Welcome

Welcome to my homepage! My name is Jan Peters. I am a Ph.D. student in the Department of Computer Science at the University of Southern California in sunny Los Angeles. I am working with Stefan Schaal, Sethu Vijayakumar (now at U. Edinburgh, UK), and Firdaus Udwadia (from the Department of Mechanical Engineering).

My research centers around the goal of bringing advanced motor skills to robotics using techniques from machine learning and optimal control. If you are interested in machine learning and optimal control, check out the learning and control sections. For motor skill representation and experiments, see the motor skill section and for my research in robotics, see the robotics tab. Furthermore, I have a longer introduction below and also check out the few publications which I have online. Attach:janOnBeach350.jpg Δ

Recently, I passed my my Quals Examination on May 2nd, 2005 at noon. We had a nice party in the free peoples republic of long beach - see the photos?. Also, I spent a great Robotics: Science and Systems conference in Boston where I am organized a workshop on Learning of Locomotion at the Robotics 2005 conference. See here for more information.

Before joining USC, I studied computer science (with a focus on artificial intelligence), and electrical engineering (majoring in automation & control) in Germany and Singapore, worked in Germany, Japan, and Singapore. Furthermore, I obtained a Dipl.-Ing. (German MSEE) from Munich University of Technology and a Dipl.-Inform. (German MSCS) from Hagen University. Furthermore, I have obtained the obligatory MSCS from University of Southern California and recently completed a Masters in Mechanical Engineering in my spare time. Check out my curriculum vitae for these subtleties.

In case that you are searching for my address or for directions on how to get to our lab and my office, look at my contact information. If you need to find my current and only valid email address fast and without exhaustive searches :), I better save you the time as it is

In general, I always like to get friendly emails - so please drop me a line. For all further contact information look right here.

Research

During the last years, my research in reinforcement learning has centered around policy search techniques. I have created a lot of examples and code showing how these methods work and how they relate to previous reinforcement learning techniques (see the policy gradients tab). Inspired by the work of Sutton, Konda, Kakade, and others, we have investigated efficient methods of learning policy gradient compatible function approximation and using it for natural policy gradient descent. This resulted in our work on Natural Actor-Critic?. Currently, we are focusing and applying these techniques.

Another area of my interests is the optimization of motor primitives as a novel way of planning robust motor plans or trajectories for robotics. Currently, I am working on two different ways of obtaining optimal motor primitives. The first of these methods relies on priming the motor primitives by imitation learning, and subsequently improving them with Natural Actor-Critic. As an alternative way, I am working on optimal motor primitives which are directly obtained.

The planning, control and motor control of humanoid robots are some of the most fascinating unsolved problems. My personal goal is to find learning control methods which will enable humanoid robots to perform more complicated tasks. I am particularly interested in biped locomotion (check out the pages in humanoid robotics?biped locomotion), and I am currently organizing a seminar on biped locomotion.

Furthermore, I am interested in the biological inspiration of motor control. Human motor control yields a variety of clues for control of humanoid robots. Particularly, I am interested in which cost functions underly human locomotion - and obviously, I see optimal control as an interesting way of studying these. Earlier in my life, I was also interested in modeling the biomechanics (e.g., the human knee), and in computational models of parts of the central nervous system.

My Life and my World

I live in Long Beach, California right at the beach. If you don't believe me that I live in one of the prettiest places in the world, check out my old? and my new pictures? from life in Long Beach ... or the gorgeous sunrises in long beach? after worked-through nights...

Attach:milliondollarhome.jpg Δ Δ Also, I enjoy rollerblading, travelling, skiing and sailing. Check out what has been going on in my life? and what my world? has to offer.

Do you want to see a Satelite Image from where I live? Check it out !

I have just recently learned from the LA Times in the homes section that I live with a million dollar view ... click on the newspaper to your right!

  

zum Seitenanfang