Machine Learning for Rapid Propagation Assessment
Project Details
-
Market
-
Test
The Challenge
Setting up effective communications networks in complex urban environments poses significant challenges, particularly in disaster or conflict zones where rapid deployment is crucial.
Traditionally, coverage prediction relies on physics-based simulators, but their time-intensive nature and reliance on high-performance computer platforms hinders quick decision-making. In time-critical scenarios, the urgent need to establish reliable communications networks necessitates a rapid and efficient alternative.
The Approach
Our approach leverages the latest in Machine Learning (ML) technology for coverage prediction, enabling rapid transmitter placement optimisation in complex urban environments on standard computing hardware.
We developed a proof-of-concept ML model suitable for the task and trained it on numerous coverage maps generated using conventional radio propagation tools – a process that required several days of compute time. Once trained, the ML model could accurately predict coverage for previously unseen terrains in a fraction of a second, using only a standard PC or laptop.
The massive speed up this technique offers paves the way to optimisation of transmitter location by facilitating rapid exploration of transmitter locations that achieve desired coverage objectives. These optimised locations still only take a few seconds to compute, offering unprecedented speed and scalability in establishing effective and robust communications networks.
The Outcome
We have developed a cutting-edge coverage prediction ML model to optimise transmitter locations in complex urban environments, that can run on a conventional laptop or PC in seconds.
The swift optimisation makes this solution particularly suited to rapid deployments or situations which are subject to rapid change.
The video at the top of this page shows the propagation of a wave from a single transmitter through an urban terrain.
