exhibits new attractors such as quasi-periodic orbits and chaos in addition to the usual limit cycles and the trivial no signal fixed point. The richer dynamic phases imply the existence of bifurcations, whose nature has not been fully analyzed. Here, to shed light on the nature of the bifurcations, we turn to a closely related system and systematically study the various bifurcations therein along different routes in parameter space. We identify the bifurcations as of the types including pitchfork, Hopf, homoclinic bifurcations, and saddle-node bifurcations of cycles. By both analytical and numerical methods, we reveal how various attractors interplay with each other and change their stability. We also quantitatively evaluate the locations where these bifurcations occur by tracking the both stable and even not easily detected unstable limit cycles. These findings deepen our understanding of the underlying mechanisms resulting in the rich dynamic phases in the coupled twin spin masers.

quant-ph↗

arXiv2024

Nonlinear spin dynamics induced by feedback under continuous Larmor frequency distributions

Nonlinear spin dynamics are essential in exploring nonequilibrium quantum phenomena and have broad applications in precision measurement. Among these systems, the combination of a bias magnetic field and feedback mechanisms can induce self-sustained oscillations at the base Larmor frequency due to nonlinearity. These features have driven the development of single-species and multiple-species spin masers. The latter, with multiple discrete Larmor frequencies, provides significant advantages for precision measurement by mitigating uncertainties in precession frequencies due to long-term drifts in experimental conditions. The self-sustained oscillations of single-species and multiple-species spin masers correspond to limit cycles and quasi-periodic orbits of the stable nonlinear dynamics of the systems respectively; the correspondence is elucidated in a recent study on a related spin system featuring two discrete intrinsic Larmor frequencies under dual bias magnetic fields. Here, we extend the study to the case that the intrinsic Larmor frequencies of individual spins of the system, given rise to by an inhomogeneous bias magnetic field, form a continuum. We show that generically the stable dynamics of the system includes limit cycles, quasi-periodic orbits, and chaos. We establish the relation between the synchronization frequency of limit cycles and the field inhomogeneity and derive an equation determining the stability of limit cycles. Furthermore, detailed characteristics of different dynamical phases, especially the robustness of limit cycles and quasi-periodic orbits against experimental fluctuations, are discussed. Our findings not only encompass the case of discrete Larmor frequencies, but also provide crucial insights for precision measurement and the exploration of continuous time crystals and quasi-crystals.

quant-ph↗

arXiv2024

The $\mathcal{PT}$-symmetry-breaking transition in a chain of trapped interacting ions

Trapped ions are an ideal platform to implement quantum simulation. Previously the parity-time reversal ($\mathcal{PT}$) symmetry-breaking transition in the paradigmatic non-Hermitian Hamiltonian $h_{PT}=J\sigma_x-i\Gamma\sigma_z$ has been observed in a single ion experiment in a passive way. In this work, we propose to study the interaction effects on the $\mathcal{PT}$-symmetry-breaking transition in a chain of $N$ trapped interacting ions. We consider an effective Ising interaction $H^\text{Ising-x}_\text{int} =\sum_{j<k}U_{jk}\sigma_x^j\sigma_x^k$ between the ions on top of $h_{PT}$. We find that sufficiently strong interaction strength can enhance the $\mathcal{PT}$-symmetric phase for even $N$ while the phase is suppressed in all the other cases. In particular, the suppression can be so strong that even infinitesimal dissipation, quantified by $\Gamma$, can turn the system into the $\mathcal{PT}$-symmetry-breaking phase. In addition, we assess the convolved effects due to the coupling and the spin phase shifts. Our findings can be readily tested in ion chain experiments.

cond-mat.quant-gas↗

arXiv2024

Environment-induced Transitions in Many-body Quantum Teleportation

Quantum teleportation is a phenomenon arising from entanglement, decisively distinguishing the classical and quantum worlds. The recent success of many-body quantum teleportation is even more surprising: although input information is initially dispersed and encoded into the many-body state in a complex way, the teleportation process can refocus this highly non-local information at the receiver's end. This success manifests intriguing capability of many-body systems in quantum information processing. Current studies indicate that information scrambling, a generic dynamic process in many-body systems, underlies the effectiveness of many-body quantum teleportation. However, this process is known to undergo a novel scrambling-dissipation transition in the presence of environments. How environments affect the quantum information processing capability of many-body systems calls for further investigation. In this work, we study many-body quantum teleportation in the presence of environments. We predict two emergent critical points that hallmark the transitions of the teleportation performance from the quantum regime to the classical regime, and finally to the no-signal regime as the system-environment coupling, quantified by $\gamma$, increases. In the quantum regime, teleportation can outperform its classical counterparts, while in the classical regime, it can be replaced by a classical channel. Our prediction is based on a generic argument harnessing the relationship between many-body quantum teleportation and information scrambling, corroborated by solvable Brownian Sachdev-Ye-Kitaev models.

quant-ph↗

arXiv2024

Environment-Induced Information Scrambling Transition with Charge Conservations

In generic closed quantum systems, the complexity of operators increases under time evolution governed by the Heisenberg equation, reflecting the scrambling of local quantum information. However, when systems interact with an external environment, the system-environment coupling allows operators to escape from the system, inducing a dynamical transition between the scrambling phase and the dissipative phase. This transition is known as the environment-induced information scrambling transition, originally proposed in Majorana fermion systems. In this work, we advance this dicovery by investigating the transition in charge-conserved systems with space-time randomness. We construct solvable Brownian Sachdev-Ye-Kitaev models of complex fermions coupled to an environment, enabling the analytical computation of operator growth. We determine the critical dissipation strength, which is proportional to $n(1-n)$ with $n$ being the density of the complex fermions, arising from the suppression in the quantum Lyapunov exponent due to the Pauli blockade in the scattering process. We further analyze the density dependence of maximally scrambled operators at late time. Our results shed light on the intriguing interplay between information scrambling, dissipation, and conservation laws.

quant-ph↗

arXiv2023

Towards the Neuromorphic Computing for Offroad Robot Environment Perception and Navigation

My research objective is to explicitly bridge the gap between high computational performance and low power dissipation of robot on-board hardware by designing a bio-inspired tapered whisker neuromorphic computing (also called reservoir computing) system for offroad robot environment perception and navigation, that centres the interaction between a robot's body and its environment. Mobile robots performing tasks in unknown environments need to traverse a variety of complex terrains, and they must be able to reliably and quickly identify and characterize these terrains to avoid getting into potentially challenging or catastrophic circumstances. To solve this problem, I drew inspiration from animals like rats and seals, just relying on whiskers to perceive surroundings information and survive in dark and narrow environments. Additionally, I looked to the human cochlear which can separate different frequencies of sound. Based on these insights, my work addresses this need by exploring the physical whisker-based reservoir computing for quick and cost-efficient mobile robots environment perception and navigation step by step. This research could help us understand how the compliance of the biological counterparts helps robots to dynamically interact with the environment and provides a new solution compared with current methods for robot environment perception and navigation with limited computational resources, such as Mars.

cs.RO↗

arXiv2023

Feedback-Induced Nonlinear Spin Dynamics in an Inhomogeneous Magnetic Field

Nonlinear effects are the root of interesting phenomena such as masers and lasers, and play a significant role in science and engineering. In spin systems, nonlinear spin dynamics is crucial for the prediction of complex dynamical behavior such as self-organizing oscillation, with applications ranging from spin masers and time crystals to precision measurement. However, when a spin system operates in a static magnetic field, how the inhomogeneity of the field affects its dynamics is a primary concern. Here we study the dynamics of a collection of spins with multiple Larmor frequencies for modeling a static inhomogeneous magnetic field, and reveal that due to the nonlinearity induced by a feedback scheme, the spin system exhibits much richer stable dynamical phases, including quasi-periodic orbits and chaos besides the usual limit cycles emerged in previous works. These phases are generally applicable to coupled nonlinear spin systems, even with more than two intrinsic Larmor frequencies or in continuum cases. Furthermore, we discuss their robustness against the experimental noises and the feasibility of realization in several spin systems. Our findings contribute to future observation of nonlinear dynamical phases and prospective applications in precision measurement.

cond-mat.quant-gas↗

arXiv2022

Dynamical Transition of Operator Size Growth in Quantum Systems Embedded in an Environment

In closed generic many-body systems, unitary evolution disperses local quantum information into highly non-local objects, resulting in thermalization. Such a process is called information scrambling, whose swiftness is quantified by the operator size growth. However, for quantum systems embedded in an environment, how the couplings to the environment affect the process of information scrambling quests revelation. Here we predict a dynamical transition in quantum systems with all-to-all interactions accompanied by an environment, which separates two phases. In the dissipative phase, information scrambling halts as the operator size decays with time, while in the scrambling phase, dispersion of information persists and the operator size grows and saturates to an $O(N)$ value in the long-time limit with $N$ the number of degrees of freedom of the systems. The transition is driven by the competition between the system intrinsic and environment propelled scramblings and the environment induced dissipation. Our prediction is derived from a general argument based on epidemiological models and demonstrated analytically via solvable Brownian SYK models. We provide further evidence which suggests that the transition is generic to quantum chaotic systems when coupled to an environment. Our study sheds light on the fundamental behavior of quantum systems in the presence of an environment.

quant-ph↗

arXiv2022

CWP: Instance complexity weighted channel-wise soft masks for network pruning

Existing differentiable channel pruning methods often attach scaling factors or masks behind channels to prune filters with less importance, and implicitly assume uniform contribution of input samples to filter importance. Specifically, the effects of instance complexity on pruning performance are not yet fully investigated in static network pruning. In this paper, we propose a simple yet effective differentiable network pruning method CWP based on instance complexity weighted filter importance scores. We define instance complexity related weight for each instance by giving higher weights to hard instances, and measure the weighted sum of instance-specific soft masks to model non-uniform contribution of different inputs, which encourages hard instances to dominate the pruning process and the model performance to be well preserved. In addition, we introduce a regularizer to maximize polarization of the masks, such that a sweet spot can be easily found to identify the filters to be pruned. Performance evaluations on various network architectures and datasets demonstrate CWP has advantages over the state-of-the-arts in pruning large networks. For instance, CWP improves the accuracy of ResNet56 on CIFAR-10 dataset by 0.32% aftering removing 64.11% FLOPs, and prunes 87.75% FLOPs of ResNet50 on ImageNet dataset with only 0.93% Top-1 accuracy loss.

cs.LG↗