Master Thesis on “Model-based Dynamic Manipulation of Deformable Objects”

Supervisor: Prof. Ville Kyrki (ville.kyrki@aalto.fi)
Advisor: Dr. Gökhan Alcan (gokhan.alcan@aalto.fi), David Blanco Mulero (david.blancomulero@aalto.fi),

Keywords: rope manipulation, deformable object modeling, model-based control.

Project Description

Task demo using the lasso environment [4]

Deformable objects such as cloths and ropes have many degrees of freedom, which creates additional challenges while manipulating them compared to rigid objects. The forces caused by the acceleration define the manipulation task as static or dynamic. Even though dynamic manipulation is more challenging due to the difficulties in characterizing high-dimensional configurations spaces, it generates some advantages by allowing control of the non-grasped points of the deformable object as well.

The goal of this thesis is to develop model-based approaches for the dynamic manipulation of a deformable object. Particularly, the focus will be on the task of throwing a lasso around a target (a bollard) with a robotic arm. In this context, the thesis is expected to include control-oriented modeling of the rope dynamics and trajectory optimization. Initial experiments will be conducted in simulation, and verified controllers will be tested with a real robot.

Deliverables

  • Review of relevant literature,
  • Realization of high-fidelity rope simulations in a selected environment (MuJoCo, PyBullet, etc.),
  • Control-oriented modeling of rope dynamics, e.g. using mesh-based models,
  • Trajectory optimization based problem formulation and method development for throwing lasso task,
  • Simulation experiments and real-world tests.

Practical Information

Pre-requisites: Python(high/medium), ML(medium), optimal control (medium)

Tools: Franka Panda Robot, PyTorch/Jax

Simulators: (up-to-change) MuJoCo, PyBullet

Start: Available immediately

References

[1] Dynamic Manipulation of Deformable Objects With Implicit Integration, https://ieeexplore.ieee.org/document/9380919

[2] Iterative Residual Policy: for Goal-Conditioned Dynamic Manipulation of Deformable Objects, https://arxiv.org/pdf/2203.00663.pdf

[3] Differentiable Cloth Simulation for Inverse Problems, https://proceedings.neurips.cc/paper/2019/file/28f0b864598a1291557bed248a998d4e-Paper.pdf

[4] Dynamic environments with deformable objects, https://openreview.net/pdf?id=WcY35wjmCBA, project website: https://github.com/contactrika/dedo