Object Detection and Reinforcement Learning Approach for Intelligent Control of UAV
Апстракт
In recent years, the development of deep learning models that can generate more accurate predictions and operate in real-time has brought both opportunities and challenges across the various domains of robotic vision. This breakthrough enables researchers to design and deploy more challenging tasks on intelligent mobile robots, which require emphasized abilities of learning and reasoning. In this paper, a new method for intelligent robot control, based on deep learning and reinforcement learning is proposed. The fundamental idea of this work is how the UAV equipped with a monocular camera can learn significant information about the object of interest in the context of its localization and navigation. For such purpose, the object detection system based on Tiny YOLOv2 architecture is employed. Furthermore, bounding box data generated by a convolution neural network is utilized for depth estimation and determining object boundaries. This information has shown how the state-space dimension...s can be significantly reduced, which was essential for further implementation of the Q-learning algorithm. In order to test the proposed framework, a model is developed in MATLAB Simulink. The simulation, which covered different scenarios, was carried out on the UAV within the 3D scene rendered by Unreal Engine. The obtained results have demonstrated the applicability of the proposed methodology for depth estimation, gathering information about the object, object-driven navigation, and autonomous localization and navigation.
Кључне речи:
Unmanned aerial vehicles / Q-learning / Intelligent control / Deep learning / Convolution neural networks / Autonomous localization and navigationИзвор:
Lecture Notes in Networks and Systems, 2022, 472 LNNS, 659-669Издавач:
- Springer Science and Business Media Deutschland GmbH
Финансирање / пројекти:
- Ministry of Education, Science and Technological Development
Колекције
Институција/група
Mašinski fakultetTY - CONF AU - Miljković, Zoran AU - Jevtić, Đorđe PY - 2022 UR - https://machinery.mas.bg.ac.rs/handle/123456789/3823 AB - In recent years, the development of deep learning models that can generate more accurate predictions and operate in real-time has brought both opportunities and challenges across the various domains of robotic vision. This breakthrough enables researchers to design and deploy more challenging tasks on intelligent mobile robots, which require emphasized abilities of learning and reasoning. In this paper, a new method for intelligent robot control, based on deep learning and reinforcement learning is proposed. The fundamental idea of this work is how the UAV equipped with a monocular camera can learn significant information about the object of interest in the context of its localization and navigation. For such purpose, the object detection system based on Tiny YOLOv2 architecture is employed. Furthermore, bounding box data generated by a convolution neural network is utilized for depth estimation and determining object boundaries. This information has shown how the state-space dimensions can be significantly reduced, which was essential for further implementation of the Q-learning algorithm. In order to test the proposed framework, a model is developed in MATLAB Simulink. The simulation, which covered different scenarios, was carried out on the UAV within the 3D scene rendered by Unreal Engine. The obtained results have demonstrated the applicability of the proposed methodology for depth estimation, gathering information about the object, object-driven navigation, and autonomous localization and navigation. PB - Springer Science and Business Media Deutschland GmbH C3 - Lecture Notes in Networks and Systems T1 - Object Detection and Reinforcement Learning Approach for Intelligent Control of UAV EP - 669 SP - 659 VL - 472 LNNS DO - 10.1007/978-3-031-05230-9_79 ER -
@conference{ author = "Miljković, Zoran and Jevtić, Đorđe", year = "2022", abstract = "In recent years, the development of deep learning models that can generate more accurate predictions and operate in real-time has brought both opportunities and challenges across the various domains of robotic vision. This breakthrough enables researchers to design and deploy more challenging tasks on intelligent mobile robots, which require emphasized abilities of learning and reasoning. In this paper, a new method for intelligent robot control, based on deep learning and reinforcement learning is proposed. The fundamental idea of this work is how the UAV equipped with a monocular camera can learn significant information about the object of interest in the context of its localization and navigation. For such purpose, the object detection system based on Tiny YOLOv2 architecture is employed. Furthermore, bounding box data generated by a convolution neural network is utilized for depth estimation and determining object boundaries. This information has shown how the state-space dimensions can be significantly reduced, which was essential for further implementation of the Q-learning algorithm. In order to test the proposed framework, a model is developed in MATLAB Simulink. The simulation, which covered different scenarios, was carried out on the UAV within the 3D scene rendered by Unreal Engine. The obtained results have demonstrated the applicability of the proposed methodology for depth estimation, gathering information about the object, object-driven navigation, and autonomous localization and navigation.", publisher = "Springer Science and Business Media Deutschland GmbH", journal = "Lecture Notes in Networks and Systems", title = "Object Detection and Reinforcement Learning Approach for Intelligent Control of UAV", pages = "669-659", volume = "472 LNNS", doi = "10.1007/978-3-031-05230-9_79" }
Miljković, Z.,& Jevtić, Đ.. (2022). Object Detection and Reinforcement Learning Approach for Intelligent Control of UAV. in Lecture Notes in Networks and Systems Springer Science and Business Media Deutschland GmbH., 472 LNNS, 659-669. https://doi.org/10.1007/978-3-031-05230-9_79
Miljković Z, Jevtić Đ. Object Detection and Reinforcement Learning Approach for Intelligent Control of UAV. in Lecture Notes in Networks and Systems. 2022;472 LNNS:659-669. doi:10.1007/978-3-031-05230-9_79 .
Miljković, Zoran, Jevtić, Đorđe, "Object Detection and Reinforcement Learning Approach for Intelligent Control of UAV" in Lecture Notes in Networks and Systems, 472 LNNS (2022):659-669, https://doi.org/10.1007/978-3-031-05230-9_79 . .