Ground Vehicle Driving by Full Sized Humanoid

Abstract

This paper presents a technical overview of the Team DRC-Hubo@UNLV’s approach to the driving task in DARPA Robotics Challenge Finals (DRC-Finals). First, hardware updates in the main body and the perception head of the contestant robot, DRC-Hubo+, were presented by comparison with the previous platform (DRC-Hubo in DRC-Trials). Then, the control system which enabled the full-sized humanoid to drive the off-the-shelf utility vehicle in the competition was provided. For this, the sensor data fusion process and the advanced driving assistant techniques were emphasized. Next, the driving pattern analysis of other type operators (such as human-drivers or pure tele-operation systems) was provided to demonstrate the performance of the developed system. Lastly, test and evaluation of the built robot and control system was provided via experimentation which DRC-Hubo+ drove the vehicle in a real-world setting. The presented approach was also verified in DRC-Finals as the author’s team was placed in the top record at the driving task in the competition.