CASE STUDY
CloudRF Signal Propagation and Tower Optimisation
TechnoLynx helped CloudRF speed up signal propagation and tower placement simulations with GPU acceleration, custom algorithms, and cross-platform support, making RF planning faster and more scalable.
The Challenge
CloudRF needed faster, more accurate signal propagation and tower-placement simulations, dealing with large terrain datasets, real-time analysis, and inconsistent results across platforms.
Slow propagation simulations at scale
Signal propagation became too slow for large datasets, and real-time analysis was difficult with detailed maps and complex urban layouts.
Heavy terrain computation
Running detailed terrain-based calculations consumed significant time and compute resources.
Tower placement is computationally expensive
Identifying optimal transmitter/receiver positions on complex terrain slowed operations considerably.
Platform inconsistencies limited deployment
Running simulations across different client environments led to inconsistencies, restricting CloudRF’s ability to offer services universally.
Project Timeline
From performance bottlenecks to GPU acceleration, placement heatmaps, and cross-platform deployment
Propagation Engine Enhancements
Developed a multi-threaded engine using height‑maps and voxel calculations, managed by a custom work‑package scheduler. Added CUDA‑optimised Cartesian & polar algorithms, with basic ray‑casting as a slower validation fallback.
Added slower reference algorithms (basic ray-casting) for validation when accuracy checks were needed.
Validation & Accuracy Checks
Tower‑Placement Heatmaps
Created heatmaps showing optimal candidate zones rather than single points. Employed Monte Carlo simulations across thousands of configurations. Achieved ~300× speedups, mainly from proprietary heatmap logic.
Implemented indoor voxel-space ray casting powered by CUDA and integrated point‑cloud-to-voxel conversion for detailed indoor modelling.
Indoor Propagation
Deployment & Platform Consistency
Containerised with Docker, used CMake for cross-platform builds, and introduced unit and integration tests to ensure reproducible results across environments.
The Solution
We delivered a multi-layer solution focused on speed, scalability, and deployment flexibility: a faster propagation engine, smarter tower placement via heatmaps + Monte Carlo exploration, indoor voxel ray-casting, and a platform-independent deployment model.
Advanced Propagation Engine
High-performance multi-threaded propagation using height maps and voxel-based terrain handling, orchestrated by a custom scheduler and accelerated via CUDA (Cartesian and polar coordinate methods).
Tower Placement Optimisation
Generated coverage heatmaps to visualise best placement zones, then ran Monte Carlo testing across thousands of configurations to identify strong patterns for coverage.
Indoor Signal Modelling
CUDA-accelerated voxel-space ray casting for indoor environments, paired with point cloud to voxel conversion for more realistic building-level simulations.
Cross-Platform Delivery
Docker containerisation for consistent deployment, CMake builds for common operating systems, and unit/integration tests to verify stability and correctness across platforms.
Technical Specifications
Execution
Multi-threaded engine + custom work-package scheduler
Terrain representation
Height maps + voxel-based calculations
GPU acceleration
CUDA-optimised algorithms (Cartesian and polar coordinate methods)
Output
Heatmaps showing candidate placement zones (coverage-oriented)
Search method
Monte Carlo testing across thousands of configurations
Method
CUDA-accelerated voxel-space ray casting
Input pipeline
Point cloud → voxel conversion for detailed indoor maps
Containerisation
Docker deployment for consistent environments
Build system
CMake cross-platform builds
Testing
Unit + integration tests verifying correctness across platforms
The Outcome
Performance improved drastically, enabling faster and more precise simulations, simpler deployment, and broader platform reach.
200–300%
Faster propagation simulations
Gains attributed to optimised scheduling and CUDA enhancements.
10–20×
Faster large height-map processing
Accelerated with GPU support when working on large terrain datasets.
~300×
Faster tower-placement heatmap generation
Largest improvement driven by proprietary heatmap algorithms (not only multithreading).
Key Achievements
Built a high-performance propagation engine using height maps, voxel calculations, and a custom work-package scheduler.
Introduced tower placement heatmaps and Monte Carlo testing to evaluate thousands of configurations and coverage patterns.
Improved indoor propagation fidelity with CUDA-accelerated voxel ray casting and point cloud conversion.
Enabled consistent deployment using Docker, cross-platform builds via CMake, and unit/integration tests for correctness.
Platform independence allowed the solution to reach new market segments.
Need Faster RF Planning and Coverage Optimisation?
Let’s discuss how GPU acceleration, scalable simulation design, and cross-platform deployment can improve signal propagation modelling and tower placement workflows for your team.
