Real-time kinematic (RTK) positioning with cars enables a 1-2 cm level accuracy

Last Updated on July 30, 2026 by pg@petergamma.org

Real-time kinematic (RTK) explained

Real-time kinematic (RTK) positioning enables autonomous vehicles and connected cars to achieve 1-2 cm lane-level accuracy, a precision critical for safe navigation where standard GPS drifts by 1-3 meters.  This technology works by using a fixed base station or a network of reference stations to send real-time correction data to a mobile rover receiver in the vehicle, eliminating errors caused by atmospheric delays and satellite clock inaccuracies. 

To maintain reliable positioning, especially in challenging environments like urban canyons or tunnels where GNSS signals may be blocked, RTK is typically fused with Inertial Navigation Systems (INS) and cameras.  This sensor fusion combines the absolute precision of RTK with the relative, continuous data from IMUs and visual SLAM, ensuring the vehicle remains accurately localized even when satellite visibility is compromised. 

Key components and benefits include:

  • Centimeter-Level Accuracy: RTK reduces positioning error from meters to 1-2 cm, enabling precise lane-keeping and automated parking. 
  • Network RTK (NTRIP): Many vehicles use internet-based NTRIP services to access correction data from thousands of global base stations, removing the need for local hardware. 
  • Sensor Fusion: Combining RTK with LiDAR, cameras, and IMUs provides redundant positioning, ensuring safety and reliability for autonomous driving stacks.