Live · RTK GNSS tracking

Every horse. Every meter.
A hundred times a second.

Navionix tracks every runner at 100Hz with RTK precision — live classification, split times, and replay, from the starting gate to the photo finish.

See it live ↓ How it works

0 Hz

Position fixes per second, per runner

cm

RTK-grade precision, fix quality 4/5

<1 s

Glass-to-glass latency, device to viewer

0+

Parallel ingest channels — a full field at once

01 · Live tracking

The whole field, live.

Every runner positioned in real time from RTK GNSS receivers. Positions, gaps in time and distance, and instantaneous speed update as the race unfolds — and multi-lap distance math stays correct even when the finish line isn’t the start line.

— 0:00.00
Live classification gap · km/h

    Live demonstration leaderboard from an anonymized real race: runners ranked by distance covered, with time gaps to the leader and instantaneous speeds, updating continuously.

    02 · Race intelligence

    The race knows itself.

    Automatic race detection

    Start-line trigger, live classification, finish detection. The system knows the race state without human input.

    Splits & sectionals

    Exact crossing timestamps at every timing line, per runner — computed with PostGIS line-crossing geometry.

    Live classification & gaps

    Positions, time and distance gaps, instantaneous and average speeds — updated continuously, race-long.

    03 · Replay

    Scrub the entire race.

    Full replay with smooth interpolation — any race, any speed, every runner, every meter. Compact server-side trajectory sampling loads replays 10–20× faster. Then go deeper: speed curves, sectional comparisons, start-offset analysis, ground covered.

    Race replay · real race · anonymized 0:00.00
    Speed curves · m/s

    04 · Telemetry

    Every stride, quantified.

    After the line, the same 100Hz stream becomes a per-runner report — start, cruising regime, finish kick, peak, path efficiency, regularity — computed automatically from the raw fixes.

    Telemetry

    GPS data · 100Hz capture · real race · anonymized

    • Best start · 4
    • Peak speed · 6
    • Best finish kick · 1
    • Tightest line · 7
    • Most regular · 6
    › How to read these metrics
    Delay
    Gap between the runner’s own start-line crossing and the official start. ≈0 for front starters.
    Start
    Mean speed over the first 15% of distance — how fast the runner gets up to pace.
    Regime
    Cruising speed: median over the central 15–85% of the trip, robust to start and finish.
    Finish
    Mean speed over the final 15% compared with regime. ▲ +% = kick to the line, ▼ −% = fading.
    Peak
    Maximum smoothed speed, with the distance and phase where it happened.
    Path
    Actual GPS path ÷ official distance (>1 = extra ground covered), plus average / max deviation from the racing line.
    Regularity
    Coefficient of variation of speed in the heart of the race (middle 50%). Lower = more metronomic.
    Drops
    Sustained speed drops (≥0.5s) below 90% of regime — flags pull-ups or lost ground.
    # Runner Delay Start Regime Finish Peak Path Regularity Drops
    1 4 Runner 4 +0.5s 50.5 km/h 48.9 km/h 47.4 km/h ▼ −3.1% vs regime 52.8 km/h @1015m · mid ×1.062 line 2.3/5.2m 3.5% 0
    2 1 Runner 1 +0.5s 49.2 km/h 49.6 km/h 48.5 km/h ▼ −2.3% vs regime 52.3 km/h @1170m · late ×1.064 line 3.2/5.9m 3.4% 0
    3 3 Runner 3 0.0s 49.0 km/h 48.7 km/h 47.5 km/h ▼ −2.4% vs regime 52.2 km/h @32m · early ×1.044 line 2.3/6.3m 2.4% 0
    4 6 Runner 6 0.0s 50.3 km/h 48.6 km/h 47.0 km/h ▼ −3.3% vs regime 53.6 km/h @44m · early ×1.041 line 2.2/7.5m 1.8% 0
    5 2 Runner 2 0.0s 50.1 km/h 47.6 km/h 43.1 km/h ▼ −9.5% vs regime 52.7 km/h @90m · early ×1.026 line 1.3/5.2m 2.2% 0
    6 8 Runner 8 +0.7s 48.1 km/h 47.2 km/h 44.9 km/h ▼ −4.9% vs regime 49.8 km/h @1081m · late ×1.017 line 1.9/5.9m 2.5% 0
    7 7 Runner 7 +1.1s 47.8 km/h 46.7 km/h 43.8 km/h ▼ −6.2% vs regime 50.3 km/h @1053m · mid ×1.009 line 1.9/6.4m 3.6% 0

    05 · Architecture

    Built for the wire.

    A Rust backend — memory-safe, predictable latency, no GC pauses under load. Multi-sentence NMEA fusion combines position, speed, heading, and per-fix accuracy into one coherent data point. Every fix lands in PostgreSQL + PostGIS: auditable, replayable. Graceful degradation and backpressure under burst load.

    1. 01

      UDP ingest

      20+ parallel channels

    2. 02

      NMEA parser

      GGA + VTG + GST fusion

    3. 03

      Race engine

      PostGIS line crossings

    4. 04

      WebSocket push

      unlimited viewers

    device → server → live viewer · <1s end-to-end

    06 · Integrations

    Plugs into your broadcast.

    Clean REST + WebSocket API. Single-binary deployment — backend and frontend ship as one artifact.

    • Broadcast graphics
    • Odds & weather feeds
    • Timing systems
    • Race-master console

    Bring Navionix to your racecourse.

    Tell us about your track and your timing setup — we’ll show you a live demo.

    Get in touch →

    navionix@ippodromoghirlandina.it