Background

Soft robots are capable of executing a variety of functions in low-cost, lightweight, and mechanically-compliant packages. As such, they could be capable of functioning in challenging environments involving uneven terrain, obstacles, and fitting within small spaces. Equipped with sensors, they may be able to assist in search-and-rescue operations in a disaster area, or retrieval of objects in spaces that are too confined or hazardous for humans or large robots to negotiate. Morin, et. al., 2012 even demonstrated the capability of soft robots to camouflage themselves in order "disappear" from the view of adversaries.

Small agile robots such as the iRobot Packbot, the Boston Dynamics RHex, and the Stanford microTugs are capable of motion in challenging environments, but due to their rigid nature cannot conform to a surface or object and are generally limited in their ability to with manipulation in small packages. Furthermore, they are typically more challenging to fabricate, limiting the possibility to deploy fleets of similar robots. Recently, Paik, Cho and Kim, 2015 have devised a soft-bodied micro robot that can gently lasso an ant in a package that is only twice the diameter of a human hair.

In keeping with the design requirement for a small soft robot design, we develop a hybrid robot where the computing hardware and pneumatic pressure system supporting the soft robot are combined with a rigid-bodied robot - a KUKA youBot. This hybrid system is shown below. This hybrid soft/rigid robot is advantageous over any one of these systems in isolation for several reasons:

  • Improved mobility. A mobile rigid base is able to shuttle the soft robot quickly between points in the workspace that are less challenging to maneuver, allowing the soft robot to traverse areas that are unreachable for the rigid-bodied robot.
  • Improved self awareness. The rigid platform can provide exteroceptive sensing capability (i.e. an RGB+D camera) for accurate localization, enabling a high degree of success in achieving tasks.
  • Synergistic functionality. The computing resources of the youBot are powerful enough to enable coordination between the two robots for accomplishing a richer set of tasks (e.g. autonomous manipulation of objects) than either platform can perform in isolation.
  • Extensibility. A rigid-bodied robot can provide a platform that can be interchanged with different soft robots or expanded to deploy several soft robots at once, providing a platform that is versatile enough to handle a wide variety of tasks.

Hybrid soft and hard robots have been explored recently as a viable design alternative to completion of search-and-rescue missions. In particular, our work is inspired by that of Stokes, et. al. 2014, in which the authors developed a hybrid robot set out to perform an object retrieval task. A soft robot design was combined with an iRobot Create robot as the rigid base to demonstrate the synergistic benefits of a hybrid soft and rigid-bodied design. In that work, there was no automation in the control of the two robots, as the two robots were independently teleoperated by a user, although many synergistic benefits were deonstrated.

The aim of our team's hybrid soft/rigid robotic system is to demonstrate synergistic benefits directly by creating a fully autonomous object retrieval task, through the use of a high-level controller that provides commands to the soft robot through the use of a controller developed in the Robot Operating System (ROS) environment. To accomplish this, we choose a rigid-body robot (youBot) with computational resources powerful enough to coordinate the two robots, features a manipulator to enhance the synergism between the robots, and is large enough to be extensible to various types and larger numbers of soft robots. We also evaluate various soft robots for effectiveness in locomotion and object grasping, and develop controllers for each. Lastly, we demonstrate vision-based navigation with our chosen soft robot design by making use of a camera mounted on-board the youBot to track both the object to be retrieved and the soft robot in real time. To our team's best knowledge, this is the first attempt at soft robot autonomy for a hybrid rigid-bodied and soft robot design.

Stephen A. Morin, Robert F. Shepherd, Sen Wai Kwok, Adam A. Stokes, Alex Nemiroski, George M. Whitesides, "Camouflage and Display for Soft Machines." Science, Vol. 337 no. 6096 pp. 828-832.

Jungwook Paek, Inho Cho, and Jaeyoun Kim, "Microrobotic tentacles with spiral bending capability based on shape-engineered elastomeric microtubes."  Scientific Reports, Vol. 5, 10768, 2015.

Adam A. Stokes, Robert F. Shepherd, Stephen A. Morin, Filip Ilievski, and George M. Whitesides, "A Hybrid Combining Hard and Soft Robots." Soft Robotics, Vol. 1, no. 1, 2014. 

The soft robot, deployed from the youBot, grasping a "minion" in an autonomous object retrieval task.

The target object: A Minion.