One of the most challenging problems in autonomous navigation is moving from GPS-friendly outdoor environments into indoor spaces where GPS signals fade. Our approach combines RTK differential positioning with LiDAR SLAM in a unified fusion engine.
At the core of the system is a continuous pose estimator that monitors the signal quality of both positioning sources simultaneously. As the vehicle approaches a building, RTK fixed-solution quality declines while SLAM features grow stronger. The fusion engine smoothly adjusts the weighting between the two, maintaining a continuous and consistent position estimate without sudden jumps.
This allows the Pangolin AT-1 to drive from an outdoor yard into a warehouse, cross the loading area, and return outdoors — all without stopping and without manual operator intervention.