UC Riverside

Recent advances have established Precise Point Positioning (PPP) as a preferred technique for providing accurate and precise position solutions for users anywhere on the globe. PPP relies on the availability of accurate corrections of satellite orbits, clocks, and hardware biases, in addition to other corrections which are estimated using a global network of reference stations. These global corrections allow PPP to achieve centimeter-level positioning accuracy using pseudorange and phase observations. However, PPP suffers from long convergence times, mainly due to residual errors from atmospheric delay corrections. Therefore, the current challenge for PPP lies not in improving its accuracy, but rather in improving the convergence time. This dissertation presents two novel approaches to improve PPP convergence time by focusing on the estimation of regional atmospheric delays.

First, a new approach is developed to enable collaboration between satellites in estimating regional atmospheric delays using measurements from a single receiver. This approach takes advantage of the increasing availability of multi-GNSS signals by using uncombined measurements from all usable satellites. By estimating the ionospheric delay as a regional Vertical Total Electron Content (VTEC) map parameterized using polynomial B-splines, the approach allows satellites to collaborate in estimating each other’s slant delays. Results show that this collaboration improves the stability and precision of ionospheric delay estimates, leading to faster convergence of PPP solutions. The method is not limited to multi-frequency users and also benefits single-frequency users.

Furthermore, the dissertation proposes a collaborative PPP approach that utilizes estimated regional atmospheric corrections from a network of local agents. This approach combines corrections from global models with estimates from local collaborating agents to derive more accurate regional corrections. The resulting collaborative estimation problem is formulated as an average consensus problem, and a distributed information-weighted consensus algorithm is developed to solve it. The algorithm allows agents to share only their atmospheric delay estimates and uncertainties with their neighbors. This distributed approach eliminates the need for a central processing center and enables agents to instantaneously achieve network accuracy upon joining the network, thereby significantly improving PPP convergence time. The collaborative PPP approach is shown to achieve instantaneous convergence for some users under favorable conditions, demonstrating its potential for real- time precise positioning applications.