Department of Computer Science
Collaborative Goal and Policy Learning from Human Operators of Construction Co-Robots
In this research, the focus is to allow autonomous robots to collaborate with each
other and with human teammates. A part of this research conducted by Dr. Crick and
his students is funded through a grant from NSF's National Robotics Initiative, and
focuses on robotic construction equipment. Humans and robots have different strengths
and skills; humans have a stronger grasp of context and big-picture decision making,
but are not actually very good at low-level control. It takes many years for an equipment
operator to become proficient at manipulating the controls of a motor grader or excavator,
but on the other hand, no robot can coordinate all of the activities on a job site.
Our research bridges the gap, allowing the robot to plan optimal control inputs within
the larger frame of a human-demonstrated task, and perform faster, safer and more
efficiently than the human could do alone.