The need for Unmanned Aerial System (UASs) is increasing in the fields of science, security, military and others. With an increasing number of challenges faced by mankind the need for system with advance autonomous and intelligent capabilities is growing rapidly. The required intelligent behavior can be achieved in an unmanned system with machine learning methods such as Reinforcement Learning (RL) theory. This theory presents algorithms which have the capability to learn how to adapt to its environment. Using RL methods, autonomous systems have the ability to use past experiences to improve future mission performance. In this research, the effects of rewards on the ability of an autonomous UAV controlled by a Reinforcement Learning agent to accomplish a target localization task were investigated. The numerical values of the reward scheme were varied. It was expected that with an increase in the reward obtained by a learning agent upon correct detection the systems will become more risk tolerant and vice-versa. It was also predicted that the systems will be more efficient and would have a tendency to locate targets faster with an increase in the sensor sensitivity values after the system achieves steady-state performance. The results were analyzed statistically to detect the effects and confirm the hypothesis.
","abstract_html":"<p>The need for Unmanned Aerial System (UASs) is increasing in the fields of science, security, military and others. With an increasing number of challenges faced by mankind the need for system with advance autonomous and intelligent capabilities is growing rapidly. The required intelligent behavior can be achieved in an unmanned system with machine learning methods such as Reinforcement Learning (RL) theory. This theory presents algorithms which have the capability to learn how to adapt to its environment. Using RL methods, autonomous systems have the ability to use past experiences to improve future mission performance. In this research, the effects of rewards on the ability of an autonomous UAV controlled by a Reinforcement Learning agent to accomplish a target localization task were investigated. The numerical values of the reward scheme were varied. It was expected that with an increase in the reward obtained by a learning agent upon correct detection the systems will become more risk tolerant and vice-versa. It was also predicted that the systems will be more efficient and would have a tendency to locate targets faster with an increase in the sensor sensitivity values after the system achieves steady-state performance. The results were analyzed statistically to detect the effects and confirm the hypothesis.</p>","abstract_has_math":false,"creators":["Virani, Hemali Dinesh"],"institution":null,"degree_name":"Master of Science in Human Factors & Systems","degree_level":"Thesis - Open Access","degree_discipline":"Human Factors and Systems","degree_department":null,"school":null,"contributors":[],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2014,"date_issued":"2014-10-01T07:00:00Z","date_published":"2014-10-01T07:00:00Z","updated_at":"2026-07-27T19:26:08Z","subjects":["UAV","drones","reinforcement learning","target localization","Aeronautical Vehicles","Aviation"],"languages":[],"rights":[],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"https://commons.erau.edu/edt/257","outbound_label":"Repository record","outbound_source":"dc:identifier"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:creator","label":"Author","values":["Virani, Hemali Dinesh"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"thesis:degree_discipline","label":"Discipline","values":["Human Factors and Systems"]},{"key":"thesis:degree_level","label":"Degree Level","values":["Thesis - Open Access"]},{"key":"thesis:degree_name","label":"Degree Name","values":["Master of Science in Human Factors & Systems"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["UAV","drones","reinforcement learning","target localization","Aeronautical Vehicles","Aviation"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["https://commons.erau.edu/edt/257"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["The need for Unmanned Aerial System (UASs) is increasing in the fields of science, security, military and others. With an increasing number of challenges faced by mankind the need for system with advance autonomous and intelligent capabilities is growing rapidly. The required intelligent behavior can be achieved in an unmanned system with machine learning methods such as Reinforcement Learning (RL) theory. This theory presents algorithms which have the capability to learn how to adapt to its environment. Using RL methods, autonomous systems have the ability to use past experiences to improve future mission performance. In this research, the effects of rewards on the ability of an autonomous UAV controlled by a Reinforcement Learning agent to accomplish a target localization task were investigated. The numerical values of the reward scheme were varied. It was expected that with an increase in the reward obtained by a learning agent upon correct detection the systems will become more risk tolerant and vice-versa. It was also predicted that the systems will be more efficient and would have a tendency to locate targets faster with an increase in the sensor sensitivity values after the system achieves steady-state performance. The results were analyzed statistically to detect the effects and confirm the hypothesis.
"]},{"key":"dc:title","label":"Title","values":["The Effects of Rewards on the Ability of an Autonomous UAV Controlled by a Reinforcement Learning Agent to Accomplish a Target Localization Task"]}]}],"canonical_facts":{"dc:creator":["Virani, Hemali Dinesh"],"dc:description.abstract":["The need for Unmanned Aerial System (UASs) is increasing in the fields of science, security, military and others. With an increasing number of challenges faced by mankind the need for system with advance autonomous and intelligent capabilities is growing rapidly. The required intelligent behavior can be achieved in an unmanned system with machine learning methods such as Reinforcement Learning (RL) theory. This theory presents algorithms which have the capability to learn how to adapt to its environment. Using RL methods, autonomous systems have the ability to use past experiences to improve future mission performance. In this research, the effects of rewards on the ability of an autonomous UAV controlled by a Reinforcement Learning agent to accomplish a target localization task were investigated. The numerical values of the reward scheme were varied. It was expected that with an increase in the reward obtained by a learning agent upon correct detection the systems will become more risk tolerant and vice-versa. It was also predicted that the systems will be more efficient and would have a tendency to locate targets faster with an increase in the sensor sensitivity values after the system achieves steady-state performance. The results were analyzed statistically to detect the effects and confirm the hypothesis.
"],"dc:identifier":["https://commons.erau.edu/edt/257"],"dc:subject":["UAV","drones","reinforcement learning","target localization","Aeronautical Vehicles","Aviation"],"dc:title":["The Effects of Rewards on the Ability of an Autonomous UAV Controlled by a Reinforcement Learning Agent to Accomplish a Target Localization Task"],"thesis:degree_discipline":["Human Factors and Systems"],"thesis:degree_level":["Thesis - Open Access"],"thesis:degree_name":["Master of Science in Human Factors & Systems"]},"updated_at":"2026-07-27T19:26:08Z"}