PulseAugur / Brief
EN
LIVE 22:15:06

Brief

last 24h
[1/1] 222 sources

Multi-source AI news clustered, deduplicated, and scored 0–100 across authority, cluster strength, headline signal, and time decay.

  1. A Machine Learning Framework for Weighted Least Squares GNSS Positioning based on Activation Functions

    Researchers have developed a new machine learning framework to improve the accuracy of Global Navigation Satellite Systems (GNSS) positioning, particularly in challenging urban environments. The system uses activation functions to transform machine learning predictions about signal quality into weights for a weighted least squares algorithm. Experiments in Hong Kong and Tokyo showed that sigmoid activation functions consistently provided the most significant improvements in positioning accuracy across various machine learning models and GNSS configurations. AI

    IMPACT Improves location accuracy in challenging environments, potentially benefiting autonomous systems and location-based services.