|M.Sc Student||Goldenberg Douglas|
|Subject||Cooperative Object Manipulation A Rigidity Approach|
|Department||Department of Aerospace Engineering||Supervisor||Professor Daniel Zelazo|
Moving objects with a single robot is a common practice; however, the weight and
dimensions of the cargo oftentimes limit the task. To aid this, we propose using multiple coordinating robots to move large objects. The main challenge in this approach is that the robots do not know the geometry of the object they are moving, but do have complete information regarding their own relative formation (that is, their position, orientation and velocity). This fact motivates solving the problem using rigidity theory, a tool recently used in formation control. Within this context, the planar motion of the object can be decoupled into pure steady state translations and rotations by utilizing the eigenvectors of the corresponding framework's rigidity matrix. This formulation is beneficial because it only requires local information to create any desired trajectory. In other words, utilizing the robots' position and velocity with respect to the system's center of mass, we are able to produce the forces needed for translations and rotations.
This work explores how the rigidity matrix can be used to find the forces needed to
shift and rotate an object, potentially allowing the robots to move that object anywhere in the plane. Finally, we demonstrate the analytical results with numerical simulations.