Presentation Title

Multi-robot informed path planning under communication constraints

Advisor Information

Prithviraj Dasgupta

Location

Dr. C.C. and Mabel L. Criss Library

Presentation Type

Poster

Start Date

4-3-2016 10:45 AM

End Date

4-3-2016 12:15 PM

Abstract

Autonomous exploration using multiple robots is a rapidly evolving technology that can reduce the dangers to humans while exploring unstructured environments such as extraterrestrial planetary surfaces, or, building structures following a disaster scene. In both cases, the environment is initially unknown, and likely dangerous for humans to enter and maneuver within. Robots can explore and provide critical information back to the human operators, allowing for more specialized deployments of human teams, reducing the risk those teams are put into. In this work, we propose a preliminary approach to exploration of unknown and unstructured environments using multiple robots in communication constrained environments. In our proposed approach, we plan to use Gaussian Processes to model the dynamics of the information gathered by the robots and the likelihood of communication. Based on these two Gaussian Processes, we will select candidate points and go to the point that best balances the objectives of information gain and communication of information to a base station, with a bias towards communications the longer the robot goes without communicating samples back to the base station. We plan to evaluate this approach in simulation and on physical robots available in the CMANTIC Robotics Lab.

This document is currently not available here.

COinS
 
Mar 4th, 10:45 AM Mar 4th, 12:15 PM

Multi-robot informed path planning under communication constraints

Dr. C.C. and Mabel L. Criss Library

Autonomous exploration using multiple robots is a rapidly evolving technology that can reduce the dangers to humans while exploring unstructured environments such as extraterrestrial planetary surfaces, or, building structures following a disaster scene. In both cases, the environment is initially unknown, and likely dangerous for humans to enter and maneuver within. Robots can explore and provide critical information back to the human operators, allowing for more specialized deployments of human teams, reducing the risk those teams are put into. In this work, we propose a preliminary approach to exploration of unknown and unstructured environments using multiple robots in communication constrained environments. In our proposed approach, we plan to use Gaussian Processes to model the dynamics of the information gathered by the robots and the likelihood of communication. Based on these two Gaussian Processes, we will select candidate points and go to the point that best balances the objectives of information gain and communication of information to a base station, with a bias towards communications the longer the robot goes without communicating samples back to the base station. We plan to evaluate this approach in simulation and on physical robots available in the CMANTIC Robotics Lab.