GNSS-based environmental context detection for navigation

Abstract

Environmental context detection is a topic of interest for the navigation community since it enables to build a context-adaptive solution. Indeed if the type of environment is known it is then possible to choose the proper data processing algorithm or to select the sensors to be used to dynamically adapt the navigation solution design itself. This paper proposes to build a supervised machine learning model which can robustly classify multiple contexts such as urban canyons, urban, trees and open-sky areas using GNSS data only. A training and test database have been built with four datasets acquired at different times in order to prove the relevance of the solution. These datasets are made available to the community for research purpose. The choices of features and classifier are also discussed and compared to others papers. Our solution achieved an average 82.40% of classification accuracy.

Vision-enhanced GNSS-based environmental context detection for autonomous vehicle navigation

“…As a first step, our previous work [5] of GNSS-only SVM classifier is applied to this new city-scale dataset. The original Table III provides the resulting F1-score for the tree test cases.…”

“…We first show the classification results of the Support-Vector Machine (SVM) classifier trained on GNSS data only. It is based on our previous work [5], but applied to larger city-scale datasets. As it is revealed that this GNSS-only SVM model has difficulty in distinguishing the two moderate contexts (Trees and Urban), this paper proposes to enhance the model by adding satellite NLOS information obtained from semantic segmentation on sky-oriented fisheye camera images.…”