Statement of Objectives

Deformable objects are widely used in many industrial, agricultural and health applications. Robotizing the tasks that involve these objects would have a huge impact. However, robotic manipulation of soft objects in a dynamic and unconstrained situation is an ambitious research topic that includes different domains such as perception, planning, control, machine learning, human-robot-interaction, etc. Hence, we propose this workshop for deformable object manipulation with the interest in soft interaction which lies in the many potential robotic systems. In particular, the workshop focuses on:

  • Grippers/tools design and grasping for manipulation of deformable objects
  • Grasping of deformable objects can be found in different health and industrial applications. This is a very challenging task due to the non-rigidity properties of the objects. This imposes challenges from different aspects such as gripper and tool design, object grasping, and how to control the grasping points during the manipulations. Some answers exist in the literature composed of having prior knowledge of the deformed object or/and designing a specific gripper or tool. It is also interesting to explore the solutions adopted in the neighboring field of soft robotics, which may be relevant to solve such problems.
  • Data management of deformable objects
  • Data management is an important aspect of robotic control. One of the recent deformable object manipulation methods is to use deep reinforcement learning. However, it is not possible to compare these methods since it is evaluated on a different task variant with different simulators or robot setups. In this topic, we highlight the importance of benchmarking, such as Real2Sim methods, and how to use these benchmarks in learning techniques (sim2real) for deformable objects manipulation.
  • Perception for manipulation of deformable objects
  • Perception approaches are inspired to imitate human perception. While manipulating an object, touch and vision sensors would assist in understanding the object, how to grasp it, and how to manipulate it. However, these sensors become more challenging to use when dealing with deformable objects. Several research works tackle this topic by using vision sensors, tactile sensors or a combination/fusion of both modalities. In recent years, multi-modal sensing has been used in different applications such as deformation state estimation/correction. In addition, perception approaches have been used for interacting with humans and the surrounding environments. This opened the opportunity for new applications where the robot manipulates deformable objects while interacting with the human and the environment.
  • Control for manipulation of deformable objects
  • Planning and control for manipulation of deformable objects are considered complex tasks due to the high nonlinearity of the deformation. Different approaches exist in the literature that tackle the control challenge which is mostly based on perception techniques, contact estimation or object behavior. And others proposed promising planning solutions such as process analysis, learning from demonstration, deep reinforcement learning, and generalization planning. In this topic, we focus on highlighting the complexity and solutions for planning and control for manipulation of deformable objects.

The workshop will present current advances and contributions in deformable object manipulation. A panel of recognized experts will present and discuss their recent and innovative research results. In addition, we propose an intervention between experts close to the subject to better understand the communication and dissemination within the community.
Akin to ROMADO’s previous workshops and to prompt discussion of current and future research, paper submissions will be requested, and a poster session will be organized for accepted papers. Furthermore, all materials (presentations and discussions, but also experimental data) will be available online.