TriFinger: An Open-Source Robot for Learning Dexterity

Wüthrich, Manuel; Widmaier, Felix; Grimminger, Felix; Akpo, Joel; Joshi, Shruti; Agrawal, Vaibhav; Hammoud, Bilal; Khadiv, Majid; Bogdanovic, Miroslav; Berenz, Vincent; Viereck, Julian; Naveau, Maximilien; Righetti, Ludovic; Schölkopf, Bernhard; Bauer, Sebastian

doi:10.48550/arxiv.2008.03596

Cited by 12 publications

(21 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The Real Robot Challenge is a manipulation challenge using the TriFinger Platform developed by Wüthrich et al [10]. The TriFinger robot consists of three identical fingers, each with three degrees of freedom, that are positioned above the robot's workspace.…”

Section: Real Robot Challengementioning

confidence: 99%

“…This approach does not rely on feedback of the object's pose, allowing us to be robust to inaccurate or missed object observations. In Phases 1 and 2, however, where pose estimations are more reliable, we also add feedback controllers: a force control policy similar to Wüthrich et al [10] that helps maintain a stable grasp on the object, and a learned residual policy [1,7] that applies corrective position and torque actions to our controller. In all phases, we combine this approach with a set of motion primitives that grasp and align the object before attempting to move it to the goal pose.…”

Section: Approachmentioning

confidence: 99%

“…Our force control pipeline is similar to the one described by Wüthrich et al [10]. We use a PD controller to turn object pose errors into a wrench applied to the object's center of mass.…”

Section: Feedback Controllers Used In Phases 1 Andmentioning

confidence: 99%

See 2 more Smart Citations

Grasp and Motion Planning for Dexterous Manipulation for the Real Robot Challenge

Yoneda,

Schaff,

Maeda

et al. 2021

Preprint

View full text Add to dashboard Cite

This report describes our winning submission to the Real Robot Challenge. The Real Robot Challenge is a three-phase dexterous manipulation competition that involves manipulating various rectangular objects with the TriFinger Platform. Our approach combines motion planning with several motion primitives to manipulate the object. For Phases 1 and 2, we additionally learn a residual policy in simulation that applies corrective actions on top of our controller. Our approach won first place in Phase 2 and Phase 3 of the competition. We were anonymously known as 'ardentstork' on the competition leaderboard. Videos and our code can be found at https://github.com/ripl-ttic/real-robot-challenge. * Equal contribution 1 For tasks involving the cuboid object (Phase 3), it is not necessary to achieve a desired rotation around the long axis of the cuboid, due to the difficulty of accurately measuring it.

show abstract

Section: Real Robot Challengementioning

confidence: 99%

Section: Approachmentioning

confidence: 99%

See 1 more Smart Citation

Grasp and Motion Planning for Dexterous Manipulation for the Real Robot Challenge

Yoneda,

Schaff,

Maeda

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Third, we extensively analyze the usefulness of disentangled representations in terms of out-of-distribution downstream generalization, both in terms of held-out factors of variation and sim2real transfer. In fact, our dataset is based on the TriFinger robot from Wüthrich et al (2020), which can be built to test the deployment of models in the real world. While the analysis in this paper focuses on the transfer and generalization of predictive models, we hope that our dataset may serve as a benchmark to explore the usefulness of disentangled representations in real-world control tasks.…”

Section: Introductionmentioning

confidence: 99%

On the Transfer of Disentangled Representations in Realistic Settings

Dittadi,

Träuble,

Locatello

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

Learning meaningful representations that disentangle the underlying structure of the data generating process is considered to be of key importance in machine learning. While disentangled representations were found to be useful for diverse tasks such as abstract reasoning and fair classification, their scalability and real-world impact remain questionable. We introduce a new high-resolution dataset with 1M simulated images and over 1,800 annotated real-world images of the same robotic setup. In contrast to previous work, this new dataset exhibits correlations, a complex underlying structure, and allows to evaluate transfer to unseen simulated and real-world settings where the encoder i) remains in distribution or ii) is out of distribution. We propose new architectures in order to scale disentangled representation learning to realistic high-resolution settings and conduct a large-scale empirical study of disentangled representations on this dataset. We observe that disentanglement is a good predictor for out-of-distribution (OOD) task performance.

show abstract

“…After an initial RRC qualifying phase, successful participants are entered into Phase 1 where they must solve the challenging 'Move Cube on Trajectory' task. In this task, a cube must be carried along a goal trajectory, which specifies the coordinates at which the cube should be positioned at each time-step, using the provided TriFinger robotic platform [3]. For final Phase 1 evaluation, participants submit their developed control policy and receive a score based on how closely it can bring the cube to a number of randomly sampled goal trajectories.…”

mentioning

confidence: 99%