Planning-grade RF documentation

Compare a small urban Cell plan. Explain the recommendation.

A.T.O.M is a local-first deterministic urban RF planning and experimentation workspace. Configure a small Cell plan, evaluate a baseline, compare a bounded azimuth recommendation, inspect its assumptions, and export the evidence. Measured real-world propagation accuracy is not yet established.

Run A.T.O.M locally Follow the first planning task

The current A.T.O.M focused workspace showing the tool rail, overlay drawer, and Ankara map — Current workspace: compact command bar, focused tools, persistent map, and contextual results.

A.T.O.M dark network workspace showing eight selected cells, propagation rays, and the map key — Current workspace: compact command bar, focused tools, persistent map, and contextual results.

Choose your starting point

Start with one comparison: configure, evaluate, compare, explain and preserve the evidence. Use the architecture guide afterward to inspect how the system produces it.

Installation and operation

Run your first baseline-to-recommendation comparison

Clone with Git LFS, start Docker Compose, configure two to eight Cells, evaluate and compare a bounded recommendation, then export evidence.

Open download guide → System design

Trace data from the browser to the RF engines

See component boundaries, request lifecycles, R-tree queries, worker limits, interference calculations, and 5G communication paths.

Open architecture guide →

What A.T.O.M evaluates

Each analysis is exposed as a distinct workflow so radio configuration, spatial evidence, and output interpretation remain connected without filling the map with permanent controls.

Sector propagation

Estimates 2.6/28 GHz propagation with the deterministic urban UMa LOS/NLOS subset and explicit legacy fallback. The 140 GHz profile is research/reference only.

Rays, Rx power, range, gaps

Demand optimization

Recommends evaluated azimuths using POI/residential demand and reach objectives. Normal networks accept two to eight Cells; bounded search does not prove a global optimum.

Before/after score and azimuths

Interference quality

Evaluates serving-cell selection and co-channel interference on an adaptive grid for modeled RSRP, SINR, RSRQ, and RSSI.

Quality surface and demand impact

5G communication path

Maps selected 5G cells to gNBs and distinguishes direct Xn coordination, AMF/N2 fallback, and UPF/N3 session traffic.

Routes, functions, sessions, events

Planning reports

Preserves the plan, assumptions, source Run and comparisons in separately downloaded reports. Portable projects contain bounded retained Scenario evidence, not every workspace artifact.

Print/PDF and Markdown

One planning loop

The application keeps the map stable while one focused tool is open. Configuration changes invalidate stale analysis, and only the latest matching request is allowed to update the workspace.

01 Set context Choose a project, mode, technology, transmit power, and cells.

02 Shape propagation Set rays, radius, azimuth, and sector beam width.

03 Run analysis Evaluate a sector, network plan, optimization, or interference surface.

04 Inspect spatially Select towers, gaps, paths, and radio-quality samples on the map.

05 Compare and export Save and compare scenarios, validate evidence, then export reproducible reports.

Know the confidence boundary

A.T.O.M supports comparisons under explicit assumptions. Normal network8, Recommendation5 and Measurement6 are independent limits. Optional terrain/path-profile evidence does not modify canonical fast-network propagation; building entry estimates one representative facade, not complete indoor coverage.

Local computation, explicit inputs

Validated dataset packs are loaded into memory. Projects preserve exact requests, model versions, and dataset identity for repeatable comparisons.

Planning estimate, not a measurement

Measured real-world propagation accuracy is not yet established. Fast fading, canonical multipath, MIMO scheduling, measured/vendor antenna diagrams and uplink behavior remain outside the model. Standards-based subsets and operational qualification do not establish field accuracy.

Reference library

Use the focused guides above for orientation, then move into the detailed API, algorithms, deployment, and release history.

REST API referenceEndpoints and contracts Algorithms and RF physicsEquations and geometry Capability detailsSupported workflows Reading propagation mapsVisual interpretation Production deploymentOperations guide Dataset Pack StudioBuild arbitrary-region packs Modeling limitsConfidence boundary Concept 4E building entryFacade-entry model reference Concept 4F.1 height-aware obstructionCenterline roof classifier Concept 4G.1 antenna and link budgetExplicit RF term contract Concept 4I.1 140 GHz auditApplicability and evidence boundary Concept 4I.2B P.1411 candidatesStrict Table 4 reference gates Concept 4I.3 measurement validationEvidence, calibration, and spatial holdouts Concept 7C durable run historyLocal execution lineage Concept 7D report artifactsLocal evidence and delivery Concept 7E Scenario UXVersions, branches, and lineage Concept 7H RF configurationPlanning, Advanced, and Research / reference FAQ and troubleshootingCommon questions Project changelogReleases and updates Contributor guideDevelopment workflow