Title

Using Opponent Modeling To Adapt Team Play In American Football

Keywords

Adaptive players; American football; Opponent modeling; Play recognition; Upper confidence bounds applied to trees (UCT)

Abstract

An issue with learning effective policies in multiagent adversarial games is that the size of the search space can be prohibitively large when the actions of both teammates and opponents are considered simultaneously. Opponent modeling, predicting an opponent's actions in advance of execution, is one approach for selecting actions in adversarial settings, but it is often performed in an ad hoc way. This chapter introduces several methods for using opponent modeling, in the form of predictions about the players' physical movements, to learn team policies. To explore the problem of decision making in multiagent adversarial scenarios, we use our approach for both offline play generation and real-time team response in the Rush 2008 American Football Simulator. Simultaneously predicting movement trajectories, future reward, and play strategies of multiple players in real time is a daunting task, but we illustrate how it is possible to divide and conquer this problem with an assortment of data-driven models. © 2014 Elsevier Inc. All rights reserved.

Publication Date

1-1-2014

Publication Title

Plan, Activity, and Intent Recognition: Theory and Practice

Number of Pages

313-341

Document Type

Article; Book Chapter

Personal Identifier

scopus

DOI Link

https://doi.org/10.1016/B978-0-12-398532-3.00013-0

Socpus ID

84902896652 (Scopus)

Source API URL

https://api.elsevier.com/content/abstract/scopus_id/84902896652

This document is currently not available here.

Share

COinS