Scholarlink

<p>Unmanned Aerial Vehicles (UAVs) are increasingly recognized as one of the key enabling technologies for various emerging applications and services. In this paper, we investigate the use of UAV technology for real-time target detection applications within indoor environments under dynamic and uncertain target movements, constrained UAV battery capacity, and the presence of indoor obstacles. An important challenge in this domain is how to support prolonged UAV operation, given their limited battery capacity and the need to return to a charging station (CHS) to recharge while achieving reliable detection of moving targets in obstacle-rich indoor environments. To this end, this paper develops an intelligent UAV-based target detection system that integrates an obstacle- and energy-aware CHS placement strategy with a reinforcement learning (RL)–based target detection mechanism. The CHS placement strategy aims to determine the optimal number and placement of CHSs by using a genetic algorithm (GA) combined with <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(A^*\)</EquationSource> </InlineEquation> pathfinding while considering a set of design constraints. Based on the output of the CHS placement strategy, we develop an adaptive RL model to interact with the operating environment and learn the target movement uncertainties over time. By optimizing CHS placement and using an RL-driven target detection algorithm, our proposed system extends UAV operation with minimal recharging interruptions, and hence allows the UAV to execute appropriate energy-efficient flying actions that ensure high detection accuracy. Simulation results indicate that, compared to a reference RL-based target detection system, our proposed system significantly improves the detection rate by up to 57%, while reducing the per-detection energy consumption by around 74%, which consequently improves the system efficiency. These results demonstrate the real-world effectiveness of the proposed target detection system in indoor scenarios.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Integrated GA-based charging station distribution and ML-driven UAV-based real-time security monitoring in consumer environments

  • 期刊论文

  • 发表日期：

    2025年11月19日

查看全文
  • Haythem Bany Salameh,
  • Samah Al-Saaideh,
  • Ahmed Al-Ajlouni

摘要

Unmanned Aerial Vehicles (UAVs) are increasingly recognized as one of the key enabling technologies for various emerging applications and services. In this paper, we investigate the use of UAV technology for real-time target detection applications within indoor environments under dynamic and uncertain target movements, constrained UAV battery capacity, and the presence of indoor obstacles. An important challenge in this domain is how to support prolonged UAV operation, given their limited battery capacity and the need to return to a charging station (CHS) to recharge while achieving reliable detection of moving targets in obstacle-rich indoor environments. To this end, this paper develops an intelligent UAV-based target detection system that integrates an obstacle- and energy-aware CHS placement strategy with a reinforcement learning (RL)–based target detection mechanism. The CHS placement strategy aims to determine the optimal number and placement of CHSs by using a genetic algorithm (GA) combined with \(A^*\) pathfinding while considering a set of design constraints. Based on the output of the CHS placement strategy, we develop an adaptive RL model to interact with the operating environment and learn the target movement uncertainties over time. By optimizing CHS placement and using an RL-driven target detection algorithm, our proposed system extends UAV operation with minimal recharging interruptions, and hence allows the UAV to execute appropriate energy-efficient flying actions that ensure high detection accuracy. Simulation results indicate that, compared to a reference RL-based target detection system, our proposed system significantly improves the detection rate by up to 57%, while reducing the per-detection energy consumption by around 74%, which consequently improves the system efficiency. These results demonstrate the real-world effectiveness of the proposed target detection system in indoor scenarios.