UAV-aided Radio Map Construction Exploiting Environment Semantics

TL;DR

This paper introduces a novel method for constructing 6D radio maps using environment semantics and virtual obstacle maps, significantly reducing measurement needs and enhancing UAV relay communication capacity.

Contribution

It proposes a joint estimation approach for environment-aware radio maps using a least-squares formulation with partial quasiconvexity, improving accuracy and efficiency.

Findings

Reduces measurement requirements for accurate radio maps

Achieves over 50% capacity gain in UAV relay scenarios

Efficient algorithm based on partial quasiconvexity

Abstract

This paper constructs a full dimensional (6D) radio map to predict the channel gain between any transmitter location and any receiver location based on received signal strength (RSS) measurements between low-altitude aerial nodes and ground nodes. The main challenge is how to describe the signal strength attenuation due to the blockage from the environment. Conventional interpolation-type approaches fail to exploit the close relation between the radio map and the geometry of the environment. This paper proposes to construct radio maps by first estimating and constructing a multi-class 3D virtual obstacle map that describes the geometry of the environment with radio semantics. Mathematically, a least-squares problem is formulated to jointly estimate the virtual obstacle map and the propagation parameters. This problem is found to have a partial quasiconvexity that leads to the development of an efficient parameter estimation and radio map construction algorithm. Numerical experiments confirm that the proposed method substantially reduces the amount of measurement required for achieving the same radio map accuracy. It is also demonstrated that in a unmanned aerial vehicle (UAV)-aided relay communication scenario, a radio map assisted approach for UAV placement can achieve more than 50% capacity gain.