5200 N Lake Rd, Merced, CA 95343

https://eecs.ucmerced.edu/
View map Add to calendar

Azin Shamshirgaran

Electrical Engineering and Computer Science (EECS)

Ph.D. Dissertation Defense

 

Multi-Robot Environmental Map Learning with Budget Constraints Using Reinforcement Learning

 

Abstract
   This dissertation explores decision-making under uncertainty in robotics systems tasked with reconstructing a scalar field through sensing. In this task, each robot must determine its next decision considering surrounding uncertainties, and environmental and physical constraints. The complexity escalates in a multi-agent scenario, as each robot must not only assess its course of action but also predict and consider other robots' movements and plans. Our interest in this problem is motivated by applications in precision agriculture, where robots are used to collect measurements to estimate domain-relevant scalar parameters such as soil moisture or nitrates concentrations. In particular, I focused on the implementation of budget-aware algorithms, therefore casting the problem as an instance of constrained optimization. For this approach to be efficient, it is necessary for robots to coordinate their efforts to avoid unnecessary duplicate work or negative interference.
   In this talk, I will present a single robot Informative path planning followed by a multi-robot scenario.
   

Biography
   Azin Shamshirgaran is a PhD candidate in the Computer Science and Engineering department at UC, Merced under the guidance of Prof. Stefano Carpin. Azin’s research interests lie in the broad field of robotics, reinforcement learning and machine learning. She has been funded by USDA-NIFA (National Robotics Initiative) and IoT4Ag Engineering Research Center (funded by NSF).

Event Details

See Who Is Interested

1 person is interested in this event

User Activity

No recent activity

University of California Merced Events Calendar Powered by the Localist Community Event Platform © All rights reserved