Research Projects

ChibiFace: A Sensor-Rich Android Tablet-Based Interface for Industrial Robotics

Tablet sports Android Lollipop OS with OpenCM PocketSphinx and OpenCV for speech and image processing. The tablet connects to a laptop with Windows 7 OS that runs Python and VLA3(Stäubli Robotics Suite) to provide the control of the Stäubli industrial robot.

Vital Sign Monitoring Framework

Non-contact extraction of vital signs using RGB and thermal images obtained from a smart phone. Using this method, heart rate, respiratory rate and forehead temperature can be measured concurrently.

MOSES: A Matlab-Based Open-Source Stochastic Epidemic Simulator

Development of a computationally efficient open-source simulator for reproducing different epidemic model behaviors for control system development to contain epidemics.

Advanced Control of Variable Stiffness Actuated Robots

Variable stiffness actuation (VSA) emerged as a new paradigm in the quest for creating anthropomorphic robots with higher dexterity, efficiency and safety. VSA robots are intrinsically compliant. Elastic elements are incorporated to the joints to store energy and also isolate the link from the motors. Due to the compliant elements, the dynamics of VSA robots is complex and highly nonlinear. This becomes a challenge for the control system design, since the controller needs to handle a nonlinear, highly constrained system with short sampling intervals.

Locomotion Strategy Selection for a Hybrid Legged Wheeled Robot

In this work we developed depth image based locomotion strategy selection framework for a hybrid mobile robot. Equipped with four legs and wheels, the hybrid mobile robot is capable of environment-based locomotion strategy switching for efficient terrain navigation. Terrain recognition is the critical component of the system which classifies depth images into terrain types in real-time and selects different locomotion mode sub-controllers accordingly. Researchers achieved 97% accuracy for the five-class terrain classification problem. The framework was evaluated in set of real-world experiments conducted in mixed terrain environment.

RRT*FN – memory efficient sampling-based motion planning algorithm

RRT (Rapidly-Exploring Random Tree) is a sampling-based algorithm for solving path planning problem. RRT provides feasible solution if time of RRT tends to infinity.
RRT* is a probabilistic algorithm for solving motion planning problem, which is similar to RRT but unlike RRT it provides higher rate of convergence to the feasible solution.
RRT*FN also is a probabilistic algorithm based on RRT*.RRT*FN inherents faster rate of convergence to the feasable solution,however RRT*FN implements it using less memory.

Inertial Motion Capture Based Tele-operation of a Mobile Robot Manipulator

This work presents a human-robot interface system incorporating a mobile manipulator and full-body inertial motion capture system. Intuitive robot tele-operation is achieved through the ability of the platform to follow the motion of a human operator. Moreover the manipulator installed on top of the mobile robot mimics the motion of the right hand of the operator.

Enhanced Object Manipulation Using Multigrasp Robotic Hand for Intelligent Industrial Automation

Even though there are various advanced robot manipulators for industrial object manipulation, one of the major obstacles for intelligent industrial automation is the lack of multigrasp end-effectors (grippers), which can manipulate complex shaped objects. Recently, both research groups and commercial companies introduced several multi-fingered hand prostheses, which replicate functionality of a human hand for grasping and holding objects, performing precise operations, etc.