Equipment for Object Recognition and Robotic Manipulation
Abstract
Major Goals: The major goal of this award is to purchase robot (mobile) manipulators, mobile robots, and vision sensors for research and STEM education in object recognition and perception-based robotic manipulation and action recognition. Accomplishments: We have purchased a number of robot manipulators, mobile robots, mobile manipulator, and vision sensors. Please see the uploaded PDF file for detailed description and images of specific equipment purchased. We have also used the equipment for research and STEM education. Training Opportunities: One PhD student has used the Franka Emika Panda robot to develop and test a novel force forecast approach to reducing object pose uncertainty in challenging and contact-rich robotic manipulation tasks. Successful real world assembly for multiple peg-in-hole tasks has been achieved through forecasting the contact force in haptic simulation that best matches the measured contact force by the torque sensors of the Panda robot and using that to reduce pose uncertainty. The multiple-peg insertion process also takes advantage of the compliant control capability of the Panda robot. The Uvify Draco-R Research drones are being used in a project on coverage of linear environment features (e.g., roads, power lines). The line coverage algorithms were developed as part of a DARPA OFFSET Sprint 2 project onAutonomous Robot Swarms for Urban Coverage. Two PhD students participated in this project. Two undergraduate students, including one female undergraduate, have begun using the above equipment. One student is using the OptiTrack motion capture system with CrazyFlie robots to perform multiple-robot coordination experiments. The other student will be working on using the Aion R1 and Jackal robots to perform multiple-robot exploration of indoor environments. Two Masters students, including one female student, worked on programming the Fetch mobile manipulator to perform autonomous manipulation tasks.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 07, 2019
- Accession Number
- AD1097218
Entities
People
- Srinivas Akella
Organizations
- University of North Carolina at Chapel Hill