HPLE
High Precision Local Element

Classically, the processing steps in the ambiguity resolution (AR) domain are the following:

• Pre-processing of the observables;
• Computation of the double differences;
• Computation of the float ambiguities;
• Computation of the ‘integer ambiguities’;
• Validation of the integer ambiguities;
• Computation of the precise position.

DIGINEXT has already worked in this domain and has processed real GNSS data. This is essential so as to adapt the algorithms, designed by research organisations, to real world conditions and integrate them in adaptive software architecture. TU Delft is a recognised expert in the AR domain (LAMBDA method) and is continuously developing innovative methods to improve the accuracy, the reliability or the TTFF. The major contribution of TU Delft in this project is the ‘innovative validation’ method aimed at providing improved results in terms of both reliability and rapidity.

CERFACS has a fresh and critical view on the work performed in the AR domain so far. It aims at developing advanced methods for the numerical simulation and the algorithmic solution of large scientific and technological problems of interest for research as well as for industry. Its background in various scientific domains, such as in algebraic graph theory, is an interesting contribution to the other partners of the consortium.

HPLE addresses the problem of improving the accuracy, reliability and the TTFF (time to first fix) in high precise positioning, between two receivers, by processing GNSS (Global Navigation Satellite System) signals and particularly Galileo signals.

Description

The methodology followed so as to design the various algorithms was to respect the following guidelines:

• Improve the quality of the processed data;
• Reduce the integer ambiguities space of search;
• Reliably validate the integer ambiguities;
• Develop a GNSS observables generator to simulate a complete range of scenarios;
• Develop a demonstrator to test the algorithms on real and simulated GNSS data.

Bearing these guidelines in mind, the partners of the consortium analysed the various steps of AR processing and then designed the algorithms. The pre-processing and the DIA procedure are designed to improve the quality of the processed data, the adjustment of the KALMAN and the RLS filters are aimed at reducing the space of search and the ambiguity validation methods (the IVM and the method applied in the RLS architecture) are developed to improve the reliability of the chosen solution. To design the GNSS observables generator, an essential part of the work consisted of analysing the way the observables could be disturbed and biased so as to make this generator as realistic as possible. All the perturbations should be present and easily configurable in the generator. The HPLE demonstrator is the final software that gathers all the developed algorithms and enables their test and validation.

Objectives

The aim of the project is to design, develop and validate an innovative GNSS demonstrator providing the users with a high precision service. To improve the accuracy, reliability and the TTFF, DIGINEXT and TU Delft have merged their knowledge to build a complete processing structure: KALMAN approach:

• Pre-processing of the observables;
• Computation of the double differences;
• Computation of the float ambiguities (KALMAN);
• Refining of the float ambiguities (DIA);
• Computation of the integer ambiguities (LAMBDA);
• Validation of the integer ambiguities (IVM);
• Computation of the precise position.

The innovative validation method (IVM) automatically assigns the threshold value to validate the ambiguities on condition of the maximum allowable failure rate delivered by LAMBDA. In the framework of this project, CERFACS aims to develop an alternative approach based on the use of a RLS filter processing the reduced differences instead of the classical double differences. This led to the following processing architecture: RLS Approach:

• Pre-processing of the observables;
• Computation of the reduced single differences;
• Computation of the float ambiguities (RLS);
• Refining of the float ambiguities (RLS DIA);
• Computation of the integer ambiguities (NEAREST);
• Validation of the integer ambiguities;
• Computation of the precise position.