ⓘ Detailed Overview

Autonomous drone landing presents a critical positioning challenge: outdoor RTK GPS loses accuracy indoors, while optical flow solutions struggle with variable lighting and surface conditions. Marvelmind's indoor positioning system bridges this gap with ultrasonic-based indoor GPS technology. The solution uses four station super-beacons deployed on a landing platform (1x1 or 2x2 meters) with beacons facing upward toward the drone's approach corridor. A mobile beacon mounted on the drone facing downward communicates with station beacons to establish precise coordinates. During the initial 30-meter approach, the wide baseline between beacons creates an elongated triangle, resulting in precise Z-axis data but degraded X-Y accuracy. As the drone descends and approaches the landing zone, the geometry improves dramatically, providing full 3D precision. For final landing phases where the drone is directly above the beacon array, Marvelmind employs Precise Z configuration: a primary submap provides accurate X-Y coordinates while a secondary submap optimizes Z-axis precision. This dual-submap approach ensures consistent accuracy throughout descent and contact. The system eliminates GPS handover latency and provides centimeter-level repeatability for autonomous warehouse operations.