This paper presents a proof of concept demonstrating the integration of Digital Twins (DTs) into simulation environments to enhance the fidelity and adaptability of autonomous systems modelling. The study explores the implementation of a DT architecture for an Uncrewed Aerial System (UAS), combining a physical quadcopter with its digital counterpart, which includes a flight dynamics model, AI-based control system, and ground control interface. The DT is constructed and continuously refined using real-world telemetry data, enabling bi-directional, real-time synchronization between physical and digital entities. This persistent linkage allows the DT to reflect dynamic state changes in the physical system, while also supporting optimizations and control feedback. The approach leverages open standards and composable, modular simulation architectures to replace selected functionalities of the Computer Generated Forces (CGF) system with DT-derived behaviors. The evolving AI flight model, trained using data from the physical asset, was progressively integrated and assessed within the simulation to validate performance gains across training iterations. Comparative analyses showed significant improvement in mission execution, particularly in precision path following and search pattern adherence. The results underscore the feasibility and advantages of embedding AI-enhanced DTs into constructive simulation environments, offering enhanced realism, system adaptability, and analytical transparency. This work lays foundational insights for the broader application of DTs in simulation-based testing, optimization, and decision support for military applications.
Climate change and the growing emphasis on environmental protection present significant challenges to environmental security and operational safety in NATO missions. This paper presents a conceptual framework for the deployment of autonomous unmanned systems to support key environmental assessment processes. The research builds on a qualitative design combining analysis, synthesis, comparison, and analogical reasoning. The approach integrates UAV capabilities across the entire lifecycle of NATO-relevant procedures, including Environmental Baseline Studies (EBS), Environmental Impact Assessments (EIA), and Environmental Close-out Studies (ECS). Focus is placed on the fusion of advanced airborne sensing technologies with onboard data processing and analytics. Key use cases include pre-deployment baseline mapping, real-time hazard surveillance during operations, and post-mission environmental verification—each contributing to enhanced force protection, faster decision-making, and strategic environmental accountability. The discussion also addresses the emerging legal concept of ecocide and its implications for NATO’s environmental responsibilities. Finally, the paper outlines future directions in autonomous environmental response, including the potential deployment of UAVs for in situ mitigation via the targeted release of decontaminating or bioremediating agents. The study concludes by emphasizing both the benefits and the current limitations of UAV-based environmental monitoring while highlighting the need for standardized validation procedures within NATO- and UN-led operations.
Target reconnaissance and subsequent identification of enemy military equipment are key elements of modern reconnaissance and intelligence systems, especially in multi-domain operations. Optoelectronic and unmanned vehicles equipped with infrared sensors enable the detection and identification of objects based on their thermal radiation, which provides an advantage when operating in conditions of reduced visibility, adverse weather, or enemy camouflage. These technologies are of fundamental importance for artillery reconnaissance, where early target detection affects the effectiveness of fire support and minimizes the risk to friendly forces. This article focuses on the design, implementation, and verification of a computer vision algorithm for detecting temperature anomalies corresponding to military equipment in thermal images. The hypothesis is based on the assumption that a rule-based algorithm based on thermal imaging data analysis can reliably detect military equipment even without the use of advanced machine learning methods, while its modular structure allows for future expansion with adaptive and intelligent features to increase accuracy in changing battlefield conditions. The text describes the principles of thermal imaging systems, the influence of the environment on detection quality, and the use of computer vision methods, including image segmentation and detection architectures. It includes an analysis of the behavior of thermal unmasking traces of military equipment and other thermal objects in various conditions. Based on experimental results, the benefits and limitations of the proposed solution and the possibilities of its integration into reconnaissance and weapon systems are discussed. The conclusion offers directions for further research, in particular the implementation of deep neural networks, the expansion of training datasets, and the fusion of data from multiple sensors to increase detection reliability.
This study examines the influence of selected contemporary armed conflicts, namely the Second Nagorno-Karabakh War (2020), the Russian aggression against Ukraine (2022 to present), and the Israeli-Palestinian conflict (2023 to present), on the planned acquisition of unmanned aerial vehicles (UAVs) within the armed forces of Central European states. Employing a comparative case study methodology, the research analyses acquisition processes in six countries: Poland, the Czech Republic, Slovakia, Hungary, Austria, and Germany. The study identifies key operational characteristics of UAVs deployment in these conflicts and surveys the specific acquisition decisions made by the selected states since 2015.
For programming the behavior of complex agents such as military platforms which contain a bunch of sensors and actuators, adapting a multi-agent programming approach would increase effectiveness by providing a complete and user-friendly control over the platform, either real or virtual. Especially in an emerging era of nontechnical user programmable products, this approach is increasingly gaining attention. BPMN provides a powerful means to define collaborative missions and processes, and is currently incorporated in agent programming for multirobot applications. In the traditional sense, the common definition of an agent includes sensors, actuators, and a decision logic providing interaction with the environment. However, when complexity of platforms increases, with increasing number and complexity of organic sensors, actuators and command and control systems, encapsulating each platform as an agent may decrease its elasticity. This paper proposes an approach to incorporate BPMN for military autonomous platform visual programming. This approach allows the platform or its model to be programmed in a user-friendly and expressive way for both end-users and model designers. This is achieved by representing the platforms’ sub-systems with discrete agents, enabling multi-agent collaborative programming at the sub-platform level. In this context, we aim to extend the widely known use of BPMN in multi-robot programming to complex military autonomous platform task planning. We first make a comparative analysis of the related state-of-the-art and the proposed approach. Then, we present the potential advantages of our approach through autonomous platform task planning for selected maritime operation scenarios.
