Current research on planning human-robot interaction is mostly based on sampling-based or human demonstration-guided motion planning, in which robots learn from a set of human demonstrations to generate plans for avoiding obstacles or performing an assistive task. However, simulating human motion is a complex problem due to the redundancy of the human musculoskeletal system; thus, it increases the cost of the solution/planning to manipulate the robot’s motion. Our research will focus on developing new task-based dynamic motion planning using single or multi-objective optimization techniques that eliminate basic physical and kinematical constraints of simulating human motion.

“Through this SuperUROP project, I want to gain professional research experience in the robotics field. I’m interested in applying my knowledge in machine learning and computer vision to improve human-robotic interaction. I hope to publish a paper by the end of the SuperUROP project if I have meaningful results to display.”