Spectrum Occupancy Heatmap

A heatmap is not a static image; it is a visualization of a complex, multi-dimensional dataset. The core axes are frequency (x-axis), time (y-axis), and power/duty cycle (color gradient). Advanced implementations add a fourth dimension by tiling these plots across a geographic grid, creating a spatial map of spectral activity. This allows an operator to see not just what frequency is busy, but where and when it is occupied.

Heatmaps typically encode one of two key metrics:

• Duty Cycle: The percentage of time a frequency bin is occupied above a defined energy threshold. This is ideal for identifying persistent transmissions like broadcast carriers.
• Power Spectral Density (PSD): The average received signal power (dBm/Hz) within each bin. This reveals not just occupancy, but the relative strength of emitters, which is critical for interference hunting and locating transmitters via power-based geolocation techniques.

To build a spatial heatmap, the area of interest is partitioned into a discrete grid, often using systems like the H3 Hexagonal Grid for its uniform area and minimal shape distortion. Sensor measurements within each cell are aggregated to produce a single occupancy value. The color of the cell on the map then represents the aggregated spectrum metric, creating an intuitive overlay on a geographic information system (GIS) that highlights hot spots of electromagnetic activity.

The 'time' axis is defined by the integration window over which the spectrum analyzer collects data. A fast sweep (e.g., 100ms) captures transient, bursty signals like radar pulses, while a slow sweep (e.g., 1 minute) provides a smoothed average for long-term trend analysis. Many heatmaps use a configurable persistence function, where a pixel's color fades over time, allowing the human eye to visually correlate infrequent, short-duration signals against a background of constant transmissions.

A raw heatmap is useless without proper calibration. A noise floor must be established for each receiver, as it varies with hardware, temperature, and location. An occupancy threshold (e.g., noise floor + 6 dB) is then set to distinguish a legitimate signal from random thermal noise. Without this, the heatmap would be saturated with false positives. Advanced systems use adaptive thresholding that dynamically adjusts the detection floor based on the local noise environment to maintain a constant false alarm rate (CFAR).

A spectrum occupancy heatmap is the primary visualization layer of a Radio Environment Map (REM). While the REM is the underlying database containing raw sensor data, propagation models, and policy constraints, the heatmap is the rendered output that makes this data actionable. By overlaying a real-time occupancy heatmap with a Digital Elevation Model (DEM) and building footprints, an operator can visually correlate a sudden drop in signal power with a terrain obstruction, instantly diagnosing a hidden node problem.

A geostatistical interpolation method that estimates unknown RF signal values at unmeasured locations by computing a weighted average of known neighboring measurements. Unlike simple linear interpolation, Kriging weights are derived from a variogram model that quantifies spatial autocorrelation—how signal similarity decays with distance. This provides the best linear unbiased predictor for constructing smooth, statistically rigorous heatmap surfaces from sparse sensor grids.

A non-parametric Bayesian approach to spatial spectrum estimation that models the entire RF environment as a distribution over functions. Beyond predicting a mean power level at each coordinate, GPR outputs a full posterior variance, directly generating a confidence interval map alongside the occupancy heatmap. This uncertainty quantification is critical for risk-aware dynamic spectrum access decisions where false negatives cause harmful interference.

The mathematical engine that translates transmitter parameters into predicted received signal strength across a geographic grid. Models range from empirical Longley-Rice terrain-sensitive predictions to deterministic ray tracing that simulates individual multipath reflections from 3D city models. Accurate propagation modeling is the difference between a theoretical heatmap and one that reflects true physical coverage contours.

A high-fidelity virtual replica of the physical electromagnetic environment, continuously synchronized with real-time sensor data. An RF digital twin integrates occupancy heatmaps, propagation models, and 3D city geometry into a unified simulation platform where operators can test spectrum policies, predict interference scenarios, and optimize network configurations before deploying changes to the live environment.

A discrete global grid system developed by Uber that partitions the Earth's surface into hierarchical hexagonal cells. Hexagons eliminate the edge-effect distortions of square grids and provide uniform adjacency—every cell has six equidistant neighbors. H3 serves as the standardized spatial indexing layer for aggregating spectrum sensor data, querying occupancy states, and rendering distortion-minimizing heatmap visualizations at global scale.

A signal processing breakthrough that enables wideband spectrum reconstruction from sub-Nyquist samples by exploiting the inherent sparsity of spectrum occupancy—most frequencies are empty most of the time. Compressed sensing dramatically reduces the hardware cost and data throughput required for building real-time occupancy heatmaps across gigahertz-wide bands, making dense sensor networks economically viable.