Project Type: Research + Course Organization: Robotics and Intelligent Vehicles Research (RIVeR) Lab Course: Mobile Robotics Robot Used: Toyota Human Support Robot (HSR) Software Used: C++ | Python | ROS Date(s): Jan 2019 - Apr 2019
Japan has among the highest rates of elderly population in the world resulting in increasing costs and needs for elderly care. Part of Toyota’s mission is to develop robots to assist in the care of the elderly leading to the development of their Human Support Robot (HSR). At Northeastern’s RIVeR lab we are working with Toyota’s HSRs to teach them how to perform activities of daily living. The goal of this particular project is to direct the HSR tidy up a house. The human operator should be able to tell the HSR, via voice commands, to clean up a particular room of the house. The HSR will then autonomously recognize out of place objects, pick them up and return them to their correct places.
For this project I was responsible for motion planning and high-level behavior. I worked with the perception team to take the object location data to calculate a motion plan to approach and grasp an object. These planned motions were validated in simulation and executed on real robot hardware to successfully pickup objects incluing mugs, bottles and snack boxes. I also took the outputs of the speech processing library and programmed the behavior of the robot to navigate to the requested room and pickup the requested object. The high level behavior was programmed using python with ROS to handle the robot/node communication. At the culmination of this project, we presented our results to Toyota along with other universities including Stanford and MIT.
Address
Phone #
Boston, MA
781 812 8630
joelynch523@gmail.com