Skip to content

Lab Machine Quick Start

This page is the default starting point for a normal lab member.

It assumes:

  • you are using the dedicated lab data-collection machine
  • the machine-level base setup was already done by an admin
  • ROS Jazzy and the required apt packages are already installed
  • you are logging into the normal collection account such as shared_account
  • that account has already been provisioned for collection use

This page is not for rebuilding the machine from scratch.

If you are provisioning or repairing the machine itself, use:

Normal Operator Workflow

For the normal shared-account path, the operator should only need to:

  1. log into shared_account
  2. open a terminal
  3. run:
collect

That command should launch the operator console directly.

The operator should not need to manually:

  • source ROS
  • activate .venv
  • type the Qt launch command
  • rebuild the viewer
  • provision the workspace

What You Usually Do Not Need To Repeat

On the already-prepared lab machine, normal users usually do not need to:

  • install ROS Jazzy
  • install Ubuntu apt packages
  • install node or npm
  • perform other machine-level admin setup

Those are machine-level concerns, not per-user setup.

What You Still Need To Do

The shared collection account still depends on prior maintainer setup, but the operator should not need to perform that setup during normal collection.

What still matters operationally is:

  • the account has the required device-group membership
  • the account's collection launcher exists
  • the expected workspace and environment already exist

The checked-in launcher is the repo-root script:

  • collect

Any account can run the script directly from the repo root:

~/spark-workspace/spark-data-collection/collect

If you want the short command name collect, expose that repo-owned script on PATH, for example:

mkdir -p ~/bin
ln -sfn ~/spark-workspace/spark-data-collection/collect ~/bin/collect
echo 'export PATH="$HOME/bin:$PATH"' >> ~/.bashrc
source ~/.bashrc

After collect launches the console, continue with:

  1. Hardware Bring-Up
  2. First Raw Demo
  3. First Published Conversion
  4. First Viewer Review

Quick Checks

Before going further, these should already work for the shared account:

type collect
id

The group list should include:

  • dialout
  • plugdev
  • video

And these machine-level commands should also already work:

source /opt/ros/jazzy/setup.bash
ros2 --help

If any of those checks fail on the shared machine, that is not a normal user workflow issue. Treat it as machine/account provisioning work and go to:

Calibration

Calibration is not part of the shortest first-smoke-test path, but it is a real subsystem and should be treated as first-class work before serious data collection: