Hand-like autonomous flying robot for airborne grasping and interaction.

To fully leverage the capabilities of mobile manipulation robots, it is imperative that they are able to autonomously execute long-horizon tasks in large unexplored environments. While large language models (LLMs) have shown emergent reasoning skills on arbitrary tasks, existing work primarily concentrates on explored environments, typically focusing on either navigation or manipulation tasks in isolation. In this work, we propose MoMa-LLM, a novel approach that grounds language models within structured representations derived from open-vocabulary scene graphs, dynamically updated as the environment is explored. We tightly interleave these representations with an object-centric action space. Given object detections, the resulting approach is zero-shot, open-vocabulary, and readily extendable to a spectrum of mobile manipulation and household robotic tasks. We demonstrate the effectiveness of MoMa-LLM in a novel semantic interactive search task in large realistic indoor environments. In extensive experiments in both simulation and the real world, we show substantially improved search efficiency compared to conventional baselines and state-of-the-art approaches, as well as its applicability to more abstract tasks.

Catalase is a characteristic enzyme of peroxisomes, of which it is the most abundant protein. This enzyme serves as a typical example of a peroxisomal enzyme and is important in the teaching of biochemistry and molecular biology. Although there is substantial information regarding catalase purification, purifying catalase for the junior‑grade undergraduate experimental course face challenges in obtaining materials and increasingly expensive purification equipment. This study presents a simple method for the purification of mouse liver catalase using ethanol‑chloroform treatment, sodium sulfate fractionation, dialysis and Sephadex G‑200 gel filtration chromatography. Catalase was purified 31.8‑fold with an 18.3% yield. The advantages of this method were its low operating environment requirements, simple procedure and reduced cost. Furthermore, the method was designed to improve students' comprehensive ability and manipulative ability and to introduce a sense of innovation in the fields of biochemistry and molecular biology during their junior year.

Abriendo Puertas, Cerrando Heridas (Opening Doors, Closing Wounds): Latinas/os Finding Work-Life Balance in Academia is the newest book in the series on balancing work and life in the academy. This volume focuses on the experiences of Latina/o students, professors, and staff/administrators in higher education and documents their testimonios of achieving a sense of balance between their personal and professional lives. In the face of many challenges they are scattered across the country, are often working in isolation of each other and must find ways to develop their own networks, support structures, and spaces where they can share their wisdom, strategize, and forge alliances to ensure collective. The book focuses on Latinas/os in colleges of education, since many of them carry the important mission to prepare new teachers, and research new pedagogies that have the power of improving and transforming education. Following the format of the work-life balance book series, this volume contains autoethnographical testimonios in its methodological approach. This volume addresses three very important guiding questions (1) What are the existing structures that isolate/discriminate against Latinas/os in higher education? (2) How can Latinas/os disrupt these to achieve work-life balance? (3) Based on their experiences, what are the transformative ideologies regarding Latinas/os seeking work-life balance?

With the advancement of deep learning, road crack segmentation has become increasingly crucial for intelligent transportation safety. Despite notable progress, existing methods still face challenges in capturing fine-grained textures in small crack regions, handling blurred edges and significant width variations, and performing multi-class segmentation. Moreover, the high computational cost of training such models hinders their practical deployment. To tackle these limitations, we propose RSA-TransUNet, a novel model for road crack segmentation. At its core is the Axial-shift MLP Attention (ASMA) mechanism, which integrates axial perception with sparse contextual modeling. Through multi-path axial perturbations and an attention-guided structure, ASMA effectively captures long-range dependencies within row-column patterns, enabling detailed modeling of multi-scale crack features. To improve the model’s adaptability to structural irregularities, we introduce the Adaptive Spline Linear Unit (ASLU), which enhances the model’s capacity to represent nonlinear transformations. ASLU improves responsiveness to microstructural variations, morphological distortions, and local discontinuities, thereby boosting robustness across different domains. We further develop a Structure-aware Multi-stage Evolutionary Optimization (SMEO) strategy, which guides the training process through three phases: structural perception exploration, feature stability enhancement, and global perturbation. This strategy combines breadth sampling, convergence compression, and local escape mechanisms to improve convergence speed, global search efficiency, and generalization performance. Extensive evaluations on the Crack500, CFD, and DeepCrack datasets—including ablation studies and comparative experiments—demonstrate that RSA-TransUNet achieves superior segmentation accuracy and robustness in complex road environments, highlighting its potential for real-world applications.

Over the past few years, virtual and mixed reality systems have evolved significantly yielding high immersive experiences. Most of the metaphors used for interaction with the virtual environment do not provide the same meaningful feedback, to which the users are used to in the real world. This paper proposes a cyber-glove to improve the immersive sensation and the degree of embodiment in virtual and mixed reality interaction tasks. In particular, we are proposing a cyber-glove system that tracks wrist movements, hand orientation and finger movements. It provides a decoupled position of the wrist and hand, which can contribute to a better embodiment in interaction and manipulation tasks. Additionally, the detection of the curvature of the fingers aims to improve the proprioceptive perception of the grasping/releasing gestures more consistent to visual feedback. The cyber-glove system is being developed for VR applications related to real estate promotion, where users have to go through divisions of the house and interact with objects and furniture. This work aims to assess if glove-based systems can contribute to a higher sense of immersion, embodiment and usability when compared to standard VR hand controller devices (typically button-based). Twenty-two participants tested the cyber-glove system against the HTC Vive controller in a 3D manipulation task, specifically the opening of a virtual door. Metric results showed that 83% of the users performed faster door pushes, and described shorter paths with their hands wearing the cyber-glove. Subjective results showed that all participants rated the cyber-glove based interactions as equally or more natural, and 90% of users experienced an equal or a significant increase in the sense of embodiment.

This work presents an enriched human-machine interface for performing effective tele-manipulation using a humanoid robot. To provide the slave with the ability to dexterously interact with the remote environment, we implement an autonomous impedance controller that regulates the slave's joints stiffness and damping according to the manipulation loading conditions. In the proposed strategy, free-space operations are performed with compliant limbs to ensure safe interactions during unforeseen collisions on the whole-arm together with a soft behaviour at the initial phase of contact. During the manipulation task, the designed controller accounts for the arm configuration and the direction/amplitude of the external force sensed at the end-effector to stiffen the humanoid arm joints, permitting to handle the task loads. Experiments performed on the compliant humanoid robot COMAN demonstrate the effectiveness of this human-inspired impedance controller during a teleoperated pick-and-place task. The interaction forces during manipulation can destabilize the floating base humanoid robot. To address this concern we also investigate the operators ability to rely on a cutaneous feedback of the slave's balance state to adjust their teleoperation strategy when required so as to complete the task while maintaining the slave's balance. A comparative study shows that the proposed vibrotactile feedback allows for a proper recognition of the slave's drop of stability during interaction tasks and that the tactile guidance leads to enhanced teleoperation performances characterized by a significantly lower number of falls.

Robots designed for in-pipe navigation, inspection, and repair require flexibility for intricate pipeline traversal and the strength to carry payloads. However, conventional wheeled in-pipe robots face challenges in simultaneously achieving both substantial flexibility and payload-carrying capacity. A superior approach involves utilizing truss robots with redundant joints and linkages for pipe shape adaptation and actuation force distribution, providing significant advantages for complex pipeline navigation and heavy payload delivery. However, the kinematics of truss robots is computationally expensive for conventional Jacobian-based algorithms due to their complicated structural constraints. To address this limitation, we propose a novel algorithm for efficient truss-robot-kinematics computation using (GP) method. Our method computes both forward kinematics and Jacobian in a propagative manner. It also guarantees geometric constraints with the Sigmoid function as the boundary. In simulation experiments, our algorithm accelerates pipe shape adaptation computation by 5.2 <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\sim$</tex-math></inline-formula> 16.4 times compared to finite difference methods. The practical feasibility of our method is assessed through physical in-pipe crawling experiments using a truss robot prototype. Additionally, the prototype's ability to carry heavy payloads is demonstrated through payload-carrying experiments, which results in 2 <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\sim$</tex-math></inline-formula> 4 times heavier payload capacity compared to two-wheeled robot approaches. We also showcase the versatility of proposed method in addressing manipulation tasks, indicating its generalizability across diverse applications. We believe this work could provide a unique algorithmic framework for truss robot kinematics formulation and computation, which will enable next generation of in-pipe robots to be more adaptive to complex environments and formidable toward real-world applications.

Recent advancements in machine learning for robotics have led to notable improvements, achieving tremendous progress in learning robotic manipulation policies using reinforcement learning, behavior cloning, and other learning frameworks. However, existing approaches remain limited by their low sample efficiency, making real-world data collection a slow and resource-intensive process. This thesis addresses these limitations by leveraging the power of geometric symmetry in physical environments. The core contribution of this work is the development of a comprehensive framework for Equivariant Policy Learning, which exploits the inherent symmetries, such as rotations and translations, present in robotic tasks. Through this approach, the proposed methods achieve significantly improved sample efficiency, enabling faster learning with far less data compared to non-equivariant baselines. Specifically, this thesis first proposes a theoretical framework of Group-Invariant MDPs and theoretically demonstrates that symmetries in the environment (i.e., symmetries that exist in the transition and reward functions) are inherent in the solutions (i.e., the optimal Q-function and optimal policy function) of the MDP. Based on this theoretical insight, this thesis proposes equivariant reinforcement learning approaches that achieve orders of magnitude higher sample efficiency than state-of-the-art baselines. Second, this work theoretically analyzes the relationships between symmetries in the model and symmetries in the environment, and proposes novel definitions of correct, incorrect, and extrinsic equivariance relationships. This analysis bridges the gap between idealized theoretical settings and practical robotic deployment scenarios. Third, this thesis proposes the integration of Equivariant Policy Learning with Diffusion Models. This combination enables sample-efficient learning for long-horizon, complex manipulation tasks, where traditional methods often struggle due to their inability to reason effectively over extended sequences of actions. The methods proposed in this work are evaluated in both simulation and real-world robotic settings, demonstrating superior performance across a range of manipulation tasks. By incorporating these advanced techniques, this work establishes a new benchmark in efficient robotic manipulation.--Author's abstract

Model free reinforcement learning suffers from the high sampling complexity inherent to robotic manipulation or locomotion tasks. Most successful approaches typically use random sampling strategies which leads to slow policy convergence. In this paper we present a novel approach for efficient exploration that leverages previously learned tasks. We exploit the fact that the same system is used across many tasks and build a generative model for exploration based on data from previously solved tasks to improve learning new tasks. The approach also enables continuous learning of improved exploration strategies as novel tasks are learned. Extensive simulations on a robot manipulator performing a variety of motion and contact interaction tasks demonstrate the capabilities of the approach. In particular, our experiments suggest that the exploration strategy can more than double learning speed, especially when rewards are sparse. Moreover, the algorithm is robust to task variations and parameter tuning, making it beneficial for complex robotic problems.

Human-machine interfaces (HMIs) are essential tools facilitating effective interactions between users and intelligent systems. However, current wearable HMIs, particularly smart gloves, often face challenges in precisely recognizing subtle finger movements and delivering real-time tactile feedback, limiting their operational effectiveness and user immersion. To overcome these challenges, we propose an intelligent glove system integrating highly stretchable liquid metal-based strain sensors and programmable electromagnetic feedback actuators to realize closed-loop immersive human-machine interaction. The strain sensors achieve superior stretchability (over 200%), linear response, minimal hysteresis, and robust durability, effectively capturing complex hand movements. Gesture recognition, performed by a convolutional neural network, achieves an accuracy of 98% across common gestures, demonstrating exceptional precision and robustness. Furthermore, the glove integrates soft programmable electromagnetic actuators that generate an output force greater than 0.25 N, providing adjustable and precise tactile feedback via pulse-width modulation. This allows users to intuitively perceive tactile sensations in human-machine interaction scenarios, including robotic manipulation and motion guidance tasks. Comprehensive experiments confirm the intelligent glove system’s superior real-time responsiveness and adaptability in dynamic environments. This advanced closed-loop HMI glove significantly enhances operator immersion, operational accuracy, and interaction realism, demonstrating strong potential in robotics, virtual reality, and remote teleoperation applications.

Cleaning high-altitude building windows is traditionally a high-risk task for human workers due to the danger of falls. To enhance safety in facade cleaning and similar tasks, the development of window-cleaning robots has become increasingly essential. However, current robots face challenges in navigating exoskeleton structures due to obstructive beams, limiting their application. This article proposes a novel biped window-cleaning robotic system (BWCRS) designed for exoskeleton structures. The system comprises two key components: a multifunctional biped climbing robot (BCR) and a specialized cleaning robot. The BCR is adept at overcoming obstacles inherent in exoskeleton structures and transitioning across various inclined window surfaces. The BCR adopts a symmetrical construction, and the ends of the feet are equipped with vacuum suction cups and electromagnets, which can switch between different actuators at the geometric base and end of the system. The system is further enhanced by air pressure sensors that monitor the suction cup adhesion in real time, ensuring stable operation. A crucial functionality of the BCR is its manipulation mode, guided by a visual camera. This mode enables precise location and manipulation of the cleaning robot, ensuring effective coverage of all window areas. The mechanical design and mathematical modeling of the BWCRS are thoroughly examined, demonstrating its robust window-cleaning capabilities. The experimental results reveal that the proposed BCR has a favorable window-climbing and manipulation ability and that the BWCRS has a favorable cleaning effect on the exoskeleton-structure windows. Note to Practitioners—This paper is motivated by the problem of window-cleaning for exoskeleton structures, but it also applies to solving window maintenance and inspection. Previous window-cleaning robots were mainly aimed at the facades or surfaces with low obstacles that cannot be applied directly to exoskeleton-structured windows. This paper proposes an exoskeleton-structure window-cleaning method based on a biped window-climbing robot and a cleaning robot. The biped robot can climb over obstacles on the exoskeleton structures and manipulate the cleaning robot to different cleaning areas. Here, we introduce the biped window-climbing robot from the perspective of mechanical design and mathematical kinematics. We then demonstrate multiple obstacle-crossing motion modes of the biped window-climbing robot. Combined with visual-based manipulation, the window-cleaning process for the exoskeleton structures by a biped window-climbing robot integrated with a cleaning robot is demonstrated. The results verify that the proposed design can complete window-cleaning tasks on exoskeleton structures.

Passive radiative cooling, as a zero‐energy, environmentally friendly refrigeration technology, holds significant application potential in mitigating global warming and related fields. However, a critical bottleneck persists: the inherent trade‐off between its optical cooling performance and mechanical robustness, which severely limits its practical deployment. This paper proposes a one‐pot ball milling process that simultaneously refines, mixes, and chemically fuses microcrystalline cellulose (MCC) and silica (SiO 2 ) to fabricate hierarchical MCC/SiO 2 hybrid particles. The engineered fillers are then seamlessly integrated into a polydimethylsiloxane (PDMS) matrix, ultimately yielding flexible, mechanically robust, and highly efficient radiative cooling composite MCC/SiO 2 /PDMS film for synergistic enhancement of optical and mechanical properties. The optimized film exhibits a record‐high solar reflectance of 94.58% and a long‐wave infrared emissivity of 95.22%. Notably, it breaks the traditional performance trade‐off: the incorporation of MCC/SiO 2 enhances the mechanical strength and toughness of PDMS by 1.6 MPa and 1.02 MJ cm −3 . Outdoor tests demonstrate an average sub‐ambient cooling effect of 8°C under direct sunlight. This work provides a novel paradigm for designing sustainable high‐performance materials through microstructural engineering of hybrid fillers.

Metasurfaces have attracted substantial attention in recent years due to the ability for the versatile manipulation of electromagnetic (EM) wave. Recently, Janus backscattering mirrors that retroreflect incident waves for one side while show transparency in the time-reversed reflection channel have been discussed. Here, the dependence of the broadband performance on the geometry parameters and the angles are investigated. The numerical simulated results validate that the bandwidth and the efficiency depend on geometry parameters, especially the tilted angle and incident angle.

Test-retest reliability of joint position sense in the carpometacarpal joint among healthy adults

We quantify the accuracy of various simulators compared to a real world robotic reaching and interaction task. Simulators are used in robotics to design solutions for real world hardware without the need for physical access. The `reality gap' prevents solutions developed or learnt in simulation from performing well, or at all, when transferred to real-world hardware. Making use of a Kinova robotic manipulator and a motion capture system, we record a ground truth enabling comparisons with various simulators, and present quantitative data for various manipulation-oriented robotic tasks. We show the relative strengths and weaknesses of numerous contemporary simulators, highlighting areas of significant discrepancy, and assisting researchers in the field in their selection of appropriate simulators for their use cases.