PhD Candidate
University of British Columbia, Ph.D. in Mechanical Engineering
Simon Fraser University, 2017, M.A.Sc in Engineering Science
Mansoura University, 2012, B.Sc. in Computers & Systems Engineering (with Distinction)

Research Description

PhD. Thesis: Towards Intuitive and Generalizable Framework for Robot Learning from Demonstration

Supervisors: Drs. Elizabeth Croft, Machiel Van der Loos

Robots are increasingly used in many areas in our daily life. The ability for non-experts to deal with these robots will become inevitable in the near future. Robot Learning from Demonstration (LfD) is concerned with programming robots to perform tasks by observing demonstrations from humans. The current state of the art in LfD uses unintuitive teaching interfaces to program the robot and lacks the ability to generalize what the robot has learned to a wider range of tasks. The goal of the proposed research is to build a simple and intuitive teaching interface for robot LfD and to build a library of simple tasks that, if combined, can allow the robot to perform several complex tasks.
We propose using innovative user interfaces for intuitively teaching robots from human demonstrations. By doing this, non-experts can easily teach a “virtual” robot, that is a model of the real one, some tasks, then these skills can be transferred to the real robot. In addition, using such interfaces allows us to collect massive data during the teaching process. These data can be structured as a generalizable taxonomy, that ranges from simple tasks to complex ones. We will use this taxonomy to teach the robot how to learn from its past experience and to be a fast-learner in new tasks.
Our proposed research democratizes access to robotics. Building a generalizable database of tasks with efficient structure will significantly increase the applicability of using robots in many areas. With my proposed teaching interfaces, robots can be more adaptable to the increasing changes needed to accommodate the requirements of customers. The increased adaptability comes from giving users more control over the behaviour of their robots in an intuitive and easy manner. In addition, such interfaces will make it easier for domain experts who do not know how to program to transfer their skills to robots.

MASc. Thesis: Feasibility of Using Force Myography (FMG) for Estimating Hand Force and Wrist Torque

Supervisor: Dr. Carlo Menon

Hand force estimation is critical for applications that involve physical human-machine interactions for force monitoring and machine control. Force Myography (FMG) is a potential technique to be used for estimating hand force/torque. The FMG signals represent the volumetric changes in the arm muscles due to muscle contraction or expansion during force/torque exertion. The aim of this thesis is to explore the suitability of FMG for hand force/torque estimation.
Studying the feasibility of using FMG for torque estimation was preliminary investigated by using 1-DOF torque sensor for labeling the FMG during torque exertion. A custom designed force-sensing resistors (FSRs) band was donned on the forearm muscle belly for measuring FMG signals, while the participants exerted torque around three axes. A regression model was created for each torque axis and trained using the corresponding data. The average R2 was 0.89 for pronation-supination, flexion-extension, and radial-ulnar deviations.
Using 1-DOF torque sensor for labeling the data needs a new custom-rig for capturing each torque axis. To overcome this limitation, a 6-DOF force/torque load cell was used for labeling the FMG data during force/torque exertion in any direction. In addition, a total number of 60 FSRs were embedded into four bands to be worn on the arm for measuring FMG signals during force/torque exertion. Healthy participants were recruited in this study and were asked to exert isometric force along three perpendicular axes, torque about the same three axes, and force and torque freely in any direction. Three cases were considered to explore the performance of the FMG bands in estimating force/torque in single- and multi- axis. These cases are: (1) 6 axes force/torque individually; (2) 3-DOF force and 3-DOF torque; and (3) 6-DOF force and torque simultaneously. In addition, a comparison between all possible combinations of the four bands was held to provide guidelines about the best placement of the FMG measurements in each case.

Contact Details

ICICS Building X015

a place of mind, The University of British Columbia

Faculty of Applied Science
5000 - 2332 Main Mall,
Vancouver, BC, V6T 1Z4, Canada
Tel: 604.822.6413
Email:
CARIS Lab
Department of Mechanical Engineering, UBC,
Vancouver, BC, Canada
Tel: 604.822.3147
Fax: 604.822.2403
See contact page for addresses.

Emergency Procedures | Accessibility | Contact UBC  | © Copyright The University of British Columbia