Abstract
Abstract The control system architecture development for a full-sized humanoid, HART (Human Assistive RoboT), is presented in this paper. HART has been designed and built by ART (Assistive Robot Team) at University of Hartford since 2017. With its unique design feature, which is kinematic re-scalability, HART can manipulate various control inputs of off-the-shelf vehicles with different sizes and structures in the market. In this study, the control system upgrade that can enable the robot to drive autonomously in trained environments is mainly described. For this, the way-point based navigation capability is built for HART using its trinocular camera system and CNN (Convolutional Neural Network) based terrain classifier. Next, HART’s sensor head, which is also redesigned for advancement of its perception capability, and end-effectors that are upgraded for more efficient manipulation of vehicles’ control inputs are introduced in this paper. Lastly, HART with the upgraded control system is tested and evaluated in human-centered environments.
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
