Skip to content

Safety

Teleoperation is for supervised, human-in-the-loop control — recovering a stuck robot or guiding it through a situation autonomy can't handle. Treat it as direct control of a real machine.

Before you take control

  • Confirm you are connected to the correct robot (check the robot name on the session).
  • Maintain clear line of sight or reliable situational awareness of the robot and its surroundings.
  • Make sure the area is clear of people, animals, and obstacles.
  • Verify the video feed is live and your gamepad is connected and responding.

While operating

  • Drive slowly and deliberately. Video and control travel over the network, so what you see lags real life.
  • Watch for latency. Rising packet loss or falling frame rate (see Video streaming) means added delay — slow down or stop.
  • Keep a stop plan. Know how to halt the robot immediately, both from the operator station and via any onboard or physical stop.
  • Don't outrun your awareness. If you lose video, lose the controller, or feel unsure, stop first and reassess.

Stop controls

  • Release the sticks to bring motion to neutral.
  • Click Disconnect to end the session and release control.
  • Use any physical emergency stop available on the robot if you need to halt it immediately.

Teleoperation is designed to fail safe: if the control connection or signal is lost, the robot stops rather than continuing on its last command. Always rely on your robot's onboard safety systems as the final line of defense.

Handing back to autonomy

When you've finished recovery or the task, return the robot to autonomous control deliberately:

  1. Bring the robot to a safe, stationary position.
  2. Restore the appropriate operating mode for autonomy.
  3. Disconnect the teleop session.
  4. Confirm on the Fleet page that the robot is reporting normally before walking away.