Which connection works for which teleop task
Pick your link and your task. The check uses published delay limits1.
- Round trip
- 10–30 ms
- Task needs
- 100 to 200 ms round trip
- Note
- Domestic. Under 80 ms intercontinental. Recommended for sessions over 4 hours.
Thresholds and speed limits published by one teleoperation vendor (April 2026), read at the top of each connection's range. Speeds between its 50, 150 and 300 ms points follow a straight line.
What one robot maker measured on its own VR setup
Delay adds up from camera to screen and from hand to gripper. These are the maker's numbers, on its own hardware2.
Rules to set before every session
Round trip, or postpone the session
Jitter; under 10 ms is healthy, above 30 pause
Sustained bandwidth; 3 to 6 Mbps for 2 to 4 cameras
Operators plug in for production sessions
With no heartbeat, the robot ramps to a stop
One teleoperation vendor's published rules1.
See station setups →Questions buyers ask
What latency is acceptable for robot teleoperation?
It depends on the task. Published vendor guidance puts contact work under 50 ms round trip, pick and place at 100 to 200 ms, and anything over 500 ms down to supervision only.
Can operators in another country drive our robots?
For pick and place and supervision, usually yes over fibre. One vendor puts intercontinental fibre under 80 ms round trip. For precise contact work, keep operators on the same local network.
Is average latency or jitter more important?
Jitter. A link that averages 40 ms but spikes to 400 ms is worse than a steady 120 ms. Measure the spread, not only the average.
What happens when the link drops?
Set a watchdog. In one vendor's rule the operator station sends a heartbeat 10 times a second; after 500 ms of silence the robot ramps to zero over 200 ms, then brakes.
Operators far from your robots?
Tell us where the robots and the people are. We read every brief and come back to you.