On the Achievability of Submeter-Accurate UAV Navigation With Cellular Signals Exploiting Loose Network Synchronization

A framework that could achieve submeter-level unmanned aerial vehicle (UAV) horizontal navigation in multipath-free environments with cellular carrier phase measurements is developed. This framework exploits the “loose” synchronization between cellular base transceiver station (BTS) clocks. It is shown through extensive experimental data that the beat frequency stability of cellular BTSs approaches that of atomic standards and that the clock deviations can be realized as a stable autoregressive moving average model. This BTS clock model is referred to as loose network synchronization. A rule-of-thumb is established for clustering the clock deviations to minimize the position estimation error, while significantly reducing the computational complexity. The presented models allow the UAV to achieve sustained carrier phase-based meter- to submeter-accurate navigation. To demonstrate the efficacy of the developed framework, this article presents three UAV flight experiments in Southern California, USA, utilizing signals from different cellular providers transmitting at different frequencies. The three experiments took place in open, semiurban environments with nearly multipath-free, line-of-sight (LOS) conditions, in which the UAV traveled 1.72, 3.07, and 0.61 km, achieving a horizontal position root mean squared error of 36.61, 88.58, and 89.33 cm, respectively, with respect to the UAV’s on-board navigation system.