This package contains the high-fidelity, physics-based radar simulation engine isolated from the main CARLA ADAS pipeline. It uses CARLA Semantic LiDAR as the primary input and synthesizes raw ADC-like radar signals.
🚀 Iteration 18: The "Knobs and Dials" Milestone
The engine has been restructured for modularity. You can now tune individual radar profiles without touching the core physics logic.
🏗️ Package Structure
main.py: Entry point for standalone execution.
ConfigureRadar.py: The Control Panel. Define hardware profiles (BW, Chirps, LUTs here).
lidar.py: Handles point cloud normalization and axis-swapping (CARLA -> Shenron).
shenron/heatmap_gen_fast.py: GPU-accelerated signal synthesis.
🛠️ Usage
Configure Hardware:
Edit ConfigureRadar.py to select/tune your radar profile (e.g., awrl1432).
Set Paths:
Edit carla_shenron_config.yaml to point to your LiDAR dataset and output directories.
Run Simulation:
# Windows (Conda environment required)pythonmain.py
📡 Physics Consistency
This package strictly enforces the 1/R⁴ Radar Range Equation. All legacy scaling multipliers have been removed to ensure the simulation matches real-world hardware intensity profiles.
Fox CARLA ADAS Simulation | Shenron Physics Engine | 2026-04-15