Ricardo B. Sousa PhD Thesis: SLAM 2D - Traditional Pose Graph Optimization (Raw Version 0.1) скачать в хорошем качестве

Ricardo B. Sousa PhD Thesis: SLAM 2D - Traditional Pose Graph Optimization (Raw Version 0.1)

This video represents a 2D Simultaneous Localization and Mapping (SLAM) framework with 2D laser scanner and wheeled odometry data. The framework uses as a front-end a 2D Laser Odometry-Only Tracker (   • Ricardo B. Sousa PhD Thesis: 2D Laser...  ) to generate the keyframes and relative odometry estimation between poses. Then, the SRRG2_SOLVER ( https://gitlab.com/srrg-software/srrg... ) is used as the backend to optimize the graph in a traditional pose graph optimization (factors are the relative odometry estimations and the poses in the graph are the optimization variables). Minor changes implemented relative to version 0.0 in terms of the tracker matcher and loop closure aligner. Left window: keyframe change and scan matching of the laser odometry tracker. Right window: SLAM 2D Pose Graph evolution. PhD Candidate Ricardo B. Sousa (0000-0003-4537-5095) ( https://sousarbarb.github.io/ ) Supervisor at Sapienza: Prof. Dr. Giorgio Grisetti (0000-0002-8038-9989), Sapienza Università di Roma Supervisors at FEUP / INESC TEC: Prof. Dr. António Paulo Moreira (0000-0001-8573-3147), FEUP & INESC TEC Senior Researcher Héber M. Sobreira, PhD (0000-0002-8055-1093), INESC TEC Prof. Dr. Manuel F. Silva (0000-0002-0593-2865), ISEP & INESC TEC Hosting Faculty: Faculty of Engineering of the University of Porto (FEUP) ( https://sigarra.up.pt/feup/en/ ) Hosting Institution: CRIIS - Centre for Robotics in Industry and Intelligent Systems ( https://www.inesctec.pt/en/centres/criis ) from INESC TEC - Institute for Systems and Computer Engineering, Technology and Science ( https://www.inesctec.pt/en )