Fast geometry-based ambiguity resolution algorithm for real-time kinematic - Archive ouverte HAL Accéder directement au contenu
Article Dans Une Revue Research Square - Preprint Année : 2023

Fast geometry-based ambiguity resolution algorithm for real-time kinematic

Thibaut Castan
  • Fonction : Auteur
  • PersonId : 1217248

Résumé

Abstract Centimeter-level positioning using real-time kinematic technology is achievable only after overcoming the step of ambiguity resolution. Fast algorithms providing such a solution tend to have low success rates, often resulting in time to first fix too long in many real-life cases. On the other hand, other algorithms, such as the LAMBDA method, provide a better success rate at the expense of more computations. The recent growth in the number of constellations available for positioning along with the multiplication of signals per satellite is challenging slower algorithms which prove to have a complexity exponential with the number of signals. This article presents a new approach to the integer ambiguity problem, showing in an experiment a success rate of 100% with a computational speed ten times higher than the MLAMBDA algorithm, the complexity of the search process being O ( n 2 ) with n the total number of signals. We consider here a short baseline model to introduce the algorithm. It is straightforward to extend the results to longer baselines using a more complete model, as the hypotheses for the algorithm to be successful are made explicit along the article. Classical techniques developed to improve ambiguity resolution, as well as computational tricks could be used to further improve the speed of this algorithm, and could hence lead to improved time to first fix in low-cost embedded equipment.

Dates et versions

hal-03946363 , version 1 (19-01-2023)

Identifiants

Citer

Thibaut Castan. Fast geometry-based ambiguity resolution algorithm for real-time kinematic. Research Square - Preprint, 2023, ⟨10.21203/rs.3.rs-2438681/v1⟩. ⟨hal-03946363⟩

Collections

INSMI
12 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More