Graduate Thesis Or Dissertation

 

Cooperative Robot Localization Using Event-Triggered Estimation Público Deposited

https://scholar.colorado.edu/concern/graduate_thesis_or_dissertations/ng451h96t
Abstract
  • It is known that multiple robot systems that need to cooperate to perform certain activities or tasks incur in high energy costs that hinder their autonomous functioning and limit the benefits provided to humans by these kinds of platforms. This work presents a communications-based method for cooperative robot localization. Implementing concepts from event-triggered estimation, used with success in the field of wireless sensor networks but rarely to do robot localization, agents are able to only send measurements to their neighbors when the expected novelty in this information is high. Since all agents know the condition that triggers a measurement to be sent or not, the lack of a measurement is therefore informative and fused into state estimates. In the case agents do not receive either direct nor indirect measurements of all others, the agents employ a covariance intersection fusion rule in order to keep the local covariance error metric bounded. A comprehensive analysis of the proposed algorithm and its estimation performance in a variety of scenarios is performed, and the algorithm is compared to similar cooperative localization approaches. Extensive simulations are performed that illustrate the effectiveness of this method.
Creator
Date Issued
  • 2017
Academic Affiliation
Advisor
Committee Member
Degree Grantor
Commencement Year
Subject
Última modificação
  • 2019-11-18
Resource Type
Declaração de direitos
Language

Relações

Itens