* Faculty       * Staff       * Contact       * Institute Directory
* Undergraduate       * Graduate       * Institute Admissions: Undergraduate | Graduate      
* Events       * Institute Events      
* Lab Manual       * Institute Computing      
No Menu Selected

* Research

Ph.D. Theses

Physics-Empowered Perception for Robot Grasping and Dexterous Manipulation

By Li Zhang
Advisor: Jeffrey Trinkle
November 15, 2013

Recent years, robot technology has been greatly advanced by sophisticated perception algorithms which usually combine the advantages of the predetermined system model and up-to-date observations from physical sensors. However, the robot's grasping and manipulation capability falls far behind its mobility counterpart. In unstructured environment, state-of-the-art robots grasp only with separate palm and finger motions, which are hence slower than human-like grasp, to overcome unavoidable uncertainties from locating objects and positioning end effectors. Still, well-designed grasps might fail due to inadvertent bumping between the robot hand and the object which causes the object tumble away from the original planned grasp trajectory. The inability to promptly detect such tumbling movements, which will also be induced during dynamic grasping acquisition, is a primary factor of the current status. The lack of knowledge on the object properties also prevents effective reactive grasping strategy as many properties exhibit heavy effects on the dynamic interactions and hence the grasping actions.

To address such weaknesses, we define the G-SL(AM)^2 problem to set requirements on perception algorithms for the ultimate goal of human-like grasping. It requires Simultaneous object Localization And Modeling of its physical properties during Manipulating the object. In this thesis, we propose two perception algorithms, which are essentially estimation frameworks, to attack the G-SL(AM)^2 problem. Both algorithms explicitly adopt a frictional dynamic model to characterize real-time interactions between the robot end effector and the object at the center of respective stochastic dynamic models. Their main difference is that one framework adopts a full-blown dynamic formulation while the other decomposes it into the sub-models of contact mode prediction and state propagation, which are originally tightly coupled in physics. Constraints are hence relaxed at certain sampled states. Both frameworks are proven effective in capturing dynamic behaviors from applications to various grasping scenarios and the relaxation model gives better computation performance.

* Return to main PhD Theses page