pg_orrery/README.md
Ryan Malloy 15a830dc40 Initial implementation of pg_orbit PostgreSQL extension
6 custom types (tle, eci_position, geodetic, topocentric, observer,
pass_event), 67 SQL functions, 2 operators (&&, <->), and a GiST
operator class for altitude-band indexing. Wraps Bill Gray's sat_code
for SGP4/SDP4 propagation with WGS-72 constants for propagation and
WGS-84 for coordinate output. All 5 regression tests pass on PG 18.
2026-02-15 17:07:07 -07:00

306 lines
12 KiB
Markdown

# pg_orbit
Orbital mechanics types and functions for PostgreSQL.
pg_orbit adds native SQL types for TLEs, orbital positions, ground stations, and
satellite passes. It wraps Bill Gray's [sat_code](https://github.com/Bill-Gray/sat_code)
(MIT) for SGP4/SDP4 propagation, provides coordinate transforms between inertial
and ground-referenced frames, predicts passes over observer locations, and supports
GiST-indexed conjunction screening on altitude bands.
Think PostGIS, but for objects in orbit.
## Installation
Requirements:
- PostgreSQL 14+ development headers (`pg_config` in PATH)
- GCC and Make
- C++ compiler (for sat_code)
```bash
git clone --recurse-submodules https://github.com/...
cd pg_orbit
make
sudo make install
```
Then in your database:
```sql
CREATE EXTENSION pg_orbit;
```
If you cloned without `--recurse-submodules`, initialize the sat_code dependency:
```bash
git submodule update --init
```
## Quick Start
```sql
-- Create a table with a TLE column
CREATE TABLE satellites (
norad_id int PRIMARY KEY,
name text NOT NULL,
tle tle NOT NULL
);
-- Insert a TLE (standard two-line format, concatenated with newline)
INSERT INTO satellites VALUES (25544, 'ISS',
'1 25544U 98067A 24001.50000000 .00016717 00000-0 10270-3 0 9002
2 25544 51.6400 208.5000 0006000 30.0000 330.0000 15.50000000400000');
-- Propagate to a point in time
SELECT sgp4_propagate(tle, now()) FROM satellites WHERE norad_id = 25544;
-- Subsatellite point (nadir)
SELECT subsatellite_point(tle, now()) FROM satellites WHERE norad_id = 25544;
-- All passes over Boulder, CO in the next 24 hours (min 10 deg elevation)
SELECT sat.name, p.*
FROM satellites sat,
LATERAL predict_passes(sat.tle, '40.0N 105.3W 1655m'::observer,
now(), now() + '24h', 10.0) p
ORDER BY pass_aos_time(p);
-- Conjunction screening with GiST index
CREATE INDEX ON satellites USING gist (tle);
SELECT a.name, b.name, tle_distance(a.tle, b.tle, now())
FROM satellites a, satellites b
WHERE a.tle && b.tle AND a.norad_id < b.norad_id
AND tle_distance(a.tle, b.tle, now()) < 10.0;
```
## Types
| Type | Size | Description |
|------|------|-------------|
| `tle` | 112 bytes | Parsed mean orbital elements (epoch, Keplerian elements, drag terms, identifiers). Stored as a fixed-size struct, not raw text. |
| `eci_position` | 48 bytes | Position (km) and velocity (km/s) in the True Equator Mean Equinox (TEME) frame. |
| `geodetic` | 24 bytes | Latitude/longitude (degrees) and altitude (km) on the WGS-84 ellipsoid. |
| `topocentric` | 32 bytes | Azimuth, elevation (degrees), slant range (km), and range rate (km/s) relative to an observer. |
| `observer` | 24 bytes | Ground station location. Accepts human-readable input: `'40.0N 105.3W 1655m'` or decimal degrees. |
| `pass_event` | 48 bytes | Satellite pass with AOS/MAX/LOS times, max elevation, and AOS/LOS azimuths. |
### Input Formats
**tle** -- Standard two-line format (lines joined by newline):
```sql
SELECT '1 25544U 98067A 24001.50000000 .00016717 00000-0 10270-3 0 9002
2 25544 51.6400 208.5000 0006000 30.0000 330.0000 15.50000000400000'::tle;
```
**observer** -- Flexible ground station input:
```sql
-- Compass notation with altitude
SELECT '40.0N 105.3W 1655m'::observer;
-- Decimal degrees (positive East, altitude in meters)
SELECT '40.0 -105.3 1655'::observer;
```
## Functions
### TLE Accessors
| Function | Returns | Description |
|----------|---------|-------------|
| `tle_norad_id(tle)` | `int4` | NORAD catalog number |
| `tle_epoch(tle)` | `float8` | Epoch as Julian date (UTC) |
| `tle_inclination(tle)` | `float8` | Inclination in degrees |
| `tle_eccentricity(tle)` | `float8` | Eccentricity (dimensionless) |
| `tle_raan(tle)` | `float8` | Right ascension of ascending node (degrees) |
| `tle_arg_perigee(tle)` | `float8` | Argument of perigee (degrees) |
| `tle_mean_anomaly(tle)` | `float8` | Mean anomaly (degrees) |
| `tle_mean_motion(tle)` | `float8` | Mean motion (rev/day) |
| `tle_bstar(tle)` | `float8` | B* drag coefficient (1/earth-radii) |
| `tle_period(tle)` | `float8` | Orbital period (minutes) |
| `tle_perigee(tle)` | `float8` | Perigee altitude (km above WGS-72 ellipsoid) |
| `tle_apogee(tle)` | `float8` | Apogee altitude (km above WGS-72 ellipsoid) |
| `tle_age(tle, timestamptz)` | `float8` | TLE age in days (positive = past epoch) |
| `tle_intl_desig(tle)` | `text` | International designator (COSPAR ID) |
### Propagation
| Function | Returns | Description |
|----------|---------|-------------|
| `sgp4_propagate(tle, timestamptz)` | `eci_position` | Propagate to a point in time. Uses SGP4 for near-earth, SDP4 for deep-space. |
| `sgp4_propagate_series(tle, start, stop, step)` | `SETOF (t, x, y, z, vx, vy, vz)` | Time series of TEME positions at regular intervals. |
| `tle_distance(tle, tle, timestamptz)` | `float8` | Euclidean distance (km) between two objects at a reference time. |
### Coordinate Transforms
| Function | Returns | Description |
|----------|---------|-------------|
| `eci_to_geodetic(eci_position, timestamptz)` | `geodetic` | TEME to WGS-84 geodetic (lat/lon/alt). Requires time for Earth rotation. |
| `eci_to_topocentric(eci_position, observer, timestamptz)` | `topocentric` | TEME to observer-relative az/el/range. |
| `subsatellite_point(tle, timestamptz)` | `geodetic` | Nadir point on WGS-84 ellipsoid. Propagates internally. |
| `ground_track(tle, start, stop, step)` | `SETOF (t, lat, lon, alt)` | Ground track as time series of subsatellite points. |
### ECI Accessors
| Function | Returns | Description |
|----------|---------|-------------|
| `eci_x(eci_position)` | `float8` | X position (km, TEME) |
| `eci_y(eci_position)` | `float8` | Y position (km, TEME) |
| `eci_z(eci_position)` | `float8` | Z position (km, TEME) |
| `eci_vx(eci_position)` | `float8` | X velocity (km/s) |
| `eci_vy(eci_position)` | `float8` | Y velocity (km/s) |
| `eci_vz(eci_position)` | `float8` | Z velocity (km/s) |
| `eci_speed(eci_position)` | `float8` | Velocity magnitude (km/s) |
| `eci_altitude(eci_position)` | `float8` | Geocentric altitude (km) |
### Topocentric Accessors
| Function | Returns | Description |
|----------|---------|-------------|
| `topo_azimuth(topocentric)` | `float8` | Azimuth in degrees (0=N, 90=E, 180=S, 270=W) |
| `topo_elevation(topocentric)` | `float8` | Elevation in degrees (0=horizon, 90=zenith) |
| `topo_range(topocentric)` | `float8` | Slant range (km) |
| `topo_range_rate(topocentric)` | `float8` | Range rate (km/s, positive = receding) |
### Pass Prediction
| Function | Returns | Description |
|----------|---------|-------------|
| `next_pass(tle, observer, timestamptz)` | `pass_event` | Next pass over observer. Searches up to 7 days. |
| `predict_passes(tle, observer, start, stop [, min_el])` | `SETOF pass_event` | All passes in a time window. Optional minimum elevation (degrees). |
| `pass_visible(tle, observer, start, stop)` | `boolean` | True if any pass occurs in the time window. |
### Pass Event Accessors
| Function | Returns | Description |
|----------|---------|-------------|
| `pass_aos_time(pass_event)` | `timestamptz` | Acquisition of signal time |
| `pass_max_el_time(pass_event)` | `timestamptz` | Maximum elevation time |
| `pass_los_time(pass_event)` | `timestamptz` | Loss of signal time |
| `pass_max_elevation(pass_event)` | `float8` | Maximum elevation (degrees) |
| `pass_aos_azimuth(pass_event)` | `float8` | AOS azimuth (degrees, 0=N) |
| `pass_los_azimuth(pass_event)` | `float8` | LOS azimuth (degrees, 0=N) |
| `pass_duration(pass_event)` | `interval` | Pass duration (LOS - AOS) |
### Operators
| Operator | Operands | Description |
|----------|----------|-------------|
| `&&` | `tle, tle` | Altitude band overlap. Necessary (not sufficient) condition for conjunction. |
| `<->` | `tle, tle` | Minimum altitude-band separation in km. Returns 0 if bands overlap. |
Both operators are supported by the GiST `tle_ops` operator class for indexed scans.
## GiST Indexing
The `tle_ops` operator class indexes TLEs by their altitude band (perigee to apogee).
This provides fast filtering for conjunction screening: only pairs whose altitude
bands overlap can possibly be close to each other.
```sql
-- Create the index
CREATE INDEX idx_tle_alt ON satellites USING gist (tle);
-- The && operator triggers index scans
EXPLAIN SELECT a.name, b.name
FROM satellites a, satellites b
WHERE a.tle && b.tle AND a.norad_id < b.norad_id;
-- KNN ordering by altitude-band distance
SELECT name, tle <-> (SELECT tle FROM satellites WHERE norad_id = 25544) AS sep
FROM satellites
ORDER BY tle <-> (SELECT tle FROM satellites WHERE norad_id = 25544)
LIMIT 20;
```
The index reduces conjunction candidate pairs from O(n^2) to the set of objects with
intersecting altitude bands, which is then refined by computing actual `tle_distance()`
at a specific time.
## Geodetic Constants
TLEs are mean elements fitted using WGS-72 constants. Using WGS-84 constants for
propagation introduces kilometer-scale position errors because the elements absorb
geodetic model biases during the fitting process.
pg_orbit enforces this:
- **Propagation (SGP4/SDP4):** WGS-72 constants only (mu, ae, J2, J3, J4, ke)
- **Coordinate output (geodetic, topocentric):** WGS-84 (a=6378.137 km, f=1/298.257223563)
These are not interchangeable. Mixing them is a silent accuracy loss.
## Building from Source
```bash
# Build (requires pg_config in PATH)
make
# Install to PostgreSQL extension directory
sudo make install
# Run regression tests against a live database
make installcheck
```
Override `pg_config` location if needed:
```bash
make PG_CONFIG=/usr/lib/postgresql/16/bin/pg_config
sudo make PG_CONFIG=/usr/lib/postgresql/16/bin/pg_config install
```
### Project Layout
```
pg_orbit.control Extension metadata (version 0.1.0)
Makefile PGXS build
sql/
pg_orbit--0.1.0.sql Type, function, operator, and GiST definitions
src/
pg_orbit.c PG_MODULE_MAGIC entry point
tle_type.c TLE input/output/binary/accessors
eci_type.c ECI position type
observer_type.c Observer type with flexible parsing
sgp4_funcs.c SGP4 propagation and distance
coord_funcs.c Coordinate transforms (TEME/geodetic/topocentric)
pass_funcs.c Pass prediction (next_pass, predict_passes)
gist_tle.c GiST operator class for altitude-band indexing
types.h Shared struct definitions and constants
lib/
sat_code/ Bill Gray's SGP4/SDP4 (MIT, git submodule)
test/
sql/ Regression test SQL
expected/ Expected output
data/
vallado_518.csv 518 verification test vectors (Vallado et al.)
```
## Testing
pg_orbit uses the standard PostgreSQL regression test framework.
```bash
make installcheck
```
Test categories:
| Suite | Coverage |
|-------|----------|
| `tle_parse` | TLE input/output round-trip, malformed input rejection |
| `sgp4_propagate` | Vallado 518 test vectors, deep-space and high-eccentricity edge cases |
| `coord_transforms` | TEME to geodetic, TEME to topocentric accuracy |
| `pass_prediction` | Known ISS passes, polar and retrograde orbits |
| `gist_index` | Index scan vs. sequential scan result equivalence |
The Vallado 518 test vectors are the standard SGP4 verification dataset. Each row
contains a NORAD ID, minutes since epoch, and expected position/velocity. All 518
must pass to machine epsilon.
## License
[PostgreSQL License](LICENSE). Copyright (c) 2025, Ryan Malloy.
The bundled sat_code library is separately licensed under the MIT license.