A Mobile Manipulation Platform for Collaborative Human-Robot Dexterous Manipulation
Abstract
In this proposal the performer aims to build a mobile robotic manipulation platform for research into side-by-side, collaborative human-robot manipulation. This technology will be directly applicable to Human-Robot Interaction (ONR Department 34), for robots working alongside humans on unscripted and complex manipulation. Targeting applications like shipboard maintenance, logistics, or emergency response, they propose to build a platform with the following components: (i) a torque-controlled robotic arm for safe side-by-side human interaction; (ii) a mobile base able to power and transport the system and to perform collision-free navigation and obstacle avoidance; (iii) a computer for running the suite of manipulation algorithms they plan to develop; (iv) depth and color cameras as the main sensors for manipulation.This platform will enable a three-pronged research plan into Human-in-the-Loop robotic manipulation. First, the performer will develop new versatile robotic hands for teleoperation and shared autonomy, which will be mounted as end-effectors on the robot arm. Second, they will develop methods for side-by-side teleoperation where the user controls the robotic end-effector by moving his or her own hand, without holding dedicated input devices (e.g. joysticks). Third, they will develop interfaces by which the user provides "cognitively valuable" data that is difficult for the robot to obtain, but requires few bits to be transmitted.The resulting platform will also be instrumental for teaching. In Introduction to Robotics, taught by PI Ciocarlie in Fall of 2014 and 2015, class assignments are entirely project-based, implemented so far on a Baxter manipulator. Hands-on robot experience was rated by students as a class highlight. This new platform, intrinsically human-safe and approachable, will allow projects to include key aspects of mobile manipulation. The new platform will also be used in teaching Advanced Robotic Manipulation, where it will enable projects including grasp and manipulation planning along with aspects of Human-Robot Interaction pertaining to manipulation.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Nov 23, 2016
- Source ID
- N000141613176
Entities
People
- Matei Ciocarlie
Organizations
- Office of Naval Research
- Trustees of Columbia University in the City of New York
- United States Navy