3 July 2025 in IEEE Robotics and Automation Letters
Dexterous grasping is a crucial technique in humanoid robot manipulation. However, existing methods still fall short in effectively detecting dexterous grasps in cluttered environments. In this work, we propose DexMGNet, a novel multi-mode dexterous grasping framework designed for such challenging scenarios. We introduce the concept of pre-grasping and redefine dexterous grasping to enhance adaptability. We propose an effective pre-grasp and grasp data sampling strategy and develop a conditional generative model for grasp and pre-grasp generation. Additionally, we integrate pre-grasp collision detection within the hand's workspace, significantly improving grasping performance in cluttered environments. Our method supports multi-mode grasping, including two-finger, three-finger, and four-finger grasps, enabling greater flexibility across diverse grasping tasks. In real-world desktop grasping experiments, our approach achieves a 93.3% success rate in single-object scenes and a 78.3% success rate in multi-object scenes, demonstrating its effectiveness and superiority.
18 November 2025 in The International Journal of Robotics Research
Seven-degree-of-freedom offset manipulators are typically large-scale space robot systems featuring two long links. The offsets at the shoulder, elbow, and wrist joints extend the joint motion range, thereby enhancing the manipulator’s workspace and dexterity to perform more complex missions. The two long links render the physical significance of the manipulator’s elbow highly pronounced, with the arm angle being closely coupled to the elbow joint. Arm angle parameterization is well-suited for configuration control of such manipulators; however, offsets pose a geometric challenge that hinders analytical redundancy resolution. This article focuses on the Experimental Module Manipulator (EMM) of the China Space Station and investigates its nullspace redundancy resolution approach via arm angle self-adaptation. First, a novel semi-analytical inverse kinematics solution via arm angle parameterization is proposed for the EMM, offering high computational accuracy and fast solving speed. Moreover, convex hulls of the EMM and the China Space Station are modeled for high-precision collision detection based on the Gilbert–Johnson–Keerthi distance algorithm. Subsequently, an obstacle avoidance strategy via arm angle self-adaptation is developed for the EMM to perform tasks safely in complex environments. Additionally, the effectiveness and practicality of both the arm angle parameterized inverse kinematics solution and the obstacle avoidance strategy are verified through comprehensive simulations. Finally, this research is applied to the EMM for successful accomplishment of on-orbit servicing to support the Shenzhou-16 astronaut’s extravehicular activities aboard the China Space Station, thereby achieving on-orbit experimental validation.
9 March 2026 in Knowledge-Based Systems
13 June 2025 in IEEE Robotics and Automation Letters
Motion generation in dynamic environments is crucial for human-machine interaction with redundant manipulators. In this context, we propose the Enhancing Safety and Manipulability (ESM) scheme, which integrates geometry-based dynamic obstacle avoidance, manipulability optimization, trajectory tracking, and joint limit avoidance into a unified scheme operating at the joint-angle level. The introduction of a flexible collision library enables the scheme to locate critical points based on geometry, while the incorporated obstacle speed allows the scheme to effectively avoid dynamic obstacles. In the ESM, manipulability is naturally set as the non-convex goal. To solve the ESM, the Accelerated Multi-agent Optimizer (AMO) is proposed, which adopts a neural network-inspired architecture and uses a meta-heuristic approach to construct activation functions, endowing the optimizer with non-convex control capabilities. Subsequently, a GPU-based parallel computing method is implemented, significantly reducing computing time. Detailed simulations, experiments, and comparisons demonstrate the framework's effectiveness and superiority.
2 September 2025 in IEEE Robotics and Automation Letters
Utilizing neural networks to predict potential regions containing optimal paths in advance and subsequently biasing the sampling probability towards these promising regions has been proven to effectively enhance the path planning efficiency of sampling-based algorithms. Undoubtedly, the accuracy of the promising regions is of paramount importance. Currently, the generalizability of many CNN- or Transformer-based models remains limited, often performing poorly in unknown environments. To enhance generalization capability, we reformulate the promising region prediction problem as a conditional generation task and address it using a diffusion model, referred to as the DiffRP (Diffusion-based Region Prediction). We propose three paradigms for generating promising regions, among which we innovatively introduce a biased noise initialization method for the diffusion process. Specifically, we bias the mean of the noise distribution using obstacle maps and design a map-conditioned denoising model to progressively generate accurate promising regions from the biased noise. Experiments on public datasets demonstrate that our proposed DiffRP method outperforms existing state-of-the-art models by 35$\sim$42% in promising region prediction accuracy. Moreover, the non-uniform sampling algorithm (DiffRP-RRT*) based on this region achieves a 3$\sim$52% reduction in sample number compared with other neural-network-driven approaches.