Supported by the European Space Agency (ESA) NAVISP program, UK-based Omnisense has successfully completed the DroneHome-2 project. The result is a highly resilient, multi-sensor navigation architecture enabling reliable autonomous return-to-home and landing operations for drones operating in GNSS-degraded or denied environments.
GNSS multipath and signal obstructions can threaten safety-critical landing phases when drones operate near marine vessels, offshore platforms, and industrial infrastructure. To maintain bounded, deterministic drone behavior, DroneHome-2 integrates a ground-based, ultra-wideband (UWB), local positioning system (LPS) with a high-performance flight navigation stack. The architecture fuses UWB pseudoranges with inertial measurement unit (IMU), GNSS, barometer, and magnetometer data within a custom 21-state extended Kalman filter.
The project dealt positively with several key regulatory and engineering hurdles. Because airborne UWB transmission is heavily restricted, drones can operate in a receive-only mode, relying on one-way time of arrival (ToA) ranging using UWB pulses sent from ground beacons.
To make this work without physical clock synchronization, the central LPS network controller runs a joint timing and location engine (JTLE). This software maintains a ground-network timing model (TM4) that calculates real-time clock offsets and drift. Corrections are shared back to the ground tags via a Wi-Fi backhaul, enabling them to broadcast UWB chirps with a mathematically aligned time of departure (ToD) to the drone.
Confirmed precision
At a recent event hosted by ESA, Omnisense researchers presented the results of trials of the DroneHome-2 system. Structured flight simulations and controlled field tests were used to evaluate both GNSS-enabled and GNSS-disabled landing scenarios. During complete GNSS denial, the navigation engine’s explicit modeling of airborne clock bias and drift enabled maintenance of a highly stable, bounded positioning accuracy of 30 to 50 centimeters. When operating in an assisted, GNSS-supported regime, landing accuracy was under 20 centimeters.
The trials showed that overall precision is geometry-dependent. Collinear beacon layouts increase dilution of precision (DOP), meaning operational landing zones need to employ ‘surround’ geometries to ensure optimal horizontal and vertical constraints.
DroneHome-2 builds on the previous DroneHome project, which was paused in 2022 following an industrial partner change. Omnisense successfully retained the core architecture, shifting focus from initial feasibility specifications to quantified, physical performance validation.
The team now plans to introduce DroneHome-2 on commercial market, starting with maritime and offshore vessel applications where relative navigation constraints are high, followed by fixed-geometry industrial applications, and eventually scaling to urban corridor operations.
