Vivek Thangavelu
Carnegie Bosch Postdoctoral Fellow, Carnegie Mellon University
I am currently building robot systems that can estimate soil contaminant distributions. An adaptive sampling algorithm iteratively samples at locations in the environment based on the sample
information and environment model based on Gaussian Processes. A wheeled robot system is equipped
with a Portable X-ray fluorescence analysis (PXRF) sensor and soil manipulation tools and deployed
in various remote industrial sites
During my doctoral studies, my work focussed on designing decentralized
algorithms for multi-robot systems that
build large-scale structures in remote and unstructured environments,
while guaranteeing certain notions of progress and completion of the cooperative task.
In a nutshell, a hetergenous team of ground-based mobile agents are tasked to build traversable
structures using a mixture of different types of building materials (rigid, compliant and found
stones) which allows them to build structures
many times the size of an individual agent.
I have experience working with both simulation and real-world
robot systems.
My work is loosely organized as below; please click on the icons for more information.
Robot constructing a navigable structure with bags
Heterogeneous robot team constructing with bags and foam blocks
Robot constructing a navigable structure with found stones
Concept design: Robot with material storage (WIP)
Stone dataset for construction, Deposition planning, and Experiments depicting the progression of the construction process
