Defense News

MND News References
Time: 10:00 on April 29, 2025

One. Subject of press conference: Explanation of the ''2025 Military Academy Cup Competition of Cybersecurity'' and ''NDU Participating in the Second Competition of Drones Used for National Defense.''

Two. Chairperson: Ministry of National Defense (MND) Spokesman, Major General SUN LI-FANG

Three. Attendees:
I.Admin. Deputy Minister of MND, Lieutenant General YANG CHI-JUNG
II.Chief of the General Staff, ROCN Command HQ, Vice- Admiral CHIU JUN-JUNG
III.Chief of the Planning Department, ROCAF Command HQ, Major General JIANG YUAN-CHI
IV.Chief of the Logistics Department, ROCAF Command HQ, Major General TIEN ZHUNG-YI
V.Commandant of Chung Cheng Institute of Technology, NDU, Rear-Admiral TSUI I-FONG
VI.Director of Information Security Education and Training Center, Chung Cheng Institute of Technology, NDU, Colonel LAN CHIEN-WU
VII.Director of School of Defense Science, Chung Cheng Institute of Technology, NDU, Colonel LEE YEN-HUNG
VIII.Deputy Director of the Information Security and Cyber Warfare Bureau, DCGS/Communications, Electronics and Information, Colonel CHANG KO-CHIN
IX.Senior Engineer of the Financial Resource Division, Dept./Resource Planning, Mr. FU TSENG-QU
X.Director of Acquisition Management Department, Armaments Bureau, Major General LEE PENG-HSIAO
XI.Deputy Chief of the Joint Operations Planning Division, DCGS/Operations and Planning, Colonel ZHUNG, KUN-CHANG
XII.Deputy Director of the Intelligence Research Center, DCGS/Intelligence, Colonel HU ZHUNG-HUA
XIII.Chief of Protection and Defense Security Division, Political Warfare Bureau, Colonel PENG TSUNG-HAN

Four. News references of press conference
To strengthen the R.O.C. Armed Forces’ independent development in the fields of cybersecurity and drone systems, the National Defense University (NDU) is organizing the “2025 Military Academy Cup Competition of Cybersecurity,” which integrates theory and practice to cultivate cyber warfare talents. In addition, the university is participating in the “Second Competition of Drones Used for National Defense,” guided by the National Science and Technology Council and hosted by National Cheng Kung University. Through the interaction of industry and academia, and cooperation between faculty and students, this initiative aims to enhance the research, development, and application of autonomous combat technologies of drones.

I.2025 Military Academy Cup Competition of Cybersecurity
The Military Academy Cup Competition of Cybersecurity this year is scheduled to be held in the third week of May (May 22–23). Participants are invited from the Army, Navy, and Air Force academies, the Central Police University, the Army Academy R.O.C., the Air Force Institute of Technology, the Management College of NDU, and the Chung Cheng Institute of Technology, NDU. Each institution will assign two instructors and select a team of six outstanding students to form a competition team, totaling eight participating teams. The event is divided into two main parts: a training camp and the actual competition.
(I)Activities of Training Camp: The event includes both comprehensive cybersecurity courses and hands-on training. The comprehensive courses focus on introducing development trends in cybersecurity technology and discussing working knowledge such as website information gathering, reverse engineering, encryption and decryption techniques, and digital forensics. The practical sessions demonstrate the methods and tools used in these technologies, with live instruction and hands-on exercises.
(II)Actual Competition: The competition will be conducted in two phases. In the first phase, theoretical and practical cybersecurity knowledge will be integrated into mission-based scenarios, where participating teams will engage in a challenge of Capture the Flag (CTF) guided by specific task prompts. In the second phase, a themed cyber range environment will be designed based on the competition topics. Each team will take control of the environment and attempt to achieve various mission objectives to earn points. Experts in cybersecurity from industry, government, and academia will be invited to serve as judges, providing both scoring and guidance.
II.NDU Participating in the Second Competition of Drones Used for National Defense
The challenge theme of this year is “Conducting Multi-Target Reconnaissance Missions in a GNSS-Denied Environment.” The scenario simulates conditions where drones are disrupted and unable to use Global Navigation Satellite System (GNSS) signals, and must operate without First Person View (FPV) controls. Under these constraints, the drones are required to autonomously navigate and conduct reconnaissance in an unknown environment, collecting target information such as quantity and location. The key technical challenges include: “autonomous navigation without GNSS,” “autonomous flight and dynamic path planning,” “onboard target reconnaissance, AI-based image recognition, and target localization.” The following section outlines the technical highlights of this competition:
(I)Dead Reckoning Technology: In a GNSS-denied environment, our team adopts the Dead Reckoning (DR) method to estimate new positions. DR is a navigation technique that calculates the current position based on a known starting point, speed, direction, and time. Starting from a known initial position, the system uses data from accelerometers and gyroscopes to obtain changes in velocity and heading. Using this information, theoretical formulas are applied to estimate the new position. This technique offers autonomy, real-time capability, and broad applicability. However, it is prone to cumulative errors, so it must be supplemented with other sensors or periodic calibration to improve accuracy.
(II)Precision Recognition Technology:The system is equipped with an Nvidia Jetson Nano micro-computing platform and a global shutter monocular camera. YOLOv8 is selected as the core object detection algorithm, and the model is further optimized using TensorRT, a deep learning inference acceleration library, to achieve high accuracy and real-time performance under limited computational power. To meet the requirements of the competition, the team built a custom dataset focused on targets such as civilian vehicles, camouflage-painted military vehicles, and fuel drums, enhancing the model's object recognition capabilities.
(III)Target Localization Technology: The YOLO object detection model generates bounding boxes around detected objects, which are used to mark and locate them. By combining DR (Dead Reckoning) and recognition technologies, the system can localize the identified targets.
Through the cybersecurity and drones competition, the goal is not only to stimulate innovative thinking and practical skills among military academy students, but also to deepen the training of cybersecurity professionals and enhance the development and application of unmanned vehicle technologies. These events demonstrate the R.O.C. Armed Forces’ commitment to adopting emerging technologies and cutting-edge advancements. In the future, the NDU will continue to promote industry-academia collaboration, cultivate outstanding defense technology talent, enhance the independent research and development capabilities of the defense industry, and ultimately strengthen overall defense capabilities.