• humanoid robots
• DRC
• teleoperation
• whole-body control
• semi-autonomous control

The aim of this thesis is to design and develop a modular software framework for controlling humanoid robots in teleoperation, in a context of disaster-response or civil defense. Over the years, natural (earthquakes, floods, etc.) or man-made disasters (nuclear reactor meltdowns, terrorist attacks, etc.) have caused several victims. The state of the art of disaster-robotics allows to deploy efficient and powerful robots in order to assist and support humans in the delicate phases of searching and rescuing survivors. In particular, with the use of teleoperation, the inclusion of a human operator (human-in-the-loop) can dramatically promote the application of humanoid robots, due to the human superior competence in critical thinking and context-awareness. This way, robots can be used as an interface between man and environment. Under these concepts, the thesis work focused on the design of a robust and efficient control architecture that brings whole-body locomotion and manipulation capabilities to the robot. Specifically, this thesis dealt with the development of a software module for teleoperating a robot while it is in a vehicle, making it able to drive. The module internal architecture is structured as a Finite State Machine, which allows to model a workflow of behaviors in an event-driven manner, providing safe and robust control in a teleoperation scenario. The effectiveness of the developed software has been validated during the DARPA Robotics Challenge Finals, occured in Pomona, CA (USA), on June 5-6 of 2015.