MAP- <i>X</i> : Massive Field Data Processing for Real-Time Wide-Area Mapping Using High-Altitude Platforms With MIMO

In this paper, we propose Massive Aerial Processing for <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">X</i> (MAP-<italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">X</i>), an innovative framework for reconstructing spatially correlated ground information, where <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">X</i> represents arbitrary types of geographically referenced sensing data. MAP-<italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">X</i> leverages distributed sensors and a high altitude platform (HAP) equipped with a planar antenna array that captures radio frequency signals transmitted simultaneously from ground sensors. MAP-<italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">X</i> incorporates two key techniques that provide significant advantages over conventional terrestrial wireless sensor network (WSN)-based methods. First, MAP-<italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">X</i> enables all sensors to transmit concurrently within a single subframe, resulting in the random superposition of signals at the receiver. This non-orthogonal transmission approach enhances estimation accuracy as the number of devices increases, addressing the limitations of restricted orthogonal channel availability. Second, MAP-<italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">X</i> introduces a novel waveform design that enhances angular resolution at the receiver. Combined with linear and machine learning (ML)-based post-processing techniques implemented at the HAP, this approach significantly improves the accuracy of field-data reconstruction. This paper details the signal model, waveform design, and efficient post-processing techniques underlying MAP-<italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">X</i>. Simulation results demonstrate that our framework surpasses the upper bound performance of orthogonal data collection combined with optimal covariance-based signal reconstruction, achieving superior latency reduction and estimation accuracy.