Adaptive-Robust Fusion Strategy for Autonomous Navigation in GNSS-Challenged Environments

Authors:

Abstract
High-precision positioning and navigation is highly important for unmanned vehicles in GNSS-challenged environments. The main aim of this paper is to develop an adaptive-robust fusion strategy for low-cost GNSS/SINS integrated systems with aiding information predicted by data predictors, which can provide reliable fusion positioning solutions when the GNSS signal is challenged. For handling the GNSS degraded problem, we make an adaptive-robust modification to the Kalman filter by introducing a new adaptive factor that can accurately adjust the estimation error covariance matrix and Kalman gain according to the real process. In addition, we design a self-organizing data-driven modeling network with a broad-deep structure for synthesizing navigation data predictors, in order to struggle with the GNSS denied problem. To testify the effectiveness and robustness of the new fusion algorithm, practical experiments with real data sets gathered from road tests in urban areas have been carried out. The results, that is with more than 80% increase in both north and east direction in GNSS-challenged area of our data sets, show that the proposed adaptive-robust fusion strategy can significantly improve the continuity and reliability of integrated navigation, and offer a more precise, robust and reliable solution for autonomous navigation in GNSS-challenged environments.
Funding Information
  • National Natural Science Foundation of China (62273051)
  • Collective Intelligence and Collaboration Laboratory (QXZ23013403)

This publication has 34 references indexed in Scilit: