Learning differentiable dynamics models for shape control of deformable linear objects

Yang, Yuxuan; Stork, Johannes A.; Stoyanov, Todor

doi:10.1016/j.robot.2022.104258

Cited by 5 publications

(11 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The term shape control of DLOs is typically used to refer to two different manipulation problems targeting the achievement of a final target shape: 1) the manipulation of a soft DLO with a sequence of pick and place actions [12]- [14], where the deformation of the DLO is held in place by the friction of the surface underneath. 2) the manipulation of elastic DLOs with one or more robotic arms and/or one end of the DLO fixed [2], [7], [10], [15], [16], where the second arm is used to achieve better control of the shape, e.g. in the situation where the DLO's stiffness and the object exhibit rigid or plastic behavior.…”

Section: A Dlo Shape Control Taskmentioning

confidence: 99%

“…Regarding the manipulation of elastic DLOs, several works proposed a learning-based framework to predict the DLO dynamics. In [2], [7], [10] the DLO dynamic is approximated with datadriven approaches trained using simulation data. In [2] the authors propose an online adaptation of the DLO model to compensate for the reality gap.…”

Section: Learned Dlo Modelsmentioning

confidence: 99%

“…Their ambiguity can make it difficult to distinguish different parts of them or to distinguish DLOs from other objects [3], especially given their relatively small size [4], [5]. Moreover, the detection of the full DLO state, including the twist along the DLO, poses also a challenge [6], [7]. Regarding manipulation, DLOs prove to be challenging due to their unpredictable configuration behavior, high dimensional statespace, and complex nonlinear dynamics [2], [7].…”

Section: Introductionmentioning

confidence: 99%

“…Moreover, the detection of the full DLO state, including the twist along the DLO, poses also a challenge [6], [7]. Regarding manipulation, DLOs prove to be challenging due to their unpredictable configuration behavior, high dimensional statespace, and complex nonlinear dynamics [2], [7]. Therefore, a deep understanding of their physical characteristics is essential for predicting and controlling their shape effectively [8].…”

Section: Introductionmentioning

confidence: 99%

“…The proposed framework can directly be applied for the manipulation of various DLOs on diverse surfaces, thanks to the data-driven approximation of the DLO dynamics conditioned on the model parameters. Therefore, there is no need to: 1) generate every time new task-specific data as in [7], 2) introduce complex online adaptation controllers as in [2], [10], 3) perform cumbersome and not intuitive parameters identification procedures as in [8], [11].…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Deformable Linear Objects Manipulation With Online Model Parameters Estimation

Caporali,

Kicki,

Galassi

et al. 2024

IEEE Robot. Autom. Lett.

View full text Add to dashboard Cite

Manipulating Deformable Linear Objects (DLOs) is a challenging task for a robotic system due to their unpredictable configuration, high-dimensional state space and complex nonlinear dynamics. This paper presents a framework addressing the manipulation of DLOs, specifically targeting the model-based shape control task with the simultaneous online gradient-based estimation of model parameters. In the proposed framework, a neural network is trained to mimic the DLO dynamics using the data generated with an analytical DLO model for a broad spectrum of its parameters. The neural network-based DLO model is conditioned on these parameters and employed in an online phase to perform the shape control task by estimating the optimal manipulative action through a gradient-based procedure. In parallel, gradient-based optimization is used to adapt the DLO model parameters to make the neural network-based model better capture the dynamics of the real-world DLO being manipulated and match the observed deformations. To assess its effectiveness, the framework is tested across a variety of DLOs, surfaces, and target shapes in a series of experiments. The results of these experiments demonstrate the validity and efficiency of the proposed methodology compared to existing methods. Project website at https://sites.google.com/view/dlo-manipulation.

show abstract

Section: A Dlo Shape Control Taskmentioning

confidence: 99%

Section: Learned Dlo Modelsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Deformable Linear Objects Manipulation With Online Model Parameters Estimation

Caporali,

Kicki,

Galassi

et al. 2024

IEEE Robot. Autom. Lett.

View full text Add to dashboard Cite

show abstract

Robotic co-manipulation of deformable linear objects for large deformation tasks

Almaghout,

Cherubini,

Klimchik

2024

Robotics and Autonomous Systems

View full text Add to dashboard Cite

Casting Manipulation With Unknown String via Motion Generation, Actual Manipulation, and Parameter Estimation

Tabata,

Miyagusuku,

Seki

et al. 2024

IEEE Access

View full text Add to dashboard Cite

In this manuscript, we propose a motion strategy for manipulating strings with unknown properties. Our approach iteratively refines its motion generation based on parameters estimated from observed string behavior, without the need for real-time feedback. This strategy has been shown effective in achieving several motion objectives using uniform strings of similar lengths. In this research, we improve upon this strategy by addressing the challenges posed by varying string lengths and non-uniform strings. For this, we utilize a non-uniform string model and address various string properties to demonstrate the feasibility of our proposed motion strategy. Experiments conducted with different string types and lengths (between 300 to 610mm), including some with non-uniform mass distributions, demonstrate our method's effectiveness. Results show that our proposed method functions effectively with various kinds of strings, regardless of length and mass distribution, without requiring precise model parameters. Unique to this approach is its ability to adapt to various string characteristics through parameter estimation and motion generation, significantly reducing the need for real-world manipulation trials. Our findings illustrate the potential of our method for use in advanced robotic applications that require handling deformable objects.

show abstract

Learning differentiable dynamics models for shape control of deformable linear objects

Cited by 5 publications

References 29 publications

Deformable Linear Objects Manipulation With Online Model Parameters Estimation

Deformable Linear Objects Manipulation With Online Model Parameters Estimation

Robotic co-manipulation of deformable linear objects for large deformation tasks

Casting Manipulation With Unknown String via Motion Generation, Actual Manipulation, and Parameter Estimation

Contact Info

Product

Resources

About