Robotics and Biology Laboratory

Soft Manipulation

Soft Hands represent a departure from traditional robot hand design, which often focuses on precise models and planning of contact points. Instead, our approach emphasizes robustness and safety through the use of soft materials and flexible mechanics. This softness allows us to leverage contact with the environment and use it for effective grasping and manipulation strategies. Our research explores hand design, the integration of sensors in soft hands, and feedback control methods for grasping and in-hand manipulation that take advantage of the hands' morphological properties. Furthermore, we view soft hands as a computational resource that can support robust manipulation behavior generation. This requires a balance between the software that controls the hand and the morphological computation provided by the body, leading to the problem of how to efficiently co-design body and control.

RBO Hand 3

Contact Persons

Steffen Puhlmann, Adrian Sieler

The RBO Hand 3 is a highly dexterous and versatile soft hand based on pneumatic actuation. It advances on its predecessor by combining 16 actuated degrees of freedom with intrinsic mechanical compliance in an anthropomorphic design. Special attention was paid to the design of the opposable thumb. The RBO Hand 3 is inherently safe and highly robust for externsive real-world experimentation. It closely replicates the functioning of its human counterpart to facilitate transfer of insights to human dexterity, and it is capable of dexterous grasping and manipulation

The RBO Hand 3 has been featured on the cover of National Geographic and the Soft Robotics Journal.

If you want to build your own RH3, you are welcome to do so. We have published the entire production process here.

Soft In-Hand Manipulation

Contact Persons

Adrian Sieler, Apoorv Vaish

Unlike most robot grippers, human hands are versatile manipulators: they are soft, compliant, dexterous, and can be used for multiple tasks at once (e.g. lifting a teacup while holding a pencil in the same hand, which requires repositioning of the pencil to free the other fingers for gripping the cup handle).

We want robotic hands to be similarly versatile. The RBO Hand 3 already solves a large part of this problem by mimicking the morphology of a human hand, but planning movements is still a hard problem. Conventional manipulation planners are designed for rigid hands with links, and they employ analytic models of the hand and a previously known object. Such models cease to be useful in our case, because it is hard to precisely model a system with a soft hand and a complex unknown object.

With these constraints in place, how can our soft robotic hands be made to grasp a pencil and move it between different fingers, or spin a cube to an arbitrary face?

If you are curious about how we address this challenge, please follow the link below.

Sensorization of Soft Hands

In robotics, sensing is a crucial aspect, particularly for soft robots. These robots are highly compliant, and therefore, their interactions with the environment cause complex deformations. Sensors play a vital role in observing and measuring such interactions, allowing the robots to react accordingly. However, traditional sensor technologies are not suitable for use with soft robotic actuators, as they are often rigid and not flexible enough to work with silicone rubber-based actuators.

To overcome this challenge, we are developing novel "soft" sensor technologies that are designed specifically for soft hands and actuators. These sensors are flexible and can stretch, bend, and twist without compromising the compliance of the actuator. We also use prior knowledge to extract relevant measurements from the high-dimensional sensor data, allowing us to sense various properties of the actuator, including deformation, contact forces, contact locations, and even object materials. In conclusion, our goal is to create sensors that are compatible with soft robots and enable them to interact with the environment effectively.

You want to learn more? Please follow the link below.


Contact Persons

Steffen Puhlmann, Apoorv Vaish

The behavior of a robotic agent is determined by its control program, shape and material composition, as well as external factors from its environment. Both control and morphology affect the behavior and thus must be chosen carefully to ensure robust and general behavior in various operating environments. Thus, given a set of tasks, control and morphology must be considered as one combined aspect in designing soft robots. Moreover, we can divide responsibilities into morphology and control simultaneously and synergistically to ensure robust behavior in the physical world. This joint programming of morphology and control is called co-design.

RBO Hand 2

Contact Persons

Raphael Deimel, Vincent Wall, Steffen Puhlmann

The RBO Hand 2 is a hand made from PneuFlex actuators mounted on a flexible, printed scaffold. The hand was developed to investigate the capabilities and limits of hands when relying only on soft, deformable structures. The unique deformability provides several advantageous benefits to robots trying to interact with the environment: 

  • very robust against blunt collisions
  • very low impact energies
  • passively compliant fingers and palm decouple contact from the robot arm, stabilizing force control
  • mechanical adaptability to object shapes simplifies finger control
  • the pneumatic actuation makes it easy to create complex hand and actuator geometries

The result of our research are several hand prototypes, which we refer to collectively as Soft Hands. RBO Hand 2 is the latest model and used in our lab for research into grasping strategies.

If you want to build your own, you are welcome to do so! We have published the CAD models for the PneuFlex actuators


RBO Hand 1

Contact Persons

Raphael Deimel, Vincent Wall, Steffen Puhlmann

The RBO Hand (published in 2013) was the first soft hand that employed PneuFlex actuators. It uses 3 pairs of parallel and straight PneuFlex actuators. The finger are also partially connected to each other. The hand has a passively bendable rubber sheet acting as a palm in opposition to the fingers.

Pneumatic Box

Contact Persons

Raphael Deimel, Vincent Wall

For the control of the pneumatic RBO Hand 2 we had to create our own hardware. We developed the "PneumaticBox", which consists of valve array, a single-board computer, and a custom printed circuit board. 

A detailed description of the hard- and software can be found in the Tutorial section.


PneuFlex Actuators

Contact Persons

Raphael Deimel, Steffen Puhlmann

We develop a set of production processes and a complete design toolchain for soft continuum actuators under the name PneuFlex. The toolchain consists of several components:

  • unified process to create PneuFlex actuators of diverse shape from a given mold
  • A method to automatically create PneuFlex molds from a parametric description of shape.
  • A method to automatically create PneuFlex molds from a description of mechanical parameters
  • A system to simulate a PneuFlex actuator or to simulate arrangements of actuators (i.e. soft hands)
  • pneumatic control hardware tailored to the requirements of finger-size soft continuum actuators
  • Controllers to enable complex, synchronized actuation across PneuFlex actuators

Human & Robot Grasping

Contact Persons

Steffen Puhlmann

We study human grasping under a variety of conditions in order to identify and characterize different grasping strategies. Specially, we are interested in strategies that are robust to be performed under different kinds of impairment, e.g. visual. Our final goal, is to transfer those robust strategies to a robot. In addition to subjects' grasping with their natural hands, we also observe them when they use soft robotic hands such as RBO Hand 2 and Pisa/IIT Hand. Our goal is to leverage their intuition in grasping to understand how robotic hands can be controlled and evaluate the transfer of their strategies to robotic hands.

Perception and Planning of Grasps

Contact Persons

Clemens Eppner, Jessica Abele

Our approach to grasping is motivated by the fact that humans don't avoid contact with the environment but rather exploit it to generate haptic feedback complementing visual feedback. This exploitation of environmental constraints simplifies the grasping problem by converting a high-dimensional configuration search problem into successive local searches guided by these environmental constraints, such as surfaces or edges.

We are developing algorithms that model the environment as a collection of environmental constraints which can be used to generate reactive feedback plans that lead to robust and reliable grasps.


Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) under Germany's Excellence Strategy - EXC 2002/1 "Science of Intelligence" - project number 390523135

German Priority Program DFG-SPP 2100 “Soft Material Robotic Systems” - project number 405033880.

Soft Manipulation (SoMa) -  funded by the European Commission in the Horizon 2020 program, award number 645599, May 2015 - April 2019.  

Alexander von Humboldt professorship - awarded by the Alexander von Humboldt foundation and funded through the Ministry of Education and Research, BMBF, July 2009 - June 2014



[english] Puhlmann, Steffen; Harris, Jason; Brock, Oliver
RBO Hand 3: A Platform for Soft Dexterous Manipulation
IEEE Transactions on Robotics :1-16
April 2022
ISSN: 1941-0468


Bhatt, Aditya; Sieler, Adrian; Puhlmann, Steffen; Brock, Oliver
Surprisingly Robust In-Hand Manipulation: An Empirical Study
Proceedings of Robotics: Science and Systems
Herausgeber: Virtual
Juli 2021


Deimel, Raphael; Brock, Oliver
A Novel Type of Compliant, Underactuated Robotic Hand for Dexterous Grasping
Robotics:Science and Systems (RSS)


Deimel, Raphael; Brock, Oliver
A Compliant Hand Based on a Novel Pneumatic Actuator
Proceedings of the IEEE International Conference on Robotics and Automation (ICRA), Seite 2047–2053
Mai 2013