• PhD position: Reactive Grasping Algorithms: Position filled! The Phd in this position will focus on the development of off-line grasp planning algorithms and on-line reactive grasp acquisition algorithms that can make optimal use of state estimates provided by a real-time perception system. This Phd student will also be the technical leader for this project.
• PhD position: Real-Time Grasp State Perception: The focus of this position is in the area of object perception using tactile, visual, and proprioceptive sensors. Recursive Bayesian estimation methods will be designed to track the motion of the object during the grasping process, estimate its most important physical attributes, and identify the contact points between the object and hand. Testing will be conducted with the Lab's Barrett Whole-Arm Manipulator outfitted with a one-degree-of-freedom gripper designed for use in seawater.

Parallel Physical Simulation
Positions are available to build experimentally verified simulation tools with the ultimate goal to maximize the mobility and manipulation capabilities of robotic systems. In particular, this project seeks to develop new algorithms that set the standard for accurate high-speed simulation of multibody systems on modern parallel computing architectures. The algorithms will be implemented in an open-source physics engine and verified in studies of robot traction on granular terrain (hard-packed sub-terrain covered by a layer of loose stones) and robot grasp acquisition studies.

• Postdoc position: Numerical Algorithms for Physical Simulation: Position filled! The postdoc holding this position will design and implement new algorithms for simulating large numbers of bodies undergoing intermittent contact with friction. The goal is for the new algorithms to exhibit super-linear convergence in solving time-stepping subproblems in single-threaded implementations. The postdoc will also be the technical leader of this project.
• PhD position: Parallel Numerical Algorithms: Position filled! This Phd student will work closely with the postdoc to design and implement a framework for running the new solvers on one or more modern parallel computing architectures. Those available for this project at a minimum are: a bank of Tesla GPUs, a Blue Gene/Q, and an Opteron cluster. The primary goal will be robust parallel solvers with excellent scalability properties.
• PhD position: Analysis of Physical Experiments: This Phd student will have two main responsibilities: (A) develop a method for stochastic characterization of the behavior of large systems of bodies in intermittent frictional contact and (B) conduct physical experiments to demonstrate the accuracy and practicality of the characterization method. The set of experiments will be designed around two themes: (1) stochastic characterization of traction available to legged robots running on granular terrain and (2) estimation of the probability of the success of a given grasp controller in acquiring a satisfactory grasp. A new Barrett Whole-Arm Manipulator and Hand with PPS tactile sensor arrays and finger joint torque sensors will be available for this project.

MS and BS Positions
Students in MS and BS programs in the Schools of Science and Engineering may join any existing project. Experience in any of the following are desirable, but not required: Matlab/Simulink, mechanical or electronic design and testing, real-time computing (e.g., xPC Target or Xenomi), probability and statistics, mechanical systems simulation (e.g., Working Model, Bullet Physics), C/C++, design of experiments, or database systems. Experience with FIRST is desirable, but not required.

Application Instructions:
Applicants should send a current resume, a statement of research interests and career goals, and the names and contact information for at least three references to Professor Jeff Trinkle at trinkle@gmail.com. Applications for the postdoc position will be accepted until the positions are filled. Applicants for the Phd positions should also apply for admission into the CS Department's graduate program.