Reducing the Risk of Slip While Transporting an Object in a Micro-Gravity Environment

David Carabis1, John Wen2

  • 11991
  • 2Rensselaer Polytechnic Institute

Details

10:00 - 10:30 | Mon 25 Sep | Ballroom Foyer | MoAmPo.34

Session: Monday Posters AM

Abstract

The manipulation of objects in space is an important task for many potential space applications, such as debris management, construction of objects in space, and satellite servicing. However, contact dynamics can make grasping and manipulation challenging in space. We investigate reducing the risk of slip by dynamically increasing the squeeze force applied to an object during transportation. To do this, we utilize a planar air bearing table integrated with a fixed dual-arm Motoman robot. This air bearing table experimentally approximates true space dynamics by creating a low friction environment. The robot grasps a free floating air bearing object, and then transports the object back and forth along a straight line in an oscillatory manner to increase the local tangent forces at the contact points and create a larger risk of slip. At each timestep, we compute the applied squeeze and move force components based on the grasp geometry and the measurements of local tangent and normal forces at the contact points. If this ratio exceeds the coefficient of friction, slip will occur. We then derive an expression for the ratio of the local tangent to the local normal force in terms of the move and squeeze force components, and the grasp geometry. This allows us to compute a squeeze force that will result in a desired tangent to normal force ratio at the contact points, assuming that the move force component has not changed significantly between two timesteps. If the measured ratio value falls above some safe threshold, this computation is made and the squeeze force is increased to reduce this ratio. Experimental results show this compensation method reducing the ratio of local tangent to local normal forces. Future work will focus on a predictive method of generating the move force component from the object trajectory, rather than relying on measurements from a previous timestep.