검색 상세

UAV-BS Operation Plan for Unified Communication and Localization in a GPS-denied Environment

초록/요약

Unmanned Aerial Vehicles (UAVs), also commonly referred to as drones, are now playing an important role in our lives. Unmanned aerial vehicles have emerged as crucial tools for military operations as well as civilian usage because of their agility, quicker deployment, and versatility. In civilian contexts, UAVs are particularly valuable for public safety and disaster response, where they can serve as airborne base stations (UAV-BSs) to restore critical communication and localization services when the cellular infrastructure is damaged/destroyed and the environment is GPS-denied. For wireless communications, UAVs can be deployed as an airborne base-station to expand network connectivity and greater capacity. When terrestrial networks are either damaged or non-existent, and the area is GPS- and wireless communications-denied, the UAV can be quickly deployed to serve communication and localization applications to ground terminals in a given target area. In this dissertation, we focus on the design of an unmanned aerial vehicle which is deployed as a flying base station or a base station in the sky to provide communication and localization services to terminals in an environment that lacks wireless connectivity and is GPS-denied. We model an intelligent UAV operation plan for unified communication and localization using reinforcement learning in a suburban environment which has no cellular communication and GPS connectivity. We proposed two models. 1) UAV-BS for unified communication and localization using RL (UCL-RL), and 2) Enhanced UCL-RL (EU) model. In UCL-RL, the UAV flies to the target area, moves in a circular fashion with a constant turning radius and sends navigation signals from different altitudes to the ground terminals. This provides a dynamic environment that includes the turning radius, the navigation signal transmission points, and the altitude of the unmanned aerial vehicle as well as the location of ground terminals where reinforcement learning is applied to learn intelligently. The UAV continuously interacts with the environment and learns the optimal height that provides the best communication and localization services to the ground terminals. Enhanced UCL-RL provides an enhanced model by extending the UCL-RL model. In the enhanced model, we consider a variable turning radius of the UAV-BS. The UAV-BS is deployed to the target area and placed within the minimum and maximum UAV-BS altitude and radius. The UAV-BS moves in a circular path with varying turning radius and varying height to send navigation signals to the terminals in the target area. The combination of the variable UAV-BS variable turning radius and height in addition to the navigation signal transmission points and the position of the ground terminals provide a dynamic environment where Deep Q-network reinforcement algorithm is applied for intelligent decision making. The UAV-BS though the DQN algorithm interacts with the environment continuously and learns optimal parameters for communication and localization. The dilution of precision is measured as a reward for the localization, and an average path loss is assessed using the air to ground model to evaluate the communication capability. We conducted simulations to evaluate the UCL-RL and EU schemes. The simulation results have shown that the models provide improved terminal positioning accuracy and guarantee communication service to the terminals in a GPS- and wireless-denied suburban environment.

more

목차

1. Introduction 1
1.1 Background 1
1.2 Contributions 2
1.3 Organization 3
2. Unmanned Aerial Vehicle (UAV) 4
2.1 Evolution of UAVs: An Overview 4
2.2 Applications of UAVs 5
2.3 Wireless Communication Roles of UAVs 5
2.4 UAV-BS for Communication and Localization 6
2.4.1 UAV-BS for Communication 7
2.4.2 UAV-BS for Localization 11
2.4.3 Dilution of Precision 15
2.5 Reinforcement Learning 17
3. UAV-BS Operation Plan for Unified Communication and Localization 23
3.1 Related Works 24
3.2 UCL-RL System Model 27
3.2.1 Positioning Scheme 28
3.2.2 Communication Scheme 32
3.2.3 Unified Communication and Localization Scheme 35
3.3 Communication and Localization using Q-Learning 35
3.4 Simulation Setup and Performance Evaluation 41
3.4.1 Simulation Environment Setup 41
3.4.2 Performance Evaluation 43
3.5 Chapter Summary 49
4. Enhanced UAV-BS Deployment Plan for Communication and Localization. 50
4.1 Related Works 50
4.2 EU System Model 52
4.2.1 EU Model Positioning Scheme 53
4.2.2 EU Model Communication Scheme 56
4.2.3 Unified Communication and Positioning 57
4.3 Communication and Localization using DQN 57
4.4 Simulation Setup and Performance Evaluation 65
4.4.1 Simulation Environment Setup 65
4.4.2 Performance Evaluation 72
4.5 Chapter Summary 77
5. Conclusions 79
References 81
Appendix 88
A.1 Derivation of Eq. (3.8) 88

more
반출 Meta View 목록