• LinkedIn
• YouTube

MODULE_T Real-Time Target Simulation in Any Environment – Always Precise, Always Dynamic

Discover MODULE_T – the autonomous target drone system for training, simulation, and tactical operations in open environments. It enables fully autonomous navigation without GPS, complex manoeuvres, swarm flights, and live integration into training and command centre systems. At the core of the innovation is a two-stage visual navigation pipeline: Priority 1 uses extremely fast landmark-based localisation (<10 ms), Priority 2 falls back on robust image-to-image matching (VPR) when needed.

Seamless Integration into Existing Systems

MODULE_T integrates seamlessly into existing training and simulation systems (Tactical Training Systems, Command & Control Software). RGB and thermal video streams are transmitted via RTSP and automatically detected via ONVIF. Sensor data (e.g. position, velocity, or swarm status values) flows in via MQTT and OPC UA – ideal for real-time alarms, pop-ups, and dynamic scenario adjustments. The drone is fully IP-based and compatible with common platforms such as Milestone, Genetec, and other command centres – without any additional proprietary solutions.

Precise Navigation without GPS

Precise Navigation without GPS in Open Environments
MODULE_T achieves reproducible flight paths with minimal deviation – even in large open areas under GPS interference (jamming, spoofing, or signal loss). This is made possible by a two-stage visual navigation pipeline:

Priority 1 (Landmark-based, 80–90% of cases): Extraction of ORB features in a bearing-guided ROI (40–60% of image area), table-based lookup (<1 ms), and PnP pose estimation. Priority 2 (Fallback, only when landmarks are absent): Global image matching via FAISS (IVF-PQ + HNSW) with AnyLoc descriptors, optionally refined by SuperPoint + LightGlue.

A Decision Margin (difference between the best and second-best match) adaptively determines which path becomes active. The offline-prepared ground database (approx. 700 MB) contains lookup tables, a FAISS index, and visibility filters, and resides entirely onboard (e.g. Jetson Orin). Waypoints and complex manoeuvres are planned in a user-defined 3D grid and executed via ArduPilot/Pixhawk 6X. Manual takeover is possible at any time.

Suitable Application Areas

MODULE_T is excellently suited for daily training and simulation missions in open environments, military training grounds, and tactical scenarios. Typical applications include:

• Complex manoeuvre simulation (evasive manoeuvres, high-speed targets)
• Swarm flights for realistic threat replication
• Target acquisition and tracking training for air defence systems
• GPS-denied operations in contested environments (jamming training)

Autonomous target drone systems like MODULE_T dramatically increase training efficiency by enabling realistic, repeatable scenarios without personnel requirements and without GPS dependency – a trend confirmed in current studies on UAV swarm navigation. The system is immediately ready for deployment in modern training environments.

Scientific Foundation and Comparison

The navigation technologies of MODULE_T are based on established methods of visual robotics and modern VPR approaches:

• Two-stage visual navigation (landmarks + image matching) for minimal latency with maximum robustness
• Bearing-guided ROI Selection to reduce computational load by 35–55%
• Lookup tables & FAISS indices (IVF-PQ + HNSW) for O(1) access to 10 million reference images
• AnyLoc / DINOv2-based global descriptors for perspective-invariant matching

Further reading:

• Chang, Y. et al. (2023). A review of UAV autonomous navigation in GPS-denied environments. Robotics and Autonomous Systems.
• Power, W. et al. (2020). Autonomous Navigation for Drone Swarms in GPS-Denied Environments Using Structured Learning. Sensors.
• Arandjelovic et al. (2016). NetVLAD: CNN architecture for weakly supervised place recognition. CVPR.
• Johnson et al. (2019). FAISS: A library for efficient similarity search. Facebook AI Research.
• Sarlin et al. (2020). SuperPoint & LightGlue: Keypoint detection and matching. CVPR.
• Arshid, K. et al. (2025). Toward Autonomous UAV Swarm Navigation: A Review of Trajectory Design Paradigms. Sensors.
• Qamar, S. et al. (2022). Autonomous Drone Swarm Navigation and Multi-target Tracking in 3D Environments. arXiv preprint.

MODULE_T Compared to Market Leaders

The table is based on publicly available specifications (as of 2026), manufacturer data, and industry comparisons.

MODULE_T surpasses traditional systems particularly in GPS-denied scenarios and genuine swarm intelligence – ideal for modern, realistic training requirements.

Would you like to experience MODULE_T live or adapt it for your training centre? Contact us – we turn any drone into an intelligent, autonomous target.

Request Documents

Error: Contact form not found.

Last updated: March 2026