Enhancing Payload Capacity with Dual-Arm Manipulation and Adaptable Mechanical Intelligence
Kim, Raymond Sunghwan
MetadataShow full item record
Manipulating large and heavy objects is a crucial task in various robotic applications such as agriculture, search and rescue, service, and manufacturing. While modern manipulators have advanced considerably, they are limited by their net load capacity. This places a fundamental limit on the weight of loads that a single manipulator can move. For a case where a large load exceeds the capacity of a single manipulator, there are two potential solutions. First, the manipulator can be replaced with a larger one to increase the maximum payload. This can be time-consuming and expensive. Alternatively, two manipulators can be used collaboratively to share the load. This enables use of existing manipulators. Cooperative manipulation with two arms has the potential to increase the net load capacity of the system. However, it is critical that proper load sharing takes place between the two arms. If this is not maintained, the load limits of one of the arms can be exceeded and lead to catastrophic failure. Ensuring load sharing can be a challenging controls and coordination problem. In this work, a method that utilizes mechanical intelligence in the form of a whiffletree is outlined. A whiffletree is a mechanical device that allows distribution of load through the use of pivot points and linkages. Whiffletrees are used in a range of applications including bionic limbs, under-actuated fingers, horse-harnesses on carriages, and wind turbine tests. Typically, a whiffletree consists of a bar pivoted at or near the center, with force applied from one direction to the pivot and from the other direction to the tips. The points on the linkage act as pivot points, allowing positional displacements for any attached loads. This method is used to design, fabricate, and assemble the dual-arm whiffletree gripper system that enables load sharing amongst two manipulators. The mechanical properties of the whiffletree allows load distribution without any force sensory feedback and enables robustness to positional displacements. As a result, the system is able to integrate a simplified, position-control based strategy. To allow ease of integration to existing robotic systems, the overall design of this work is easily attachable/detachable with various types of customizable grippers using pneumatic tool changers. Physical experiments were conducted to illustrate the enhanced load capacity of a robotic system using the dual-arm whiffletree gripper. Specifically, two UR5 manipulators, each with 5kg maximum payload, are utilized to re-position a 7kg load. This load would exceed the capacity of a single arm, and the experimental results show that the forces on each arm remain below this level and are evenly distributed.