Interview
July 30, 2025
Introducing KAI's next-gen simulation ecosystem powered by Unreal Engine
Korea Aerospace Industries (KAI) is a comprehensive aerospace enterprise that has led South Korea's aviation power by developing major aircraft and unmanned aerial vehicles. KAI has recently expanded into the space sector, with projects including the development of the Light Armed Helicopter (LAH) and the KF-21 next-generation fighter jet, as well as the assembly of satellites and projectiles.
Korea Aerospace Industries (KAI) is a comprehensive aerospace solutions enterprise that has been at the forefront of the aviation and defense industries for the past 40 years, independently designing and manufacturing a broad range of aerospace systems, including the KT-1 basic trainers, T-50 advanced jet trainers, Surion maneuver helicopters, and Songgolmae unmanned aerial vehicles.
In recent years, the simulation industry has been rapidly reshaping itself around state-of-the-art digital technologies including digital twins, AI, and VR/AR. KAI has adopted Unreal Engine to keep pace with these changes, driving digital innovation across the aviation industry in the process.
In this interview with Danny Lee, Director and Head of M&S Research Department at KAI, we take a deeper dive into the topics discussed in a session that was presented at Unreal Fest Seattle 2024, which examined KAI's technological innovations, focusing on how the enterprise is responding to the rapidly changing simulation industry, as well as its simulation integration strategy, real-world applications, and future vision regarding Unreal Engine.
In recent years, the simulation industry has been rapidly reshaping itself around state-of-the-art digital technologies including digital twins, AI, and VR/AR. KAI has adopted Unreal Engine to keep pace with these changes, driving digital innovation across the aviation industry in the process.
In this interview with Danny Lee, Director and Head of M&S Research Department at KAI, we take a deeper dive into the topics discussed in a session that was presented at Unreal Fest Seattle 2024, which examined KAI's technological innovations, focusing on how the enterprise is responding to the rapidly changing simulation industry, as well as its simulation integration strategy, real-world applications, and future vision regarding Unreal Engine.
How is the simulation industry changing, and how is KAI responding to those changes?
Danny Lee, Director & Head of M&S Research Department: Recently, the simulation industry has seen rapid development and undergone structural changes. Thanks to the introduction of cloud-based simulation, high-performance resources can be accessed anytime and anywhere. By combining digital twins, AI, and machine learning technologies, simulation is no longer a mere reproduction tool—it is evolving into a tool for prediction and optimization.
In addition, VR/AR/MR technologies make training more immersive and realistic, enabling simulations that resemble real-world environments. Software design based on microservices architecture significantly improves flexibility and scalability as well.
To proactively respond to this digital transformation, KAI is integrating traditional legacy simulation systems with Unreal Engine. We have three core strategies: first, accelerating technology verification and application through rapid prototyping with Unreal Engine. Second, realizing seamless integration with existing systems through standardized interfaces. Third, promoting a long-term ecosystem by designing a platform for sustainable content development. These strategies enable us to maximize the value of our existing assets while responding flexibly and efficiently to the rapidly changing technology landscape.
How is Unreal Engine impacting the ever-changing simulation industry?
Lee: Unreal Engine is playing a very important role in the evolution of the simulation industry. First off, its high-quality real-time 3D graphics allow us to implement realistic, immersive simulation environments that significantly improve the efficiency of training and testing. Also, Unreal Engine supports robust VR/AR/MR integration, enabling real-world and experience-based learning in a broad range of industries.
Unreal Engine's modular architecture and open ecosystem facilitate integration with existing legacy systems, providing the flexibility to quickly apply new technologies and features. In particular, it connects seamlessly to the latest technologies such as digital twins, AI, and machine learning, which maximizes the efficiency of designing, maintaining, and operating complex systems.
For companies like KAI, Unreal Engine is more than just a tool. It is a core technology for developing sustainable simulation content and building new simulation ecosystems.
Image courtesy of KAI
KAI has showcased a broad range of projects that demonstrate the integration of Unreal Engine into existing systems. Could you tell us a little bit about them?
Lee: KAI is actively adopting Unreal Engine in its aircraft training system to develop realistic and efficient simulators. A typical example would be the VR simulator, which helps pilots familiarize themselves with procedures and controls through a VR device before entering a full-flight simulator. Unreal Engine has been used to create a virtual cockpit that is identical to that of a real aircraft, allowing pilots to repeatedly practice from takeoffs and landings to emergency procedures and operating avionics without an instructor.
Although traditional simulators provide realistic aircraft-level controls and training effects, there has been a limit to their mass distribution due to high construction and operation costs and the need for dedicated facilities. We have introduced VR technology to overcome these challenges, and Unreal Engine serves as an excellent replacement for video generators, instrument panels, input/output devices, and more. It also enables us to achieve the effects of a large demonstration system with a single VR HMD, which would normally require multiple devices.
Image courtesy of KAI
In addition, we are combining our own dynamics model and avionics system with Unreal Engine's real-time rendering to provide a training environment that more closely imitates real piloting. We are also reproducing the 3D terrain of the Korean Peninsula based on ultra-precise mapping, such as the Geographic Information System (GIS) and Digital Elevation Model (DEM), to support pilots learning the terrain of the mission area.
Maintenance training is another field in which Unreal Engine is being utilized as a core platform. The FA-50 maintenance training simulator, unveiled at I/ITSEC 2024, was designed not only for practicing inspection and parts replacement in a VR environment, but also to allow users to create their own training programs.
This offers an alternative that overcomes the limitations of conventional document-based training, flat-screen computer-based training (CBT), and repetitive scenario-based practice. The Surion helicopter’s Virtual Flight Testing (VFT), showcased at the same event, provides an immersive training environment that reflects real-world aircraft performance and terrain data through the use of a digital twin and high-resolution visualization.
Maintenance training is another field in which Unreal Engine is being utilized as a core platform. The FA-50 maintenance training simulator, unveiled at I/ITSEC 2024, was designed not only for practicing inspection and parts replacement in a VR environment, but also to allow users to create their own training programs.
This offers an alternative that overcomes the limitations of conventional document-based training, flat-screen computer-based training (CBT), and repetitive scenario-based practice. The Surion helicopter’s Virtual Flight Testing (VFT), showcased at the same event, provides an immersive training environment that reflects real-world aircraft performance and terrain data through the use of a digital twin and high-resolution visualization.
When developing the aforementioned KAI simulations and systems, how has Unreal Engine helped the creation process, productivity, and final outcome?
Lee: KAI's simulation production pipeline has been significantly transformed with the introduction of Unreal Engine. Datasmith enables us to easily import 3D models from design tools such as CATIA, making it possible to quickly build virtual cockpits and airframe models based on actual designs, as well as greatly reduce production time without additional modeling. Furthermore, overall productivity has improved since our own flight dynamics engine and avionics simulation software have been connected to Unreal Engine in real time, seamlessly integrating our backend systems with the visualization frontend.
Unreal Engine's rendering, sound, and animation features were the core tools in the development of the VR simulator, in which pilots particularly rely on visual and auditory information to make situational judgments. Physically based rendering (PBR) enables the realistic reproduction of materials such as metal, glass, and instrument panels, while the particle system and material nodes provide the flexibility to adjust visual effects such as smoke and air distortion. Using MetaSounds, the sound reacts to engine RPM and environmental changes in real time, providing pilots with sensations that are similar to real flight.
In addition, we leveraged Animation Blueprints to create visually intuitive interlocking animations between the cockpit, instrument panel, and flight controls. Features such as Sky Atmosphere, Volumetric Clouds and Height Fog enhanced the immersion of the atmospheric representation and spatial perception training.
Image courtesy of KAI
In the case of terrain creation, Unreal Engine's Large World Coordinates (LWC) enabled us to maintain precision when moving at high speeds, even over thousands of kilometers of terrain. Having access to the full source code allowed us to implement the coordinate conversion, system integration, and precise terrain structuring for AI training systems.
In this process, real-world terrain data, aerial photographs, and altitude information were integrated into Unreal Engine. We used precise terrain information based on GIS and DEM to create realistic, challenging scenarios including complex flight paths, low flying training, and target navigation. As a result, KAI was able to successfully build a next-generation aircraft simulation platform that satisfies all the criteria: ultra-large terrain data, ultra-precise location-based training, and precise coordinate integration with external systems.
In addition, project extensibility and content creation flexibility have significantly improved through a wide range of plugins, hardware interfaces, form management tool integration, RealityScan (formerly known as RealityCapture), and assets from the Fab marketplace.
In this process, real-world terrain data, aerial photographs, and altitude information were integrated into Unreal Engine. We used precise terrain information based on GIS and DEM to create realistic, challenging scenarios including complex flight paths, low flying training, and target navigation. As a result, KAI was able to successfully build a next-generation aircraft simulation platform that satisfies all the criteria: ultra-large terrain data, ultra-precise location-based training, and precise coordinate integration with external systems.
In addition, project extensibility and content creation flexibility have significantly improved through a wide range of plugins, hardware interfaces, form management tool integration, RealityScan (formerly known as RealityCapture), and assets from the Fab marketplace.
Image courtesy of KAI
We heard KAI is also adopting AI agents into Unreal Engine for photorealistic, large-scale tactical training. Could you elaborate on that?
Lee: KAI is working on integrating reinforcement learning-based AI agents into real-world training scenarios for the development of a next-generation tactical training simulator. Particularly in complex battlefield environments where a broad range of weapon systems and platforms are operated simultaneously, there is an important need for the technology to organically integrate them in a single simulation space.
Although existing commercial simulator solutions have many constraints when it comes to external system integration and customization, Unreal Engine is able to overcome such restrictions with full source code access in C++. This openness has allowed KAI to integrate its own AI agents precisely to secure real benefits, even in tactical training scenarios that require complex interactions.
This integration goes beyond simply utilizing AI—it refers to a structure in which human pilots and AI are able to train and interact in the same simulation environment. Thanks to Unreal Engine, what would have been difficult to implement with traditional solutions is now made possible. Ultimately, Unreal Engine provides a platform that integrates AI, real-time simulation, and data feedback, playing an essential role in the implementation of KAI's next-generation tactical training system.
Could you tell us about the future direction of the simulation ecosystem and the vision of KAI?
Lee: In the future, the simulation ecosystem will develop around openness, sustainability, and personalization. Customized training systems based on AI and big data, high-performance simulations without geographical constraints in cloud environments, and immersive, real-time feedback systems utilizing VR/AR and wearable technologies are predicted to become the norm.
In the midst of such changes, KAI aims to build a technology-integrated platform and its own simulation ecosystem, laying the foundation for the sustainable growth of South Korea's simulation industry. We are actively utilizing Unreal Engine as a simulation engine—rather than just a development tool. We are developing a simulation content pipeline that enables the rapid production of high-quality content centered on the platform.
KAI's vision is to connect with the global simulation ecosystem beyond South Korea. With the open nature and technology of Unreal Engine as the basis, our goal is to create a simulation platform that can be shared across industries, which leads to a healthy and scalable ecosystem where a broad range of industries, organizations, and developers can work together.
Based on this vision and new direction, KAI aims to develop simulation technology beyond a mere training tool into a core infrastructure for improving product development, maintenance, and operational efficiency.