ORB-ODOM: Stereo and odometer sensor fusion for simultaneous localization and mapping

• Autor(a) principal: Caldato, Brenno A. C. [UNESP]
• Data de Publicação: 2017
• Outros Autores: Filho, Ricardo Achilles [UNESP], Castanho, José Eduardo C. [UNESP]
• Tipo de documento: Artigo de conferência
• Idioma: eng
• Título da fonte: Repositório Institucional da UNESP
• Texto Completo: http://dx.doi.org/10.1109/SBR-LARS-R.2017.8215301; http://hdl.handle.net/11449/171104

ORB-SLAM2 is one of the better-known open source SLAM implementations available. However, the dependence of visual features causes it to fail in featureless environments. With the present work, we propose a new technique to improve visual odometry results given by ORB-SLAM2 using a tightly Sensor Fusion approach to integrate camera and odometer data. In this work, we use odometer readings to improve the tracking results by adding graph constraints between frames and introduce a new method for preventing the tracking loss. We test our method using three different datasets, and show an improvement in the estimated trajectory, allowing a continuous tracking without losses.