The inherent difficulty and limited scalability of collecting manipulation data using multi-fingered robot hand hardware platforms have resulted in severe data scarcity, impeding research on data-driven dexterous manipulation policy learning. To address this challenge, we present a hand-agnostic manipulation transfer system. It efficiently converts human hand manipulation sequences from demonstration videos into high-quality dexterous manipulation trajectories without requirements of massive training data. To tackle the multi-dimensional disparities between human hands and dexterous hands, as well as the challenges posed by high-degree-of-freedom coordinated control of dexterous hands, we design a progressive transfer framework: first, we establish primary control signals for dexterous hands based on kinematic matching; subsequently, we train residual policies with action space rescaling and thumb-guided initialization to dynamically optimize contact interactions under unified rewards; finally, we compute wrist control trajectories with the objective of preserving operational semantics. Using only human hand manipulation videos, our system automatically configures system parameters for different tasks, balancing kinematic matching and dynamic optimization across dexterous hands, object categories, and tasks. Extensive experimental results demonstrate that our framework can automatically generate smooth and semantically correct dexterous hand manipulation that faithfully reproduces human intentions, achieving high efficiency and strong generalizability with an average transfer success rate of 73%, providing an easily implementable and scalable method for collecting robot dexterous manipulation data. Refer to the arXiv version for the appendix.

Dexterous grasping and manipulation with multifingered robotic hands presents a significant challenge due to their high degrees of freedom and the need for task-specific adaptations. Existing methods usually adopt single-task learning framework or focus on simple stable wrap grasping, limiting their efficiency and generalization ability when encountering new task or precise functional grasping pose. In this article, we introduce MetaGrasp, a novel approach that defines dexterous functional grasping as a multitask reinforcement learning (RL) problem based on hand grasp pose classification. Our method features a unique gradual skill curriculum learning (GSCL) framework, which structures the learning process into three stages: beginner, intermediate, and advanced curriculum learning according to the level of difficulty. MetaGrasp leverages this hierarchical learning structure to develop a versatile, adaptive grasping policy that can grasp objects based on hand grasp pose and object point cloud inputs. Taking five hand grasp types as research cases, the trained policy with our MetaGrasp can be easily adpated to grasp different object instances from different object categories according to functional grasp intentions specified by one expert demonstration without requiring extensive system interaction. We categorize the dexterous functional grasping tasks of a five-fingered robotic hand into multiple tasks based on hand poses for RL, and to combine meta imitation learning (IL) with curriculum learning. The experimental results show that the MetaGrasp has better one-shot generalization ability on new grasp tasks, and outperforms state-of-the-art single-task dexterous grasping methods.

Humans achieve contact-rich dexterous grasping through the synergy of visual and tactile information. However, the high-dimensional action space of high DoF multi-fingered hands poses significant challenges to this operation. In this study, we address this complexity by controlling the robotic hand at the reduced dimensional level of individual fingers instead of the entire hand, and develop a finger-based multi-agent deep reinforcement learning strategy by regarding the wrist, arm, and each finger of the hand as intelligent agents. We commence by applying a single-agent reinforcement learning algorithm to guide the whole hand to reach the feasible approaching direction and distance to the object. Then, we develop neuroscience-inspired visuo-tactile fusion networks to train multiple agents to control their assigned fingers by effectively leveraging visual and tactile feedback. This enables dynamic and collaborative adjustments of finger-object interactions, ultimately achieving precise contact with specific areas of the objects. The grasping results on 8 objects show that our approach can achieve stable and compliant grasps. To the best of our knowledge, this is the first work that employs a finger-based multi-agent reinforcement learning approach to control the dexterous grasping process under the guidance of both visual and tactile feedback.

An efficient 3D spatial spectrum prediction method is essential for UAV networks operating in highly heterogeneous spectrum environments, enabling UAVs to proactively navigate toward areas with abundant available spectrum and make decisions to access to idle ones in advance. This paper introduces a novel Multi-Scale Temporal model for 3D spatial spectrum prediction (MST-3DSSP) that comprehensively captures complex correlations across 3D spatial, frequency, and multi-scale temporal domains. Specifically, the proposed model incorporates a 3D Spatial-Frequency Fusion (3DS-FF) module to extract and fuse 3D spatial and frequency features, along with a MultiScale Temporal Extraction (MS-TE) module that combines BiLSTM and Transformer blocks to capture both small scale and large scale temporal dependencies. These two modules enable the model to understand the complex correlations across 3D spatial, frequency, and multi-scale temporal domains, thereby allowing for more accurate spectrum predictions. Extensive experiments on real-world spectrum datasets demonstrate that MST-3DSSP significantly outperforms existing spectrum prediction methods, achieving higher prediction accuracy and reduced errors, thus providing a robust solution for improving spectrum efficiency in UAV networks.

Integrated sensing and communication (ISAC) technology has been envisioned to revolutionize the future intelligent vehicle networks (IVNs). Recently, due to the mobility and flexible deployment, unmanned aerial vehicle (UAV) has been regarded as a promising aerial ISAC platform in future IVNs. In this paper, comprehensively considering all the performance on the throughput, sensing accuracy, and sensing rate, we propose a Multi-Agent joint Trajectory control and RadCom task schedule (MA-TRC) scheme for the ISAC-UAV assisted IVN. To make UAVs achieve the optimal decisions autonomously and adaptively, we propose a Game-Embedding Multi-Agent Deep Reinforcement Learning (GE-MADRL) approach. Specifically, considering the complex action space with both discrete and continuous decision variables of UAVs, we develop a multi-agent Parametrized deep Q-network (MAPDQN) based solution, which can help UAVs learn the dynamic and uncertain environment to adaptively obtain the MA-TRC scheme. Furthermore, since UAVs work in a distributed decision making manner, the potential conflicting decisions will impact the network performance. To avoid the decision conflicts among UAVs during the network parameter training, a distributed two-stage Game method is designed as an action adjuster embedded in MAPDQN, by which the learning convergence performance will be further improved and the strategy conflicts can be avoided.

Robot grasp dataset is the basis of designing the robot's grasp generation model. Compared with the building grasp dataset for Low-DOF grippers, it is harder for High-DOF dexterous robot hand. Most current datasets meet the needs of generating stable grasps, but they are not suitable for dexterous hands to complete human-like functional grasp, such as grasp the handle of a cup or pressing the button of a flashlight, so as to enable robots to complete subsequent functional manipulation action autonomously, and there is no dataset with functional grasp pose annotations at present. This paper develops a unique Cost-Effective Real-Simulation Annotation System by leveraging natural hand's actions. The system is able to capture a functional grasp of a dexterous hand in a simulated environment assisted by human demonstration in real world. By using this system, dexterous grasp data can be collected efficiently as well as cost-effective. Finally, we construct the first dexterous functional grasp dataset with rich pose annotations. A Functional Grasp Synthesis Model is also provided to validate the effectiveness of the proposed system and dataset. Our project page is: https://hjlllll.github.io/DFG/.

Intelligent robotic manipulation is a challenging study of machine intelligence. Although many dexterous robotic hands have been designed to assist or replace human hands in executing various tasks, how to teach them to perform dexterous operations like human hands is still a challenge. This motivates us to conduct an in-depth analysis of human behavior in manipulating objects and propose an object-hand manipulation representation. This representation provides an intuitive and clear semantic indication of how the dexterous hand should touch and manipulate an object based on the object's own functional areas. At the same time, we propose a functional grasp synthesis framework, which does not require real grasp label supervision, but relies on the guidance of our object-hand manipulation representation. In addition, in order to obtain better functional grasp synthesis results, we propose a network pre-training method that can make full use of easily obtained stable grasp data, and a network training strategy to coordinate the loss functions. We conduct object manipulation experiments on a real robot platform, and evaluate the performance and generalization of our object-hand manipulation representation and grasp synthesis framework.

Estimating complete human and hand silhouettes from a monocular RGB image has important roles in many visual applications such as image editing and action recognition. However, complicated scenes, non-rigid articulated deformations and occlusions make it a difficult task. To overcome the difficulties of inferring the invisible part of the human body due to objects’ occlusion, we propose a pose guided two stage deep neural network model. It leverages the key points semantic prior knowledge to help the model envision occluded areas, which is more in line with human’s visual recognition characteristics. The proposed model is validated on both hand and human datasets with varied occlusions. The experimental results are satisfying and better than the state-of-the-art methods.

In recent years, many dexterous robotic hands have been designed to assist or replace human hands in executing various tasks. But how to teach them to perform dexterous operations like human hands is still a challenging task. In this paper, we propose a grasp synthesis framework to make robots grasp and manipulate objects like human beings. We first build a dataset by accurately segmenting the functional areas of the object and annotating semantic touch code for each functional area to guide the dexterous hand to complete the functional grasp and post-grasp manipulation. This dataset contains 18 categories of 129 objects selected from four datasets, and 15 people participated in data annotation. Then we carefully design four loss functions to constrain the model, which successfully generates the functional grasp of dexterous hand under the guidance of semantic touch code. The thorough experiments in synthetic data show our model can robustly generate functional grasp, even for objects that the model has not see before.

This paper focuses on the topic of vision based hand pose estimation from single depth map using convolutional neural network (CNN). Our main contributions lie in designing a new pose regression network architecture named CrossInfoNet. The proposed CrossInfoNet decomposes hand pose estimation task into palm pose estimation sub-task and finger pose estimation sub-task, and adopts two-branch cross-connection structure to share the beneficial complementary information between the sub-tasks. Our work is inspired by multi-task information sharing mechanism, which has been few discussed in hand pose estimation using depth data in previous publications. In addition, we propose a heat-map guided feature extraction structure to get better feature maps, and train the complete network end-to-end. The effectiveness of the proposed CrossInfoNet is evaluated with extensively self-comparative experiments and in comparison with state-of-the-art methods on four public hand pose datasets. The code is available.