{"title":"Situation-aware deep reinforcement learning for autonomous nonlinear mobility control in cyber-physical loitering munition systems","authors":"Hyunsoo Lee;Soyi Jung;Soohyun Park","doi":"10.23919/JCN.2025.000001","DOIUrl":null,"url":null,"abstract":"With the rapid development of autonomous mobility technologies, drones are now widely used in many applications, including military domain. Particularly in battlefield conditions, designing a deep reinforcement learning (DRL)-based autonomous control algorithm presents significant challenges due to the need for real-time and adjustable nonlinear trajectory planning. Therefore, this paper introduces a novel situation-aware DRL-based autonomous nonlinear drone mobility control algorithm tailored for cyber-physical loitering munition applications. The proposed DRL-based drone mobility control algorithm is crafted with a focus on real-time situation-aware operations, enabling it to navigate through many obstacles encountered on the battlefield efficiently. For efficient observation and intuitive fast understanding of time-varying real-time situations, this paper presents an algorithm that works on a cyber-physical virtual battlefield environment using Unity. In detail, our proposed DRL-based nonlinear drone mobility control algorithm utilizes situation-aware sensing components that are implemented with a Raycast function in Unity virtual scenarios. Based on the gathered situation-aware information, the drone can autonomously and nonlinearly adjust its trajectory during flight. Thus, this approach is obviously beneficial for avoiding obstacles in complex and unpredictable battlefields. Our visualization- based performance evaluation shows that the proposed algorithm outperforms other mobility control algorithms, with an average performance nearly twice as high when the obstacle density is 50%. This superiority is further evidenced by the detailed trajectory planning presented.","PeriodicalId":54864,"journal":{"name":"Journal of Communications and Networks","volume":"27 1","pages":"10-22"},"PeriodicalIF":2.9000,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10923673","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of Communications and Networks","FirstCategoryId":"94","ListUrlMain":"https://ieeexplore.ieee.org/document/10923673/","RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"COMPUTER SCIENCE, INFORMATION SYSTEMS","Score":null,"Total":0}
引用次数: 0
Abstract
With the rapid development of autonomous mobility technologies, drones are now widely used in many applications, including military domain. Particularly in battlefield conditions, designing a deep reinforcement learning (DRL)-based autonomous control algorithm presents significant challenges due to the need for real-time and adjustable nonlinear trajectory planning. Therefore, this paper introduces a novel situation-aware DRL-based autonomous nonlinear drone mobility control algorithm tailored for cyber-physical loitering munition applications. The proposed DRL-based drone mobility control algorithm is crafted with a focus on real-time situation-aware operations, enabling it to navigate through many obstacles encountered on the battlefield efficiently. For efficient observation and intuitive fast understanding of time-varying real-time situations, this paper presents an algorithm that works on a cyber-physical virtual battlefield environment using Unity. In detail, our proposed DRL-based nonlinear drone mobility control algorithm utilizes situation-aware sensing components that are implemented with a Raycast function in Unity virtual scenarios. Based on the gathered situation-aware information, the drone can autonomously and nonlinearly adjust its trajectory during flight. Thus, this approach is obviously beneficial for avoiding obstacles in complex and unpredictable battlefields. Our visualization- based performance evaluation shows that the proposed algorithm outperforms other mobility control algorithms, with an average performance nearly twice as high when the obstacle density is 50%. This superiority is further evidenced by the detailed trajectory planning presented.
期刊介绍:
The JOURNAL OF COMMUNICATIONS AND NETWORKS is published six times per year, and is committed to publishing high-quality papers that advance the state-of-the-art and practical applications of communications and information networks. Theoretical research contributions presenting new techniques, concepts, or analyses, applied contributions reporting on experiences and experiments, and tutorial expositions of permanent reference value are welcome. The subjects covered by this journal include all topics in communication theory and techniques, communication systems, and information networks. COMMUNICATION THEORY AND SYSTEMS WIRELESS COMMUNICATIONS NETWORKS AND SERVICES.