This review article provides an overview of the use of inertial and visual sensors and discusses their prospects in arctic navigation of autonomous vehicles. We also review the fusion algorithms used so far for integrating vehicle localization measurements as well as the map matching algorithms relating position coordinates with the road infrastructure. The review reveals that the conventional fusion and map matching methods are not enough for navigation in challenging environments, like urban areas and Arctic environment. We also provide new results from testing inertial and optical sensors in vehicle positioning in snowy conditions. We find that the fusion of Global Navigation Satellite Systems (GNSS) and Inertial Navigation System (INS) does not provide the accuracy required for automated driving, and the use of optical sensors is challenged by snow covering the road markings. While extensive further research is needed to solve these problems, fusion of GNSS, INS and optical sensors seems to be the best option due to their complementary nature.
A positioning approach combining satellite measurements with a map representing the ground-truth trajectory is developed with the main objective of improving the availability of solutions for a mobile vehicle. For the positioning model, the Precise Point Positioning (PPP) technique is augmented with an alternative map-matching to find a probable space where the true vehicle or platform position is located. Then, by using a selection criterion based on the precise carrier phase residuals, the best candidate position within the space can be determined. This process provides an accurate initial position to the PPP filter, different from the standard PPP approach that relies on a point position using the less accurate pseudorange observables. A controlled experiment of a mobile receiver navigating over a pre-defined trajectory was conducted. The results show that the approach offers an instantaneous initial convergence, eliminating the re-convergences during two GNSS obstructions of 32 and 17 seconds, while constantly keeping the solution on the correct trajectory, even when tracking 3 to 2 satellites. This approach outperforms the standard PPP and RTK solutions in terms of convergences and re-convergences. These results are corroborated when comparing the average and standard deviation of residuals to the standard PPP model. For the pseudorange residuals, improvements of 17.5 cm and 24.3 cm in the average and standard deviation respectively were achieved. The carrier phase residuals standard deviation of the proposed approach was 3 cm better than that of the standard PPP.
RESUMOAtualmente, a necessidade de coordenadas confiáveis tem sido um dos objetivos da comunidade científica e prática. Desta forma, a análise de robustez de uma rede geodésica, tem como objetivos, com base nos erros máximos não detectados, analisar se a rede é "robusta" ou não. A rede será robusta se a influência destes erros for pequena, caso contrário é "fraca", ou seja, "não robusta". Esta análise se faz com a fusão de duas técnicas, uma que trata da análise estatística de confiabilidade e outra sobre a análise geométrica de deformações. A análise de confiabilidade fornecerá o erro máximo que não pode ser detectado por testes, após o ajustamento. Depois de encontrar estes erros, a análise geométrica de deformações determinará o potencial de deformação que esta rede terá com base nestes erros. Ressalta-se, ainda, que a análise de robustez não é dependente de datum, refletindo somente na geometria da rede e na acurácia das observações (VANÌCEK et al., 2001). Portanto, este trabalho tem como propósito contribuir com as investigações científicas sobre redes geodésicas, checando a mesma, com base em sua geometria e observações. Palavras-chave: Robustez; Confiabilidade; Deformação. ABSTRACTCurrently, the need of reliable coordinates has been one of the main objectives of the scientific and practice community. Thus, the robustness analysis of a geodetic network, aims, at analyzing if the network is "robust" or not, based on the maximum undetectable errors. The network will be robust if the influence of these errors is
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.