Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A Framework for Joint Grasp and Motion Planning in Confined Spaces

Martin Rudorfer, Jiří Hartvich, Vojtěch Vonásek

TL;DR

The paper addresses grasping in confined spaces by introducing an open, end-to-end framework that jointly considers grasp and motion planning. It provides 20 benchmark scenarios (4 environments × 5 difficulty levels) with 200 annotated 6-DoF grasps each, plus two baseline planners and tools for evaluation, simulation, and scenario sharing. Experimental results show that different planners excel in different scenarios and that difficulty levels meaningfully affect performance, reinforcing the need for integrated approaches. By making all components publicly available, the framework aims to improve reproducibility and accelerate research in constrained-space robotic manipulation.

Abstract

Robotic grasping is a fundamental skill across all domains of robot applications. There is a large body of research for grasping objects in table-top scenarios, where finding suitable grasps is the main challenge. In this work, we are interested in scenarios where the objects are in confined spaces and hence particularly difficult to reach. Planning how the robot approaches the object becomes a major part of the challenge, giving rise to methods for joint grasp and motion planning. The framework proposed in this paper provides 20 benchmark scenarios with systematically increasing difficulty, realistic objects with precomputed grasp annotations, and tools to create and share more scenarios. We further provide two baseline planners and evaluate them on the scenarios, demonstrating that the proposed difficulty levels indeed offer a meaningful progression. We invite the research community to build upon this framework by making all components publicly available as open source.

A Framework for Joint Grasp and Motion Planning in Confined Spaces

TL;DR

The paper addresses grasping in confined spaces by introducing an open, end-to-end framework that jointly considers grasp and motion planning. It provides 20 benchmark scenarios (4 environments × 5 difficulty levels) with 200 annotated 6-DoF grasps each, plus two baseline planners and tools for evaluation, simulation, and scenario sharing. Experimental results show that different planners excel in different scenarios and that difficulty levels meaningfully affect performance, reinforcing the need for integrated approaches. By making all components publicly available, the framework aims to improve reproducibility and accelerate research in constrained-space robotic manipulation.

Abstract

Robotic grasping is a fundamental skill across all domains of robot applications. There is a large body of research for grasping objects in table-top scenarios, where finding suitable grasps is the main challenge. In this work, we are interested in scenarios where the objects are in confined spaces and hence particularly difficult to reach. Planning how the robot approaches the object becomes a major part of the challenge, giving rise to methods for joint grasp and motion planning. The framework proposed in this paper provides 20 benchmark scenarios with systematically increasing difficulty, realistic objects with precomputed grasp annotations, and tools to create and share more scenarios. We further provide two baseline planners and evaluate them on the scenarios, demonstrating that the proposed difficulty levels indeed offer a meaningful progression. We invite the research community to build upon this framework by making all components publicly available as open source.
Paper Structure (14 sections, 7 figures, 1 table)

This paper contains 14 sections, 7 figures, 1 table.

Table of Contents

  1. Introduction
  2. Related Works
  3. Grasp Planning
  4. Motion Planning
  5. Integrated Grasp and Motion Planning
  6. Benchmarks and Datasets
  7. Problem Formulation
  8. Framework description
  9. Overview
  10. Scenarios
  11. Grasp Annotations
  12. Baseline Planners
  13. Experiments
  14. Conclusions

Figures (7)

  • Figure 1: Four different environments: (a) grasping a screwdriver on a shelf, (b) grasping a ball underneath a table, (c) grasping a box through a narrow gap, (d) navigating through a narrow opening to grasp the banana. Each target object has a set of possible grasp candidates. The arm is mounted on a mobile base that can move within the grey area.
  • Figure 2: Visualization of the gripper frame. The y-axis (green) represents the grasp axis along which the fingers close, and the z-axis (blue) represents the gripper's approach direction.
  • Figure 3: Overview of our framework. All code is openly available and all relevant files for the benchmark scenarios are shared with the research community. Using our tools, other researchers can contribute their own scenarios in the same format.
  • Figure 4: Visualization of the difficulty levels in each environment. In environments 01 and 02, the object moves further into the confined space, while in environments 03 and 04 the constriction gets narrower.
  • Figure 5: Examples of trajectories found by the J${}^{+}$-RRT planner.
  • ...and 2 more figures