{"id":{"repo_id":"unr","oai_identifier":"oai:scholarwolf.unr.edu:11714/3101"},"canonical_url":"https://search.dev.ndltd.org/etd/unr/oai:scholarwolf.unr.edu:11714/3101","repository":{"repo_id":"unr","name":"University of Nevada - Reno","base_url":"https://scholarwolf.unr.edu/server/oai/request"},"display":{"title":"Intent Recognition in Multi-Agent Domains","abstract":"Intent recognition is an extremely important aspect of social robotics. The ability to recognize and react to intentions is not only an integral part of social interaction, but is also useful in adversarial domains, as discussed in this thesis. It is especially important to be capable of performing intent recognition in multi-agent settings, both in terms of recognizing low-level intentions for individual agents, and of recognizing coordinated agents of multiple agents. This is because in real-world domains, it is rare to be performing tasks which require intent recognition in sparsely populatedenvironments. In this thesis, we present solutions to various aspects of the intent recognition problem. First, we introduce a framework for testing intent recognition systems in the multi-agent domain. This framework is modular in nature, making it easy to make changes to individual pieces of the system, and includes an open-source naval simulator, a low-level intent recognition module, a high-level intent recognition module, and a module for controlling the movements of the simulated agents.This thesis also presents an extension of the work presented in [17, 18] to the multi-agent domain. We solve the problem of applying the hidden Markov formulation of the intent recognition problem to multiple agents while still operating in real-time by parallelizing various steps of the intent recognition algorithm, and relax the constraint that the topology of the HMMs must be designed by hand. In addition to this application of a low-level intent recognition system to the multi-agent domain, we also present a method for recognizing intentions which require cooperation between multiple agents. We do this by encoding high-level intentions in an activation network [1]. This approach addresses some of the drawbacks of traditional plan recognition techniques, including the difficulties presented by searching large plan libraries, and the difficulties in recognizing multiple plans which are being performed simultaneously.","abstract_html":"Intent recognition is an extremely important aspect of social robotics. The ability to recognize and react to intentions is not only an integral part of social interaction, but is also useful in adversarial domains, as discussed in this thesis. It is especially important to be capable of performing intent recognition in multi-agent settings, both in terms of recognizing low-level intentions for individual agents, and of recognizing coordinated agents of multiple agents. This is because in real-world domains, it is rare to be performing tasks which require intent recognition in sparsely populatedenvironments. In this thesis, we present solutions to various aspects of the intent recognition problem. First, we introduce a framework for testing intent recognition systems in the multi-agent domain. This framework is modular in nature, making it easy to make changes to individual pieces of the system, and includes an open-source naval simulator, a low-level intent recognition module, a high-level intent recognition module, and a module for controlling the movements of the simulated agents.This thesis also presents an extension of the work presented in [17, 18] to the multi-agent domain. We solve the problem of applying the hidden Markov formulation of the intent recognition problem to multiple agents while still operating in real-time by parallelizing various steps of the intent recognition algorithm, and relax the constraint that the topology of the HMMs must be designed by hand. In addition to this application of a low-level intent recognition system to the multi-agent domain, we also present a method for recognizing intentions which require cooperation between multiple agents. We do this by encoding high-level intentions in an activation network [1]. This approach addresses some of the drawbacks of traditional plan recognition techniques, including the difficulties presented by searching large plan libraries, and the difficulties in recognizing multiple plans which are being performed simultaneously.","abstract_has_math":false,"creators":["Bigelow, Daniel"],"institution":null,"degree_name":null,"degree_level":"Master's Degree","degree_discipline":null,"degree_department":null,"school":null,"contributors":[],"advisors":["Nicolescu, Monica"],"committee_chairs":[],"committee_members":["Louis, Sushil","Kozubowski, Tomasz"],"year":2013,"date_issued":"2013","date_published":"2013","updated_at":"2026-07-27T21:45:45Z","subjects":["Intent Recognition","Plan Recognition"],"languages":[],"rights":["In Copyright(All Rights Reserved)"],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"http://hdl.handle.net/11714/3101","outbound_label":"Handle","outbound_source":"dc:identifier.uri"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor.advisor","label":"Advisor","values":["Nicolescu, Monica"]},{"key":"dc:contributor.committeemember","label":"Committee Member","values":["Louis, Sushil","Kozubowski, Tomasz"]},{"key":"dc:creator","label":"Author","values":["Bigelow, Daniel"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.accessioned","label":"Dc Date Accessioned","values":["2018-05-01T12:28:02Z"]},{"key":"dc:date.available","label":"Dc Date Available","values":["2018-05-01T12:28:02Z"]},{"key":"dc:date.issued","label":"Date","values":["2013"]},{"key":"dc:type","label":"Dc Type","values":["Thesis"]},{"key":"thesis:degree_level","label":"Degree Level","values":["Master's Degree"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Intent Recognition","Plan Recognition"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:rights","label":"Dc Rights","values":["In Copyright(All Rights Reserved)"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier.uri","label":"Identifier URI","values":["http://hdl.handle.net/11714/3101"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["Intent recognition is an extremely important aspect of social robotics. The ability to recognize and react to intentions is not only an integral part of social interaction, but is also useful in adversarial domains, as discussed in this thesis. It is especially important to be capable of performing intent recognition in multi-agent settings, both in terms of recognizing low-level intentions for individual agents, and of recognizing coordinated agents of multiple agents. This is because in real-world domains, it is rare to be performing tasks which require intent recognition in sparsely populatedenvironments. In this thesis, we present solutions to various aspects of the intent recognition problem. First, we introduce a framework for testing intent recognition systems in the multi-agent domain. This framework is modular in nature, making it easy to make changes to individual pieces of the system, and includes an open-source naval simulator, a low-level intent recognition module, a high-level intent recognition module, and a module for controlling the movements of the simulated agents.This thesis also presents an extension of the work presented in [17, 18] to the multi-agent domain. We solve the problem of applying the hidden Markov formulation of the intent recognition problem to multiple agents while still operating in real-time by parallelizing various steps of the intent recognition algorithm, and relax the constraint that the topology of the HMMs must be designed by hand. In addition to this application of a low-level intent recognition system to the multi-agent domain, we also present a method for recognizing intentions which require cooperation between multiple agents. We do this by encoding high-level intentions in an activation network [1]. This approach addresses some of the drawbacks of traditional plan recognition techniques, including the difficulties presented by searching large plan libraries, and the difficulties in recognizing multiple plans which are being performed simultaneously."]},{"key":"dc:format","label":"Dc Format","values":["PDF"]},{"key":"dc:title","label":"Title","values":["Intent Recognition in Multi-Agent Domains"]}]}],"canonical_facts":{"dc:contributor.advisor":["Nicolescu, Monica"],"dc:contributor.committeemember":["Louis, Sushil","Kozubowski, Tomasz"],"dc:creator":["Bigelow, Daniel"],"dc:date.accessioned":["2018-05-01T12:28:02Z"],"dc:date.available":["2018-05-01T12:28:02Z"],"dc:date.issued":["2013"],"dc:description.abstract":["Intent recognition is an extremely important aspect of social robotics. The ability to recognize and react to intentions is not only an integral part of social interaction, but is also useful in adversarial domains, as discussed in this thesis. It is especially important to be capable of performing intent recognition in multi-agent settings, both in terms of recognizing low-level intentions for individual agents, and of recognizing coordinated agents of multiple agents. This is because in real-world domains, it is rare to be performing tasks which require intent recognition in sparsely populatedenvironments. In this thesis, we present solutions to various aspects of the intent recognition problem. First, we introduce a framework for testing intent recognition systems in the multi-agent domain. This framework is modular in nature, making it easy to make changes to individual pieces of the system, and includes an open-source naval simulator, a low-level intent recognition module, a high-level intent recognition module, and a module for controlling the movements of the simulated agents.This thesis also presents an extension of the work presented in [17, 18] to the multi-agent domain. We solve the problem of applying the hidden Markov formulation of the intent recognition problem to multiple agents while still operating in real-time by parallelizing various steps of the intent recognition algorithm, and relax the constraint that the topology of the HMMs must be designed by hand. In addition to this application of a low-level intent recognition system to the multi-agent domain, we also present a method for recognizing intentions which require cooperation between multiple agents. We do this by encoding high-level intentions in an activation network [1]. This approach addresses some of the drawbacks of traditional plan recognition techniques, including the difficulties presented by searching large plan libraries, and the difficulties in recognizing multiple plans which are being performed simultaneously."],"dc:format":["PDF"],"dc:identifier.uri":["http://hdl.handle.net/11714/3101"],"dc:rights":["In Copyright(All Rights Reserved)"],"dc:subject":["Intent Recognition","Plan Recognition"],"dc:title":["Intent Recognition in Multi-Agent Domains"],"dc:type":["Thesis"],"thesis:degree_level":["Master's Degree"]},"updated_at":"2026-07-27T21:45:45Z"}