AIM-SIM is a real-time, low-latency tactical shooting simulator designed for military and police training. It seamlessly merges a standard replica firearm, an integrated laser targeting system, and a Unity-based scenario engine to create a highly repeatable, data-driven combat environment.
Our core philosophy is simple: achieve military-standard accuracy using consumer-grade components. By utilizing an ESP32 microcontroller and standard optical sensors instead of expensive proprietary hardware, AIM-SIM minimizes maintenance costs and entry barriers without compromising precision.
1. Smart Weaponry (Hardware-in-the-Loop)
Networked Firearm: An ESP32 microcontroller transforms a standard physical replica into a smart device that operates as its own Wi-Fi Access Point (AIM-SIM-PRO).
Dynamic Firing Logic: The weapon autonomously manages complex firing modes (Single, Burst, Full Auto) and emits a mathematically optimized 40 ms laser pulse per shot.
Web Integration: Instructors can connect directly to the weapon via a browser-based interface to instantly adjust magazine capacities and firing rates.
2. Computer Vision & Low-Latency Telemetry
Sub-Pixel Precision: Instead of using basic bounding boxes, our custom Python/OpenCV pipeline calculates the exact center of the laser at a sub-pixel level using Image Moments.
Adaptive Tracking: Dynamic thresholding adapts to ambient light for flawless daytime and Night-Ops performance, while Kalman Filters predict laser trajectories to prevent tracking loss.
Zero-Latency: Real-time data including normalized coordinates and confidence scores is streamed to the simulation engine via the UDP protocol.
3. Spatial Mapping & TEVAL Validation
Homography Transformation: We use a 4-point interactive calibration system that solves an 8-unknown equation via Gaussian Elimination. This instantly converts the perspective-distorted camera feed into a perfect flat 2D plane.
Military-Standard Validation: Accuracy is strictly proven through our Test & Evaluation (TEVAL) protocol. The system calculates the Root Mean Square Error (RMSE) to dynamically approve or reject the calibration.
4. Unity-Based Tactical Engine
Combat Modules: The state-machine-driven 3D engine executes core training drills, including Dynamic Assault, Threat Discrimination, and Snap Shooting.
Raw Ballistics: Hit detection bypasses traditional gaming mechanics. We execute strict 3D ballistic raycasting over UDP coordinates with precise body and head hitboxes.
Immersive Realism: The engine dynamically mixes hostile and civilian targets to test situational awareness, while simulating operator breathing and tactical lighting.
5. After Action Review (AAR) & Analytics
Millisecond Precision: The AAR system records every event, tracking shots, misses, and reaction times down to the millisecond.
Strict Grading: Operators are evaluated on a military-standard ranking system (Rank S to F). Hitting a single civilian results in an immediate scenario failure.
Institutional Memory: All session data is exported via CSV formats, ready for seamless integration into institutional ERP systems or Power BI to track long-term personnel development.