Research Opportunities in Robot Manipulation

If you are currently an undergraduate or graduate student at the University of Utah interested in conducting research with me, then please the rest of this page and contact me! We conduct research on improving and enhancing robot manipulation in unstructured environments like your home or offices in the School of Computing. Much of our work focuses on how robots can learn to manipulate objects they have never seen before or otherwise have incomplete models or knowledge of. We validate our work through extensive experiments on real robot hardware.

As such we have many opportunities for undergraduate students to help run experiments on the robots, build new tools to help us more easily run, evaluate, and understand our experiments, or to update and extend our existing existing software in motion planning, machine learning, and robot perception. Some example past student projects include finding unknown objects in 3D scans of the environment, placing objects in cluttered scenes, and learning in simulation to improve performance on real robots.

Some specific topics ready for an interested student:

  • Data collection, training, and evaluation of an existing neural network for grasping
  • GPU implementation to improve efficiency of existing motion planning and collision detection algorithms
  • Modeling simple robots using probabilistic programming languages
  • Fast fitting of 3D geometric primitives in vision

Required Qualifications: Opportunities exist for students with a wide variety of backgrounds at different points in their studies. If you are comfortable running and building code in Linux, or are willing to learn how with little guidance, we can likely find something for you to do. For more advanced projects we expect familiarity with C++ or Python as well as no fear of linear algebra, probability, and calculus.

Time Expectations: We expect a minimum commitment of 6 hours a week in the lab, with flexible scheduling depending on the specific project. If initial work goes well, there are opportunities for larger contributions for course credit.

How to Get Involved: If you are interested in becoming involved, first take a look at our website and recent projects. Then email Professor Hermans with a brief statement describing who you are, why you want to work with the group, and optionally any specific projects you'd like to be involved with. Please include a list of all CS and other relevant courses you've taken. Here's a secrete code to include in your email to show you read this: gT7L8g. In general, for master's students, I will require you to take my Motion Planning or Artificial Intelligence class and receive a grade of “A” before taking an independent study with me.

Prospective Students

If you are a prospective student interested in doing research in robotics at the University of Utah, please read about our graduate admissions process and our robotics group. Please only contact me directly if you have specific questions about my research or our graduate program, which can not be answered from reading these websites. If you still wish to send me an email, please include the secret code above to show you have read this page. Unfortunately, I am not able to respond to all emails about applications to our graduate program. If you are applying to the School of Computing and are interested in my research group, please state this in your statement of purpose and list me as a potential advisor in the application system.

I do not have summer internship positions or visiting researcher positions at this time.

Prospective TAs

If you are a current Utah student interested in being a TA for one of my courses, you should not contact me directly. Instead please apply via the TA pool.