Multi-agent planning for coordinated robotic weed killing

McAllister, Wyatt Spalding

Permalink

https://hdl.handle.net/2142/108240 Copy

Description

Title

Multi-agent planning for coordinated robotic weed killing

Author(s)

McAllister, Wyatt Spalding

Issue Date

2020-05-08

Director of Research (if dissertation) or Advisor (if thesis)

Chowdhary, Girish

Doctoral Committee Chair(s)

Chowdhary, Girish

Committee Member(s)

• Davis, Adam
• Srikant, Rayadurgam
• Belabbas, Mohamed Ali

Department of Study

Electrical & Computer Eng

Discipline

Electrical & Computer Engr

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Date of Ingest

2020-08-27T00:49:52Z

Keyword(s)

Multi-agent, coordinated robotics, industrial agriculture

Abstract

This work presents techniques for predictive modeling of weed growth, as well as an improved planning index to be used in conjunction with these techniques, for the purpose of improving the performance of coordinated weeding algorithms being developed for industrial agriculture. We demonstrate that the evolving Gaussian process method applied to measurements from the agents can predict the evolution of the field within the realistic simulation environment Weed World. In addition to prediction, this method provides physical insight into the seed bank distribution of the field. In this work we extend the evolving Gaussian process model in two important ways. First, we have developed a model that has a bias term, and we show how it is connected to the seed bank distribution. Secondly, we show that one may decouple the component of the model representing weed growth from the component which varies with the seed bank distribution, and adapt the latter online. We compare this predictive approach with one that relies on known properties of the weed growth model, and show that the evolving Gaussian process method gives better results, even without assuming this model information. Finally, we use an improved planning index, entropic value-at-risk (EVaR) in conjunction with the Whittle index, which allows a balanced trade-off between exploration and exploitation, and ensures model improvement when used with these various prediction schemes.

Graduation Semester

2020-05

Type of Resource

Thesis

Permalink

http://hdl.handle.net/2142/108240

Copyright and License Information

Copyright 2020 Wyatt McAllister

Owning Collections