Speeding Up The Estimation Of Floated Ambiguities For Sub-Decimeter Kinematic Positioning At Sea

Using carrier phase DGPS to find the precise position of a ship or a buoy in the high seas presents the problem that phase ambiguities cannot be resolved with confidence over very long baselines. The ambiguities may be "floated" instead. However, the convergence of the Kalman filter in a kinematic solution with floated ambiguities (i. e. solving for real-valued Lc biases) is often quite slow. But it could be speeded up by imposing the dynamic constraint that on a ship, or a buoy, changes in elevation are likely to be quite gradual, once waves and faster variations are smoothed out. To test this idea, two GPS data sets, one from a buoy in the US and another from a ship in Japan, have been processed in differential kinematic mode, floating the ambiguities with and without a constraint on the mean height. Each set contains several hours of uninterrupted data from a rover and from a nearby coastal site. In the US, the site was in Duck, North Carolina; in Japan, it was in Yaizu. In the US, high-rate data from distant receivers (350 km-1100 km) were also available for long-range solutions.