Скачать или смотреть Reliable fusion of black-box estimates of underwater localization

This video presents the experimental setup where the BAR-F Neural Network heuristically performs adaptable fusion of information for estimating the underwater vehicle localization. The experiment is shown from different perspectives, including an on-board camera fixed to the surfboard, the SONAR acquisitions, which field of view is delimited by a line in magenta. The ground truth trajectory performed by the ROV, the BAR-F estimation, and the USBL sensor readings are shown on the map. The comparison on the evolution of the accumulated localization error between the different methods explored and our proposal is presented.

This work has been presented in the IROS2018 conference. For more details about the article, please consult the following reference:

Reliable fusion of black-box estimates of underwater localization. 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) doi: 10.1109/IROS.2018.8593593. Available at: https://ieeexplore.ieee.org/abstract/...

The GNU Octave implementation sources is available at: https://github.com/henferch/BAR-F