End-to-end dexterous manipulation with deliberate interactive estimation

Hudson, Nicolas; Howard, Thomas M.; Ma, Jie; Jain, Abhinandan; Bajracharya, Max; Myint, Steven; Kuo, Calvin; Matthies, Larry; Backes, Paul; Hebert, Paul D. N.; Fuchs, Thomas J.; Burdick, Joel W.

doi:10.1109/icra.2012.6225101

Cited by 41 publications

(23 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Recently this setting has received attention by teams participating in the DARPA ARM challenge [10]- [12] and the same study on humans when grasping without vision [13]. Those methods mostly depend on human expertise when designing the feedback controllers.…”

Section: Related Workmentioning

confidence: 98%

POMDP manipulation via trajectory optimization

Vien

Toussaint

2015

2015 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

View full text Add to dashboard Cite

Efficient object manipulation based only on force feedback typically requires a plan of actively contact-seeking actions to reduce uncertainty over the true environmental model. In principle, that problem could be formulated as a full partially observable Markov decision process (POMDP) whose observations are sensed forces indicating the presence/absence of contacts with objects. Such a naive application leads to a very large POMDP with high-dimensional continuous state, action and observation spaces. Solving such large POMDPs is practically prohibitive. In other words, we are facing three challenging problems: 1) uncertainty over discontinuous contacts with objects; 2) high-dimensional continuous spaces; 3) optimization for not only trajectory cost but also execution time. As trajectory optimization is a powerful model-based method for motion generation, it can handle the last two issues effectively by computing locally optimal trajectories. This paper aims to integrate advantages of trajectory optimization into existing POMDP solvers. The full POMDP formulation is solved using sample-based approaches, where each sampled model is quickly evaluated via trajectory optimization instead of simulating a large number of rollouts. To further accelerate the solver, we propose to integrate temporal abstraction, i.e. macro actions or temporal actions, into the POMDP model. We demonstrate the proposed method on a simulated 7 DoF KUKA arm and a physical Willow Garage PR2 platform. The results show that our proposed method could effectively seek contacts in complex scenarios, and achieve near-optimal performance of path planing.

show abstract

Section: Related Workmentioning

confidence: 98%

POMDP manipulation via trajectory optimization

Vien

Toussaint

2015

2015 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

View full text Add to dashboard Cite

show abstract

“…For objects with a known polygonal mesh model, experience databases can be built offline and serve as grasp look-up table once this object has been detected in the scene. In [19,13,8] it has been shown that in this case robust grasping and manipulation can be achieved by applying force control and exploiting constraints in the environment. However, to transfer successful grasps between different objects of which only partial and noisy information is known, remains a challenge.…”

Section: Introductionmentioning

confidence: 99%

Optimizing for what matters: The top grasp hypothesis

Kappler

Schaal

Bohg

2016

2016 IEEE International Conference on Robotics and Automation (ICRA)

View full text Add to dashboard Cite

In this paper, we consider the problem of robotic grasping of objects when only partial and noisy sensor data of the environment is available. We are specifically interested in the problem of reliably selecting the best hypothesis from a whole set. This is commonly the case when trying to grasp an object for which we can only observe a partial point cloud from one viewpoint through noisy sensors. There will be many possible ways to successfully grasp this object, and even more which will fail. We propose a supervised learning method that is trained with a ranking loss. This explicitly encourages that the top-ranked training grasp in a hypothesis set is also positively labeled. We show how we adapt the standard ranking loss to work with data that has binary labels and explain the benefits of this formulation. Additionally, we show how we can efficiently optimize this loss with stochastic gradient descent.In quantitative experiments, we show that we can outperform previous models by a large margin.

show abstract

“…Each task was tested separately in static environments with only minor focus on obstacle avoidance but on grasping, sensor calibration and force feedback. Several teams have published their approaches [8], [9], [10]. In 2014, the EU-funded project EuRoC 1 [3] has been started.…”

Section: Introductionmentioning

confidence: 99%

A Flexible Robotic Framework for Autonomous Manufacturing Processes: Report from the European Robotics Challenge Stage 1

Hildebrandt

Schuetz

Wahrmann

et al. 2016

2016 International Conference on Autonomous Robot Systems and Competitions (ICARSC)

View full text Add to dashboard Cite

Abstract-Common industrial automation approaches consist on heavy and fixed robotic manipulators working in separated and closed production lines. Recent advances in sensing and control are leading to flexible and versatile robotic manipulator platforms, which could work in barrier-free production areas and might be the next advance in industrial production. In our contribution, we present a sophisticated and completely autonomous software framework for handling complex pickand-place tasks using state-of-the-art tools and algorithms. It covers applicable solutions for 3D image processing, motion planning, grasping as well as error handling and task scheduling strategies. Its competitiveness in terms of robustness and performance has been proven by earning the first place among 39 teams from all over Europe in the simulation stage of the EuRoC. At the challenge 2 of this EU-funded project, a mobile robotic platform shall be applied for intelligent, flexible and highly automated production systems. We discuss our results as well as the applicability of our framework to real industrial applications.

show abstract

End-to-end dexterous manipulation with deliberate interactive estimation

Cited by 41 publications

References 21 publications

POMDP manipulation via trajectory optimization

POMDP manipulation via trajectory optimization

Optimizing for what matters: The top grasp hypothesis

A Flexible Robotic Framework for Autonomous Manufacturing Processes: Report from the European Robotics Challenge Stage 1

Contact Info

Product

Resources

About