Grasping and Manipulation Engineer Intern

About Willow Garage

Willow Garage intends to revolutionize robotic mobile manipulation and to jump start a broad scale industry in that area. We are working towards this goal through fostering robust open hardware and software platforms; doing research internally and in close cooperation with top research groups across the globe to systematically remove the key bottlenecks to robotic capability; and enabling or spinning off companies in different application areas. We are well funded and able to do this with staying power.

We have assembled a team of international leaders in a wide range of robot fields from machine perception, machine learning, planning, control, grasping, human-robotic interaction and on to mechanical and electrical engineering. We have a great work environment with large and fluid collaborations across areas of expertise. We're open, we publish, run workshops, attend conferences and collaborate widely with top researchers around the world. We have soccer games and free lunches, and company outings from ski trips to zero-g flights.

Primary Function

Demonstrate new grasping and manipulation abilities on the PR-2 platform, in real-life scenarios and using sensor data from the on-board sensor-suite.

Roles and Responsibilities

Possible projects:

  • Teleoperation for grasping and manipulation tasks
    • Teleoperation via web interfaces
    • Shared autonomy for manipulation (robot asks for help when uncertain, or makes suggestions that the user can accept or overrule)
  • Combine novel object recognition results with probabilistic grasp planning framework
    • Feed recognition results to probabilistic grasp planner, which reasons about uncertainty in object recognition and plans grasps accordingly
    • Study interplay between components, propose and implement improvements for each of them as well as for the complete system
  • Interface for learning from demonstration
    • For specific grasps of objects
    • For entire manipulation tasks
    • As a way of specifying task or geometric constraints
  • Implement existing research for primitive decomposition of objects (mesh, point cloud)
  • Integrate object primitive detector, implement grasp planner for object primitives
  • Implement existing research on visual servoing

Skills / Job Requirements

  • experience using real-life sensor data and dealing with error / uncertainty
  • excellent C++ and/or Python coding abilities
  • comfortable operating under Linux (Ubuntu)

Nice to have Skills

  • knowledge of the fundamentals of robotic grasping (e.g. forward and inverse kinematics, grasp and arm Jacobian, form- and force- closure, friction models, etc.)
  • experience with tactile sensors
  • experience controlling robot arms
  • experience using vision / stereo vision / laser range finders
  • experience with ROS
  • experience with one of the following software packages: OpenRave, GraspIt!, Gazebo, OpenCV

Application Materials

  • Cover letter
  • Resume
  • Downloadable code sample(s)
  • Two letters of recommendation
  • Please indicate internship time frame (Summer, Fall, Winter, Spring)

 

To Apply

Send your Application Materials and links to any projects you'd like us to take a look at to jobs@willowgarage.com

Willow Garage is an equal opportunity employer