https://www.linkedin.com/pub/jim-mainprice/1b/948/4b2https://github.com/jmainpri?tab=repositorieshttps://www.researchgate.net/profile/Jim_Mainprice https://scholar.google.com/citations?user=ToU9KBUAAAAJ
 jim dot mainprice at tuebingen dot mpg dot de


I am currently working in Pr. Dr. Stefan Schaal's research group called the Autonomous Motion Department (AMD) at the Max Planck Institute for Intelligent Systems in Tuebingen, Germany. I am generally interested in motion generation; either when addressing complex robot motion planning problems or when studying human motion generation. The primary purpose of my research is enabling the seamless integration of high degree-of-freedom robots in the human environment. Predicting and mimicking human behavior is key to this integration, as our behavior encodes fundamental properties about social space sharing. Hence I currently investigate software components able to learn probabilistic models of human motion, and use these models to generate safer, more efficient and more comfortable robot behavior.

Prior I worked two years under the supervision of Pr. Dmitry Berenson in the Arc-Lab at Worcester Polytechnic Institute (WPI) based in Massachusetts, USA. I investigated motion planning in the context of human-robot collaboration and participated in the valve turning task of the DARPA Robotics Challenge (DRC). 

At WPI I investigated how to account for a prediction of "human workspace occupancy" when planning robot manipulation motion in close proximity to a human. I have then investigated Inverse Optimal Control (IOC) to balance the elementary terms of the cost function used by the motion planner in collaborative scenarios. This was made possible by gathering a library of human collaboration motions using a motion capture system. This work was published in IEEE Transaction on Robotics.

I received my PhD in robotics at LAAS-CNRS (Laboratoire d'Analyse et d'Architecture des Systèmes), in the RIS team lead by Dr. Rachid Alami. My advisor was Dr. Thierry SiméonI studied engineering problems that arise when autonomous robots move in a home or public environment. These environments are typically unstructured and require a high-degree of autonomy from the robot. My research was focused on motion planning for dexterous manipulation as part of Dexmart (a UE-FP7 project) but more precisely I was researching motion planning algorithms that would take the human presence explicitly into account in order to increase safety and comfort in human-robot interactions.