Embodied AI · Non-Line-of-Sight Teleoperation

Non-line-of-sight teleoperation
first-person remote operation

A cloud-rendered communication link: teleoperation commands flow down, first-person video flows up. Put on a NOLO headset, look through the robot's stereo cameras, and drive the robot to perform fine work — shelf restocking, sales, assembly, inspection, maintenance and care — from across a wall, a floor or a city. Head-and-hand 6DoF pose is solved in 2ms, with end-to-end average latency of about 70ms on a LAN.

Cloud-Rendered Link Head & Hand 6DoF Pose Sub-Millimeter Tracking

View Specifications

Operator wearing a NOLO VR headset with motion controllers teleoperating a robot — Control the robot remotely, right from home

4KUltra HDRecommended resolution, robot stereo module

~70msEnd-to-end average latency, measured on LAN

2msHead-and-hand 6DoF capture & processing

Operate precisely without seeing the robot.
Freeing teleoperation from “on-site line of sight” to “any distance”.

Conventional teleoperation requires the operator and the robot to share the same line of sight. NOLO's non-line-of-sight teleoperation is built on a first-person video stream and millimeter-level head-and-hand pose: put on a headset and you “become the robot” — driving it through real work such as shelf restocking, sales, assembly, inspection, maintenance and care from across a wall, a floor or a city. It is about getting the job done, not about collecting data.

Line-of-Sight Constraint

Conventional teleoperation relies on the operator looking directly at the robot — cross a wall, a floor or a city and it stops working. Every deployment has to be designed around “being able to see”.

Latency Stacks Up

Capture, encoding, transmission, decoding, display, feedback — delay in any single stage collapses the “feel” of remote control.

NOLO SOLUTION

A cloud-rendered link that brings the site to your eyes

The robot's stereo module streams a first-person stereo view to the operator's headset in real time; head-and-hand 6DoF pose is solved by our in-house StarTraq® sub-millimeter tracking, strictly synchronized with the video and sent over the network to the robot. Controller actions are mapped to robot commands in real time through sub-millimeter tracking — what you see is what you reach, and what you move is what arrives.

Teleoperation commands down, perception data up — one cloud-rendered link, closed loop
Head-and-hand 6DoF pose captured and processed in 2ms, strictly time-aligned with video frames

BIDIRECTIONAL LINK

A cloud-rendered communication link — control down, perception up

The robot side and the operator side close the loop of “seeing” and “acting” over the same link — no separate perception and control channels required.

OPERATOR SIDE

Headset renders the first-person stereo view in real time
Head-and-hand 6DoF pose capture and processing
6DoF motion controllers

COMMUNICATION LINK

Downlink: teleoperation commands sent to the robot
Uplink: perception data (stereo video stream) returned to the operator
Pose prediction lead time compensates network jitter

ROBOT SIDE

Stereo module captures the stereo view
Encodes and streams 4K Ultra HD imagery
Receives pose commands and drives the actuators

CORE CAPABILITIES

Millisecond-level latency — the capability base for non-line-of-sight teleoperation

Sub-Millimeter Pose Tracking

In-house StarTraq® positioning holds sub-millimeter accuracy and millisecond-level latency in complex real environments — the robot goes exactly where you aim.

<1mm accuracy ｜ <5ms latency ｜ 250Hz refresh

2ms Head-and-Hand 6DoF

Head and both hands are solved per frame with just 2ms of capture and processing, strictly time-aligned with video frames — commands and motion stay in sync at millisecond level.

2ms capture & processing ｜ strictly synced with video

Pose Prediction Lead Time

Future motion is predicted from the current data stream, compensating delay introduced by the network and display chain for smooth, drift-free control.

<11ms lead time ｜ smooth, no drift

First-Person Stereo Display

The robot's stereo module streams a first-person stereo view to the headset, restoring true depth perception — a full sense of presence from a thousand miles away.

4K Ultra HD

Dexterous Controller Interaction

Motion controllers are tracked by StarTraq® at sub-millimeter precision; 6DoF pose is regenerated frame by frame, and button and trigger actions map to robot commands in real time — precise grasping, aimed exactly where you point.

Sub-millimeter tracking ｜ real-time action mapping

SPECIFICATIONS

Core Specifications

Video & Display Recommendations

VIDEO & OPTICS

Robot Stereo Module	Resolution no lower than 4K Ultra HD (combined stereo)
Video Link	Image capture & encoding → network transmission → terminal decoding → screen presentation
Screen Presentation	Display interval approx. 11ms (set by frame rate)
Terminal Decoding	Video decode latency approx. 8ms
Communication Link	A cloud-rendered link: teleoperation commands down, perception data up

Pose & Tracking Metrics

TRACKING & CONTROL

Positioning	StarTraq® (acoustic-optical-RF and visual-inertial fused tracking)
Pose Accuracy	<1mm accuracy, <5ms latency, 250Hz refresh rate
Head-and-Hand 6DoF	Capture and processing in 2ms, strictly synchronized with video frames
Pose Prediction	Lead time <11ms, predicting future motion from current data
Motion Controllers	Sub-millimeter 6DoF tracking of both hands; button and trigger actions map to control commands in real time

Make every remote task as reliable as being on site

Book a demo or discuss your non-line-of-sight teleoperation deployment with us. Based on your robot platform and network conditions, the NOLO team will advise on everything from hardware selection and link design to deployment tuning.

Business & Partnership

400-668-6017 bd@nolovr.com

NOLO Non-Line-of-Sight Teleoperation

Non-line-of-sight teleoperationfirst-person remote operation