Topside control station for the underwater robot (ROV) built by Triton Robotics, a student team from Seattle that competes in the MATE ROV Competition (Ranger class). This software ran the team's 2026 vehicle, which placed 2nd overall at the 2026 MATE ROV World Championship.
The vehicle's software is split across three repositories:
| Repository | Runs on | Role |
|---|---|---|
| TritonOS | Raspberry Pi on the ROV | Control loop, thruster output, sensor telemetry, camera streams |
| TritonPilot (this repo) | Pilot computer | Controller input, live video and telemetry, logging |
| TritonAnalysis | Separate analysis laptop | Stereo measurement, 3D reconstruction, mission tools |
Written and maintained by Thomas Gust, Triton's software and electrical lead, with contributions from @mr-meepstu.
TritonPilot is the topside operator application for Triton's ROV. It runs on the pilot computer, reads the game controller, publishes pilot commands to the ROV, receives live telemetry, starts and stops onboard camera streams, displays GStreamer video panes, records stream/sensor logs, and exposes operator-facing management tools.
Mission-specific analysis applets are intentionally not part of this
repository. During competition, detection, measurement, and scoring workflows
run from the sibling TritonAnalysis repository on a separate analysis
computer. TritonPilot should stay focused on live vehicle operation and data
logging.
main_topside.py starts the PyQt6 GUI and wires together the main topside
services:
- Controller input through
pygame - Pilot command publishing to TritonOS on
tcp://<ROV_HOST>:6000 - Sensor telemetry subscription from TritonOS on
tcp://<ROV_HOST>:6001 - Video stream control through the TritonOS video RPC endpoint on port
5555 - Management/calibration tools through the TritonOS management RPC endpoint on
port
5556 - Local UDP video receive pipelines using GStreamer
- Stream-log and raw-sensor CSV recording
The normal runtime relationship is:
Xbox-style controller
|
v
TritonPilot PilotPublisherService
|
v
PilotFrame JSON over ZeroMQ
|
v
TritonOS onboard control loop
Telemetry, management RPC, and video streaming are separate paths so problems can be diagnosed independently.
main_topside.py Main GUI entry point
config.py Runtime endpoints, controller tuning, and UI defaults
data/streams.json Camera stream definitions expected by the pilot UI
gui/ PyQt windows, panels, video tabs, instruments, raw sensors
input/ Controller discovery and PilotFrame publishing
telemetry/ Sensor subscriber and topside attitude estimator
video/ GStreamer receive path, camera manager, frame correction
recording/ JSONL stream logs, raw CSV telemetry, save locations
stereo/ Stereo calibration and disparity helpers
network/ Management RPC, local network selection, ZMQ helpers
schema/ Shared pilot-control wire schema
tools/ Controller, telemetry, video, and tether diagnostics
tests/ Hardware-free unit and GUI behavior tests
docs/ Maintained setup, operations, and architecture docs
- Documentation Index
- Setup Guide
- Desktop App Build
- Network Guide
- Analysis Transfer Link
- Operations Guide
- Architecture Overview
- Subsystem Reference
- Configuration Guide
- Testing And Troubleshooting
On Windows, from the repository root:
powershell -ExecutionPolicy Bypass -File .\setup_windows.ps1
.\.venv\Scripts\activate
python .\main_topside.pyFor a manual Python-only development environment:
python -m venv .venv
.\.venv\Scripts\activate
python -m pip install --upgrade pip
python -m pip install -r requirements.txt
python -m pip install pytest
python -m pytestVideo reception needs the external gst-launch-1.0 executable. The Windows
setup script can install and validate GStreamer; see the
Setup Guide for details.
The normal tethered layout is:
- Pilot computer tether adapter:
192.168.1.1 - ROV Ethernet:
192.168.1.4 - Pilot commands: ROV port
6000 - Sensor telemetry: ROV port
6001 - Video RPC: ROV port
5555 - Management RPC: ROV port
5556 - UDP video receive ports on the pilot computer:
5000through5003 - Optional network diagnostics: ROV port
7700
If the ROV address differs, set:
$env:ROV_HOST="192.168.1.4"See Network Guide for tether routing, firewall requirements, stream ports, and diagnostics.
Use TritonPilot to operate the vehicle and save operational data. Use TritonAnalysis to interpret mission-specific images, videos, measurements, and manually entered task data. A healthy workflow is:
- Start TritonOS on the ROV.
- Start TritonPilot on the pilot computer.
- Verify controller, telemetry, management RPC, and video.
- Record the stream/raw sensor logs needed by the team.
- Hand saved logs or measurements to the analysis computer, either by USB
drive or the integrated read-only
Analysis Transfer Link. TritonPilot shows the
transfer server state as
Analysis Sharein the status bar.
TritonPilot is allowed to display raw data and diagnostics, but it should not grow mission-scoring applets. That separation keeps the piloting station predictable under competition pressure.
TritonPilot can send arm/disarm edges, hold-mode requests, gain changes, and manipulator commands. The ROV-side TritonOS service remains responsible for arming checks, command freshness, output limits, and hardware safety, but the pilot computer is still part of the live control chain.
Before water tests, verify the controller mapping, BACK/ROV/ARM gain settings, telemetry freshness, video orientation, and save location. During bench tests, keep the vehicle secured and remove props whenever thrusters may spin unexpectedly.