Search Results (45)

Dexterous bimanual manipulation remains a challenging task in reinforcement learning (RL) due to the vast state–action space and the complex interdependence between the hands. Conventional end-to-end learning struggles to handle this complexity, and multi-agent RL often faces limitations in stably acquiring cooperative movements. To address these issues, this study proposes a hierarchical progressive policy learning framework for dexterous bimanual manipulation. In the proposed method, one hand’s policy is first trained to stably grasp the object, and, while maintaining this grasp, the other hand’s manipulation policy is progressively learned. This hierarchical decomposition reduces the search space for each policy and enhances both the connectivity and the stability of learning by training the subsequent policy on the stable states generated by the preceding policy. Simulation results show that the proposed framework outperforms conventional end-to-end and multi-agent RL approaches. The proposed method was demonstrated via sim-to-real transfer on a physical dual-arm platform and empirically validated on a bimanual cube manipulation task. Full article

Contact-force perception is a critical component of safe robotic grasping. With the rapid advances in embodied intelligence technology, humanoid robots have enhanced their multimodal perception capabilities. Conventional force sensors face limitations, such as complex spatial arrangements, installation challenges at multiple nodes, and potential interference with robotic flexibility. Consequently, these conventional sensors are unsuitable for biomimetic robot requirements in object perception, natural interaction, and agile movement. Therefore, this study proposes a sensorless external force detection method that integrates SuperPoint-Scale Invariant Feature Transform (SIFT) feature extraction with finite element analysis to address force perception challenges. A visual analysis method based on the SuperPoint-SIFT feature fusion algorithm was implemented to reconstruct a three-dimensional displacement field of the target object. Subsequently, the displacement field was mapped to the contact force distribution using finite element modeling. Experimental results demonstrate a mean force estimation error of 7.60% (isotropic) and 8.15% (anisotropic), with RMSE < 8%, validated by flexible pressure sensors. To enhance the model’s reliability, a dual-channel video comparison framework was developed. By analyzing the consistency of the deformation patterns and mechanical responses between the actual compression and finite element simulation video keyframes, the proposed approach provides a novel solution for real-time force perception in robotic interactions. The proposed solution is suitable for applications such as precision assembly and medical robotics, where sensorless force feedback is crucial. Full article

This paper presents the design and implementation of an autonomous mobile robot system able to deliver objects from one location to another with minimal hardware requirements. Unlike most existing systems, our robot uses only a single camera—mounted on its robotic arm—to guide both its movements and the pick-and-place process. The robot detects target signs and objects, automatically navigates to desired locations, and accurately grasps and delivers items without the need for complex sensor arrays or multiple cameras. The main innovation of this work is a unified visual control strategy that coordinates both the vehicle and the robotic arm through homography-based visual servoing. Our experimental results demonstrate that the system can reliably locate, pick up, and place objects, achieving a high success rate in real-world tests. This approach offers a simple yet effective solution for object delivery tasks and lays the groundwork for practical, cost-efficient mobile robots in automation and logistics. Full article

The first web space of the hand plays a fundamental role in daily hand function, facilitating crucial movements, such as pinching, grasping, and opposition. The structural anomalies of acquired defects of this anatomical region, whether secondary to trauma, burns, or post-oncological surgical resections, necessitate meticulous reconstructive strategies to ensure both functional restoration and aesthetic integrity. Given the complexity and variability of first web defects, a broad spectrum of reconstructive techniques has been developed, ranging from skin grafting and local flap reconstructions to advanced microsurgical approaches. This review comprehensively examines the existing literature on first web reconstruction techniques, analyzing their indications, advantages, and limitations. Additionally, it explores innovative techniques and emerging trends in the field, such as tissue engineering, regenerative medicine, and composite tissue allotransplantation, which may revolutionize future reconstructive strategies. The primary objective is to provide clinicians with an evidence-based guide to selecting the most appropriate reconstructive strategy tailored to individual patient needs. Furthermore, we incorporate our institutional experience in managing first web defects, highlighting key surgical principles, patient outcomes, and challenges encountered. Through this analysis, we aim to refine the understanding of first web reconstruction and contribute to the ongoing evolution of hand surgery techniques. Full article

The aging population has created a significant challenge affecting the world; social and healthcare systems need to ensure elderly individuals receive the necessary care services to improve their quality of life and maintain their independence. In response to this need, developing integrated digital solutions, such as IoT based wearable devices combined with artificial intelligence applications, offers a technological platform for creating Ambient Intelligence (AI) and Assisted Living (AAL) environments. These advancements can help reduce hospital admissions and lower healthcare costs. In this context, this article presents an IoT application based on MEMS (micro electro-mechanical systems) sensors integrated into a state-of-the-art microcontroller (STM55WB) for recognizing the movements of older individuals during daily activities. human activity recognition (HAR) is a field within computational engineering that focuses on automatically classifying human actions through data captured by sensors. This study has multiple objectives: to recognize movements such as grasping, leg flexion, circular arm movements, and walking in order to assess the motor skills of older individuals. The implemented system allows these movements to be detected in real time, and transmitted to a monitoring system server, where healthcare staff can analyze the data. The analysis methods employed include machine learning algorithms to identify movement patterns, statistical analysis to assess the frequency and quality of movements, and data visualization to track changes over time. These approaches enable the accurate assessment of older people’s motor skills, and facilitate the prompt identification of abnormal situations or emergencies. Additionally, a user-friendly technological solution is designed to be acceptable to the elderly, minimizing discomfort and stress associated with using technology. Finally, the goal is to ensure that the system is energy-efficient and cost-effective, promoting sustainable adoption. The results obtained are promising; the model achieved a high level of accuracy in recognizing specific movements, thus contributing to a precise assessment of the motor skills of the elderly. Notably, movement recognition was accomplished using an artificial intelligence model called Random Forest. Full article

This study presents a topological design and modeling framework for 3D-printed robotic grippers, tailored for combined precision and coarse robotics assembly. The proposed methodology leverages topology optimization to develop multi-scale-compliant mechanisms, comprising a symmetrical continuum structure of five beams. The proposed methodology centers on the hybrid kinematics for precision and coarse operations of the gripper, parametrizing beam deformations in response to a defined set of boundary conditions and varying input loads. The research employs topology analysis to draw a clear correlation between input load and resultant motion, with a particular emphasis on the mechanism’s capacity to integrate both fine and coarse movements efficiently. Additionally, the paper pioneers an innovative solution to the ubiquitous point-contact problem encountered in grasping, intricately weaving it with the stiffness matrix. The overarching aim remains to provide a streamlined design methodology, optimized for manufacturability, by harnessing the capabilities of contemporary 3D fabrication techniques. This multifaceted approach, underpinned by the multiscale grasping method, promises to significantly advance the domain of robotic gripping and manipulation across applications such as micro-assembly, biomedical manipulation, and industrial robotics. Full article

Grasping objects, from simple tasks to complex fine motor skills, is a key component of our daily activities. Our approach to facilitate the development of advanced prosthetics, robotic hands and human–machine interaction systems consists of collecting and combining surface electromyography (EMG) signals and contextual data of individuals performing manipulation tasks. In this context, the identification of patterns and prediction of hand grasp types is crucial, with cylindrical grasp being one of the most common and functional. Traditional approaches to grasp prediction often rely on unimodal data sources, limiting their ability to capture the complexity of real-world scenarios. In this work, grasp prediction models that integrate both EMG signals and contextual (task- and product-related) information have been explored to improve the prediction of cylindrical grasps during reaching movements. Three model architectures are presented: an EMG processing model based on convolutions that analyzes forearm surface EMG data, a fully connected model for processing contextual information, and a hybrid architecture combining both inputs resulting in a multimodal model. The results show that context has great predictive power. Variables such as object size and weight (product-related) were found to have a greater impact on model performance than task height (task-related). Combining EMG and product context yielded better results than using each data mode separately, confirming the importance of product context in improving EMG-based models of grasping. Full article

The COVID-19 pandemic brought telepresence systems into the spotlight, yet manually controlling remote robots often proves ineffective for handling complex manipulation tasks. To tackle this issue, we present a machine learning-based assistive manipulation approach. This method identifies target objects and computes an inverse kinematic solution for grasping them. The system integrates the generated solution with the user’s arm movements across varying inverse kinematic (IK) fusion levels. Given the importance of maintaining a sense of body ownership over the remote robot, we examine how haptic feedback and assistive functions influence ownership perception and task performance. Our findings indicate that incorporating assistance and haptic feedback significantly enhances the control of the robotic arm in telepresence environments, leading to improved precision and shorter task completion times. This research underscores the advantages of assistive manipulation techniques and haptic feedback in advancing telepresence technology. Full article

Visual servo grasping technology has garnered significant attention in intelligent manufacturing for its potential to enhance both the flexibility and precision of robotic operations. However, traditional approaches frequently encounter challenges such as task failure when visual features move outside the camera’s field of view (FoV) and system instability due to interaction matrix singularities, limiting the technology’s effectiveness in complex environments. This study introduces a novel control strategy that leverages an asymmetric time-varying performance function to address the issue of visual feature escape. By strictly limiting the range of feature error, our approach ensures that visual features consistently remain within the camera’s FoV, thereby enhancing both transient and steady-state system performance. Furthermore, we have developed an adaptive damped least squares controller that dynamically adjusts the damping term to mitigate numerical instability resulting from interaction matrix singularities. The effectiveness of our method has been validated through grasping experiments involving significant rotations around the camera’s optical axis and other complex movements. Full article

The aim of this contribution is to present a segmentation method for the identification of voluntary movements from inertial data acquired through a single inertial measurement unit placed on the subject’s wrist. Inertial data were recorded from 25 healthy subjects while performing 75 consecutive reach-to-grasp movements. The approach herein presented, called DynAMoS, is based on an adaptive thresholding step on the angular velocity norm, followed by a statistics-based post-processing on the movement duration distribution. Post-processing aims at reducing the number of erroneous transitions in the movement segmentation. We assessed the segmentation quality of this method using a stereophotogrammetric system as the gold standard. Two popular methods already presented in the literature were compared to DynAMoS in terms of the number of movements identified, onset and offset mean absolute errors, and movement duration. Moreover, we analyzed the sub-phase durations of the drinking movement to further characterize the task. The results show that the proposed method performs significantly better than the two state-of-the-art approaches (i.e., percentage of erroneous movements = 3%; onset and offset mean absolute error < 0.08 s), suggesting that DynAMoS could make more effective home monitoring applications for assessing the motion improvements of patients following domicile rehabilitation protocols. Full article

In this research, five systems were developed to classify four distinct motor functions—forward hand movement (FW), grasp (GP), release (RL), and reverse hand movement (RV)—from EEG signals, using the WAY-EEG-GAL dataset where participants performed a sequence of hand movements. During preprocessing, band-pass filtering was applied to remove artifacts and focus on the mu and beta frequency bands. The initial system, a preliminary study model, explored the overall framework of EEG signal processing and classification, utilizing time-domain features such as variance and frequency-domain features such as alpha and beta power, with a KNN model for classification. Insights from this study informed the development of a baseline system, which innovatively combined the common spatial patterns (CSP) method with continuous wavelet transform (CWT) for feature extraction and employed a GoogLeNet classifier with transfer learning. This system classified six unique pairs of events derived from the four motor functions, achieving remarkable accuracy, with the highest being 99.73% for the GP–RV pair and the lowest 80.87% for the FW–GP pair in intersubject classification. Building on this success, three additional systems were developed for four-way classification. The final model, ML-CSP-OVR, demonstrated the highest intersubject classification accuracy of 78.08% using all combined data and 76.39% for leave-one-out intersubject classification. This proposed model, featuring a novel combination of CSP-OVR, CWT, and GoogLeNet, represents a significant advancement in the field, showcasing strong potential as a general system for motor imagery (MI) tasks that is not dependent on the subject. This work highlights the prominence of the research contribution by demonstrating the effectiveness and robustness of the proposed approach in achieving high classification accuracy across different motor functions and subjects. Full article

The kinematics of the human hand exhibit complex and diverse characteristics unique to each individual. Various techniques such as vision-based, ultrasonic-based, and data-glove-based approaches have been employed to analyze human hand movements. However, a critical challenge remains in efficiently analyzing and classifying hand grasp types based on time-series kinematic data. In this paper, we propose a novel sparse coding feature extraction technique based on dictionary learning to address this challenge. Our method enhances model accuracy, reduces training time, and minimizes overfitting risk. We benchmarked our approach against principal component analysis (PCA) and sparse coding based on a Gaussian random dictionary. Our results demonstrate a significant improvement in classification accuracy: achieving 81.78% with our method compared to 31.43% for PCA and 77.27% for the Gaussian random dictionary. Furthermore, our technique outperforms in terms of macro-average F1-score and average area under the curve (AUC) while also significantly reducing the number of features required. Full article

Pea plants (Pisum sativum L.) can perceive the presence of potential supports in the environment and flexibly adapt their behavior to clasp them. How pea plants control and perfect this behavior during growth remains unexplored. Here, we attempt to fill this gap by studying the movement of the apex and the tendrils at different leaves using three-dimensional (3D) kinematical analysis. We hypothesized that plants accumulate information and resources through the circumnutation movements of each leaf. Information generates the kinematical coordinates for the final launch towards the potential support. Results suggest that developing a functional approach to grasp movement may involve an interactive trial and error process based on continuous cross-talk across leaves. This internal communication provides evidence that plants adopt plastic responses in a way that optimally corresponds to support search scenarios. Full article

Background: A fundamental grasp of the variability observed in healthy individuals holds paramount importance in the investigation of neuropsychiatric conditions characterized by sex-related phenotypic distinctions. Functional magnetic resonance imaging (fMRI) serves as a meaningful tool for discerning these differences. Among deep learning models, graph neural networks (GNNs) are particularly well-suited for analyzing brain networks derived from fMRI blood oxygen level-dependent (BOLD) signals, enabling the effective exploration of sex differences during adolescence. Method: In the present study, we introduce a multi-modal graph isomorphism network (MGIN) designed to elucidate sex-based disparities using fMRI task-related data. Our approach amalgamates brain networks obtained from multiple scans of the same individual, thereby enhancing predictive capabilities and feature identification. The MGIN model adeptly pinpoints crucial subnetworks both within and between multi-task fMRI datasets. Moreover, it offers interpretability through the utilization of GNNExplainer, which identifies pivotal sub-network graph structures contributing significantly to sex group classification. Results: Our findings indicate that the MGIN model outperforms competing models in terms of classification accuracy, underscoring the benefits of combining two fMRI paradigms. Additionally, our model discerns the most significant sex-related functional networks, encompassing the default mode network (DMN), visual (VIS) network, cognitive (CNG) network, frontal (FRNT) network, salience (SAL) network, subcortical (SUB) network, and sensorimotor (SM) network associated with hand and mouth movements. Remarkably, the MGIN model achieves superior sex classification accuracy when juxtaposed with other state-of-the-art algorithms, yielding a noteworthy 81.67% improvement in classification accuracy. Conclusion: Our model’s superiority emanates from its capacity to consolidate data from multiple scans of subjects within a proven interpretable framework. Beyond its classification prowess, our model guides our comprehension of neurodevelopment during adolescence by identifying critical subnetworks of functional connectivity. Full article

Alleviating the burden on amputees in terms of high-level control of their prosthetic devices is an open research challenge. EMG-based intention detection presents some limitations due to movement artifacts, fatigue, and stability. The integration of exteroceptive sensing can provide a valuable solution to overcome such limitations. In this paper, a novel semiautonomous control system (SCS) for wrist–hand prostheses using a computer vision system (CVS) is proposed and validated. The SCS integrates object detection, grasp selection, and wrist orientation estimation algorithms. By combining CVS with a simulated EMG-based intention detection module, the SCS guarantees reliable prosthesis control. Results show high accuracy in grasping and object classification (≥97%) at a fast frame analysis frequency (2.07 FPS). The SCS achieves an average angular estimation error ≤18° and stability ≤0.8° for the proposed application. Operative tests demonstrate the capabilities of the proposed approach to handle complex real-world scenarios and pave the way for future implementation on a real prosthetic device. Full article