Real-Time Grasp Planning for Multi-Fingered Hands by Finger Splitting

Abstract

Grasp planning for multi-fingered hands is computationally expensive due to the joint-contact coupling, surface nonlinearities and high dimensionality, thus is generally not affordable for real-time implementations. Traditional planning methods by optimization, sampling or learning work well in planning for parallel grippers but remain challenging for multifingered hands. This paper proposes a strategy called finger splitting, to plan precision grasps for multi-fingered hands starting from optimal parallel grasps. The finger splitting is optimized by a dual-stage iterative optimization including a contact point optimization (CPO) and a palm pose optimization (PPO), to gradually split fingers and adjust both the contact points and the palm pose. The dual-stage optimization is able to consider both the object grasp quality and hand manipulability, address the nonlinearities and coupling, and achieve efficient convergence within one second. Simulation results demonstrate the effectiveness of the proposed approach. The simulation video is available at [1].

Grasp Planning for Customized Grippers by Iterative Surface Fitting

“…The palm optimization optimizes for the optimal palm transformation R * , t * with fixed δθ j , while the finger optimization solves for the optimal finger displacement δθ j with fixed (R, t). The derivation of IPFO for multi-fingered hands is similar with our previous work on finger splitting [17]. Compared with one DOF case, the finger optimization for multi-fingered Algorithm 3 Grasp Planning Algorithm 1: Input: ∂O, ∂F, center# K, sample# K s , d 0 2: Init: C ← k-means(∂O, K),regret = 0 K ,trial = 0 K 3: for It = 1, · · · , K s do 4: Guided sampling:…”

Multifingered Grasping Based on Multimodal Reinforcement Learning

“…A lot of previous work has discussed the research on multifingered grasping. Li et al [7] proposed a probabilistic model to address robust dexterous grasping under shape uncertainty, and Fan et al [8] proposed a finger splitting strategy to plan precision grasps for multifingered hands from parallel grasps. However, both examples require prior knowledge of the object model.…”

A Learning Framework for Robust Bin Picking by Customized Grippers

“…More specifically, RCNN-ISF searches ROI in the two-dimensional image plane, while the end-to-end learning searches over higher dimension depending on the grasps and grippers. For example, a grasp planning for a eight-DOF hand with three fingers has 32 dimensions [21]. Therefore, the end-to-end learning requires much more data than the proposed method.…”