Why Konan AI Pilot is
your best choice?
Konan AI Pilot is an AI combat solution capable of
autonomously handling even high-risk missions.
It can collaborate with pilots and teams, learning and
understanding missions just like a human
Konan AI Pilot combines reinforcement learning–based AI with strategic decision-making algorithms,
enabling autonomous operation and mission execution in real combat scenarios.
Through real-time feedback loops and Behavior Tree–based decision-making structures,
it realizes an AI combat system that extends from copilots to fully autonomous combat pilots.
Video - Flying Toward a Better World,
Konan AI Pilot
FEATURE
-
1. Digital Twin Framework
- Real-to-virtual mapping technology applying physical simulation and aerodynamics
- 3D rendering technology suitable for realistic fighter jet visualization
- Utilization of diverse sensor data such as radar, infrared sensors, and cameras
- Models applicable to BVR(Beyond Visual Range) and WVR(Within Visual Range)combat scenarios
-
2. Hybrid Reinforcement Learning
- Mimicking Top Gun-style fighter jet maneuvers
- Reinforcement learning through self-play
- Parallel learning utilizing heterogeneous CPUs and GPUs
- Reinforcement learning capable of understanding Multi-Mission Operations
-
3. High-Fidelity Simulation
- Light Weight Simulation → Enables fast-speed simulations
- High Fidelity Simulation → Provides realistic experience for mission readiness and battlefield training
- Edge Cases Simulation → Testing for unexpected and extreme situations
- Technology for experiments and performance metrics to validate trained AI pilots
-
4. Voice Recognition Command & Control
ARCHITECTURE
EXPECTED IMPACT
By utilizing Konan AI Pilot, highly realistic mission rehearsals
and edge-case simulations can be conducted at low cost.
- Low-cost execution of highly realistic mission rehearsals
- Low-cost execution of edge case simulations
- Establishment of a metaverse training system that overcomes time and space
constraints - Development of flight training data-based transfer learning
- Significant reduction of pilot training time
VISION
Konan AI Pilot aims to ultimately establish a manned-unmanned
teaming (MUM-T) system based on Digital Twin technology.
- 2024. 1Q
- Single Agent AI Pilot Phase 1 – WVR development / Combat flight simulator advancement Phase 1 – EO sensor application
- 2024. 2Q
- Multi Agent AI Pilot Phase 1 – BVR (2 vs 2) development / Combat flight simulator advancement Phase 2 – Radar simulation application
- 2024. 4Q
- Single Agent AI Pilot Phase 2 – BVR development / Combat flight simulator advancement – IR sensor application
- 2025. 4Q
- Combat situation awareness system debriefing / Multi Agent AI Pilot / Combat flight simulator advancement Phase 3 – Lidar sensor application
- 2028.4Q
- Launch of manned-unmanned teaming (MUM-T) system / Execution of specific missions such as reconnaissance, surveillance, cooperative engagement, and ground strike operations based on MUM-T