UAV Target Tracking in Urban Environments Using Deep Reinforcement\n Learning

2020/07/21 by Sarthak Bhagat, Bhagat, Sarthak, Sujit PB +1 · 1 citation
Computer Science · Engineering · #FOS: Computer and information sciences #FOS: Electrical engineering #Robotics (cs.RO) #Robotics and Sensor-Based Localization #Systems and Control (eess.SY) #UAV Applications and Optimization #Video Surveillance and Tracking Methods #electronic engineering #information engineering

paper · pdf · doi:10.48550/arxiv.2007.10934

openalex publication_date 2020/07/21 · openalex created_date 2022/07/26 · openalex updated_date 2026/07/28

Abstract

Persistent target tracking in urban environments using UAV is a difficult\ntask due to the limited field of view, visibility obstruction from obstacles\nand uncertain target motion. The vehicle needs to plan intelligently in 3D such\nthat the target visibility is maximized. In this paper, we introduce Target\nFollowing DQN (TF-DQN), a deep reinforcement learning technique based on Deep\nQ-Networks with a curriculum training framework for the UAV to persistently\ntrack the target in the presence of obstacles and target motion uncertainty.\nThe algorithm is evaluated through several simulation experiments qualitatively\nas well as quantitatively. The results show that the UAV tracks the target\npersistently in diverse environments while avoiding obstacles on the trained\nenvironments as well as on unseen environments.\n

Cited by

• Occlusion-Aware Ground Target Tracking by a Dubins Vehicle using Visibility Volumes

Related

• Target-Following Double Deep Q-Networks for UAVs
• Online Deep Reinforcement Learning for Autonomous UAV Navigation and Exploration of Outdoor Environments
• Autonomous UAV Navigation: A DDPG-based Deep Reinforcement Learning Approach
• A Vision Based Deep Reinforcement Learning Algorithm for UAV Obstacle\n Avoidance
• Autonomous Quadrotor Landing using Deep Reinforcement Learning