Wireless Channel Prediction in Partially Observed Environments

TL;DR

This paper presents a novel method for predicting wireless channel characteristics in environments with partial visual observations, combining ray tracing and machine learning to interpolate between statistical and deterministic models.

Contribution

It introduces a heuristic algorithm that estimates RF channels from partial environment data, capturing uncertainty and enabling adaptive predictions.

Findings

The method interpolates between statistical and deterministic models based on environment observability.

It captures the uncertainty of predictions depending on the explored environment region.

Demonstrated effectiveness in simulated indoor robotic navigation scenarios.

Abstract

Site-specific radio frequency (RF) propagation prediction increasingly relies on models built from visual data such as cameras and LIDAR sensors. When operating in dynamic settings, the environment may only be partially observed. This paper introduces a method to extract statistical channel models, given partial observations of the surrounding environment. We propose a simple heuristic algorithm that performs ray tracing on the partial environment and then uses machine-learning trained predictors to estimate the channel and its uncertainty from features extracted from the partial ray tracing results. It is shown that the proposed method can interpolate between fully statistical models when no partial information is available and fully deterministic models when the environment is completely observed. The method can also capture the degree of uncertainty of the propagation predictions depending on the amount of region that has been explored. The methodology is demonstrated in a robotic navigation application simulated on a set of indoor maps with detailed models constructed using state-of-the-art navigation, simultaneous localization and mapping (SLAM), and computer vision methods.