Embodied AI Glossary中文

GNSS / Real-Time Kinematic Positioning

GNSS / RTK 定位GNSS / RTKAdvanced

Combines satellite navigation with base-station differential correction to achieve centimeter-level positioning outdoors.

GNSS is the umbrella term for global navigation satellite systems, including GPS, BeiDou, GLONASS, and Galileo; used alone, positioning error is typically at the meter level. RTK (real-time kinematic positioning) sets up a base station at a known location, streams its observed satellite carrier-phase data to a mobile receiver in real time, differences the two to cancel out most shared errors, and then resolves the carrier-phase integer ambiguity, achieving centimeter-level accuracy; network RTK uses multiple base stations to widen coverage. Its limitation is that it needs a clear view of the sky — it fails indoors, between buildings, and under tree cover due to blockage and multipath reflection, so real systems are usually fused with an IMU, wheel odometry, or lidar SLAM. In embodied AI, this is mainly used on outdoor robots: inspection quadrupeds, agricultural and lawn-mowing robots, drones, and self-driving vehicles.

ExampleAn outdoor inspection quadruped robot carries an RTK antenna on its back, fusing it with IMU and lidar for localization, and autonomously patrols a campus following preset latitude-longitude waypoints.

Also called
RTK, Carrier-Phase Differential GPS, Satellite Positioning
Related
Inertial Measurement Unit · Multi-Sensor Fusion · State Estimation · LiDAR SLAM · Inspection Robot · Autonomous Driving
Sources
Wikipedia: Real-time kinematic positioning

See it in the full glossary →