Title: Dynamic Modeling, Design and Control of Wire-Borne Underactuated Brachiating Robots: Theory and Application

Date: Wednesday, June 23, 2021
Time: 2:00 pm - 4:00 pm (EST)
Location: BlueJeans meeting (https://bluejeans.com/816526798)

Siavash Farzan
Robotics Ph.D. Candidate
School of Electrical and Computer Engineering
Georgia Institute of Technology

Committee:
Dr. Ai-Ping Hu (Advisor) - Georgia Tech Research Institute, Georgia Institute of Technology
Dr. Jonathan Rogers (Advisor) - School of Aerospace Engineering, Georgia Institute of Technology
Dr. Seth Hutchinson - School of Interactive Computing, Georgia Institute of Technology
Dr. Nader Sadegh - School of Mechanical Engineering, Georgia Institute of Technology
Dr. Anirban Mazumdar - School of Mechanical Engineering, Georgia Institute of Technology

Abstract:
The ability of mobile robots to locomote safely in unstructured environments will be a cornerstone of robotics of the future. Introducing robots into fully unstructured environments is known to be a notoriously difficult problem in the robotics field. As a result, many of today's mobile robots are confined to prepared level surfaces in laboratory settings or relatively controlled environments only. One avenue for deploying mobile robots into unstructured settings is to utilize elevated wire networks. The research conducted under this thesis lays the groundwork for developing a new class of wire-borne underactuated robots that employs brachiation — swinging like an ape — as a means of locomotion on flexible cables.
Executing safe brachiation maneuvers with a cable-suspended underactuated robot is a challenging problem due to the complications induced by the cable dynamics and vibrations. This thesis studies, from concept through experiments, the dynamic modeling techniques and control algorithms for wire-borne underactuated brachiating robots, to develop advanced locomotion strategies that enable the robots to perform energy-efficient and robust brachiation motions on flexible cables. High-fidelity and approximate dynamic models are derived for the robot-cable system, which provide the ability to model the interactions between the cable and the robot and to include the flexible cable dynamics in the control design. An optimal trajectory generation framework is presented in which the flexible cable dynamics are explicitly accounted for when designing the optimal swing trajectories. By employing a variety of control-theoretic methods such as robust and adaptive estimation, control Lyapunov and barrier functions, semidefinite programming and sum-of-squares optimization, a set of closed-loop control algorithms are proposed. A novel hardware brachiating robot design and embodiment are presented, which incorporate unique mechanical design features and provide a reliable testbed for experimental validation of the wire-borne underactuated brachiating robots. Extensive simulation results and hardware experiments demonstrate that the proposed multi-body dynamic models, trajectory optimization frameworks, and feedback control algorithms prove highly useful in real world settings and achieve reliable brachiation performance in the presence of uncertainties, disturbances, actuator limits and safety constraints.