Goal directed multi-finger manipulation: Control policies and analysis

  title={Goal directed multi-finger manipulation: Control policies and analysis},
  author={Sheldon Andrews and Paul G. Kry},
  journal={Comput. Graph.},

Figures and Tables from this paper

Catching and Throwing Control of a Physically Simulated Hand
A nominal controller for animating an articulated physics-based human arm model, including the hands and fingers, to catch and throw objects is designed and training agents are used to improve the robustness of the nominal controller.
Relaxed-rigidity constraints: kinematic trajectory optimization and collision avoidance for in-grasp manipulation
The method to perform in-grasp manipulation uses kinematic trajectory optimization which requires no knowledge of dynamic properties of the object, and is general enough to generate motions for most objects the robot can grasp.
Relaxed-Rigidity Constraints: In-Grasp Manipulation using Purely Kinematic Trajectory Optimization
A novel approach to performing in-grasp manipulation planning: the problem of moving an object with reference to the palm from an initial pose to a goal pose without breaking or making contacts, using kinematic trajectory optimization.
Applications of machine learning in sensorimotor control
This thesis focuses on methods for learning to control systems without prior knowledge of the dynamics of the system or its environment, and presents two algorithms that combine Gaussian process regression and stochastic gradient descent and addresses the trajectory-tracking problem for dynamical systems.
Towards learning hierarchical skills for multi-phase manipulation tasks
This paper presents an approach for exploiting the phase structure of tasks in order to learn manipulation skills more efficiently and was successfully evaluated on a real robot performing a bimanual grasping task.
A Comparison of Action Spaces for Learning Manipulation Tasks
This paper compares learning performance across three tasks, four action spaces, and using two modern reinforcement learning algorithms to lend support to the hypothesis that learning references for a task-space impedance controller significantly reduces the number of samples needed to achieve good performance across all tasks and algorithms.
Deep Dynamics Models for Learning Dexterous Manipulation
It is shown that improvements in learned dynamics models, together with improvements in online model-predictive control, can indeed enable efficient and effective learning of flexible contact-rich dexterous manipulation skills -- and that too, on a 24-DoF anthropomorphic hand in the real world, using just 4 hours of purely real-world data to learn to simultaneously coordinate multiple free-floating objects.
Generalization in Dexterous Manipulation via Geometry-Aware Multi-Task Learning
It is shown that a single generalist policy can perform in-hand manipulation of over 100 geometrically-diverse realworld objects and generalize to new objects with unseen shape or size and it is found that multi-task learning with object point cloud representations not only generalizes better but even outperforms the single-object specialist policies on both training as well as held-out test objects.
Learning to Use Chopsticks in Diverse Styles
This paper focuses on chopsticks-based object relocation tasks, which are common yet demanding, and automatically discovers physically valid chopsticks holding poses by Bayesian Optimization and Deep Reinforcement Learning, which works for multiple gripping styles and hand morphologies without the need of example data.


Policies for Goal Directed Multi-Finger Manipulation
A method for one-handed task based manipulation of objects using a mid-level multiphase approach to break the problem into three parts, providing an appropriate control strategy for each phase and resulting in cyclic finger motions that accomplish the task.
Contact-invariant optimization for hand manipulation
This work is an extension of the recent contact-invariant optimization method, which introduced auxiliary decision variables directly specifying when and where contacts should occur, and optimized these variables together with the movement trajectory.
Multi-modal Motion Planning for a Humanoid Robot Manipulation Task
The new method, Random-MMP, randomly samples mode transitions to distribute a sparse number of modes across configuration space to enable a humanoid robot to push an object on a flat surface.
Robust finger gaits from closed-loop controllers
  • M. Huber, R. Grupen
  • Engineering
    IEEE/RSJ International Conference on Intelligent Robots and Systems
  • 2002
An approach to finger gaiting is presented which constructs behavior on-line by activating combinations of reusable feedback control laws with formal stability and convergence properties drawn from a control basis.
Synthesis of detailed hand manipulations using contact sampling
A new method for creating natural scenes with human activities that involve both gross full-body motion and detailed hand manipulation of objects is introduced, and a randomized sampling algorithm is proposed to search for as many as possible visually diverse solutions within the computational time budget.
Generalization of human grasping for multi-fingered robot hands
An imitation learning approach for learning and generalizing grasping skills based on human demonstrations is presented, which learns low-dimensional latent grasp spaces for different grasp types which form the basis for a novel extension to dynamic motor primitives.
Physically based grasping control from example
This paper presents a controller for physically based grasping that draws from motion capture data, and it adds compensation for movement of the arm and for gravity to make the behavior of passive and active components less dependent on the dynamics of arm motion.
Interaction capture and synthesis
  • SIGGRAPH 2005
  • 2005
Optimizing walking controllers
A modified version of the SIMBICON controller is optimized for characters of varying body shape, walking speed and step length, which produces a number of important features of natural walking, including active toe-off, near-passive knee swing, and leg extension during swing.
An overview of dexterous manipulation
  • A. Okamura, Niels Smaby, M. Cutkosky
  • Psychology
    Proceedings 2000 ICRA. Millennium Conference. IEEE International Conference on Robotics and Automation. Symposia Proceedings (Cat. No.00CH37065)
  • 2000
An overview of research in dexterous manipulation is presented, which includes grasp planning and quality measures, and looks at mid- and low-level control frameworks, and then compares manipulation versus exploration.