How TART works

TART reads raw satellite telemetry and returns plots, statistics and written findings. It does not guess what your data means: everything it says about a channel comes from the dictionary you supply, and everything it says about the orbit comes from the element set you supply.

Three inputs, in order. A telemetry file. A dictionary that names and bounds each parameter. An element set that places the satellite in space. The first is required; the other two unlock progressively more of the catalog.

Importing telemetry

Drop a CSV, TSV or SQLite export. TART reads every row, discards those it cannot parse, and tells you how many and why. Rows are never silently dropped.

Adding a second file merges it into the same dataset. Overlapping days are kept once; the combined total appears under the file list.

The dictionary

One row per parameter: a description, a unit, optional limits, and an optional mapping to a standard name. Descriptions make reports readable. Limits enable alarm overlays. Mappings are what let a generic analysis recognise your battery as a battery.

parameter,description,unit,min,max,standard_name BAT1_VOLTAGE,Battery 1 terminal voltage,mV,7000,8400,battery_voltage BAT1_TEMP,Battery 1 pack temperature,degC,-10,45,battery_temperature

Download the draft, fill it in a spreadsheet, upload it back. TART proposes mappings where a name is close to a standard one and shows its confidence; ties are never resolved for you.

Element sets

Give TART a NORAD ID or a TLE file and it computes the orbit, the ground track and the eclipse intervals for your window. Elements more than a few weeks from your telemetry drift; TART shows the age so you can judge it.

The catalog

Each analysis declares what it needs. If the requirement is met the row is ready. If the requirement is a mapping you have not made, the row is blocked and links to the dictionary. If your file simply does not carry the telemetry, the row is listed as not applicable and stays visible — the absence is information too.

Reading a result

Values are binned by day. A hatched column means no contact that day — not a zero. The MIN/MAX band shows the spread inside each bin; the line is the mean. Findings sit under the plot in plain language, and the provenance row at the foot records exactly which file, dictionary and spec version produced what you are looking at.

Every result exports as a PNG for a slide or a self-contained HTML file for a report.