Real-time Sensor Fusion Techniques in Indoor and Urban Environments
Students should have strong programming skills in Python and C++. Knowledge about Linux and docker is useful, but not necessary. The multi-sensor backpack system has 2 x VLP16 lidars, Ladybug5+ HD camera, BD990 Trimble GNSS, and XSens MTI 630R IMU.
Simultaneous Localization and Mapping (SLAM) techniques, especially Extended Kalman Filtering techniques, can be used for real-time situational awareness. The goal of this thesis is to review the state of the art in SLAM techniques and develop improved methods for sensor fusion in indoor and urban environments using a multi-sensor system.
The objective of this thesis is to develop advanced SLAM techniques for mobile mapping in indoor and urban environments using open datasets and a multi-sensor system provided by the laboratory. This will involve a review of the state of the art in SLAM, including recent developments in optimization methods, sensor fusion, and deep learning. The student will then develop and implement improved SLAM methods in C++ and Python, using datasets provided by the laboratory or collected by the student. The performance of the new methods will be evaluated and compared against existing SLAM techniques. Specific objectives are discussed with the student; and may include extending the method in [1] so that it can be used with multiple lidars and cameras to create more accurate spatial representations. In this line of work, machine learning and semantics can be leveraged [2].
[1] Lin, J., & Zhang, F. (2022, May). R 3 LIVE: A Robust, Real-time, RGB-colored, LiDAR-Inertial-Visual tightly-coupled state Estimation and mapping package. In 2022 International Conference on Robotics and Automation (ICRA) (pp. 10672-10678). IEEE.
[2] Hughes, N., Chang, Y., & Carlone, L. (2022). Hydra: A real-time spatial perception engine for 3d scene graph construction and optimization. arXiv preprint arXiv:2201.13360.