Failure-resilient Graph-based SLAM for Autonomous Robotic Exploration in GNSS-Denied Environments

Abstract

In this thesis, motivated by the autonomous exploration problem, we address the problem of autonomous robot localization using Simultaneous Localization and Mapping (SLAM) with Light Detection and Ranging (LiDAR) perception enhanced by black-box visual odometry in scenarios where laser scan matching can be ambiguous because of a lack of sufficient features in the scan. We propose to develop a novel localization method based on the Graph SLAM approach that benefits from fusing data from multiple types of sensors to overcome the drawbacks of using only LiDAR data. The proposed localization method uses a failure detection model based on the quality of the LiDAR scan matching and inertial measurement unit (IMU) data. The failure model improves LiDAR-based localization by an additional localization source, including low-cost black-box visual odometers like the Intel RealSense T265. The proposed method is compared to the state-of-the-art localization system LIO-SAM in cluttered and open urban areas. Based on the performed experimental deployments, the proposed failure detection model with black-box visual odometry sensor yields improved localization performance measured by the absolute trajectory and relative pose error indicators. Furthermore, we developed elevation mapping and traversability estimation to employ the proposed localization method in autonomous robotic exploration that is based on the frontier-based exploration strategy. The proposed localization method has been experimentally validated within the developed exploration framework in the outdoor field experimental deployments in the campus backyard, where it allows building successfully aligned map of the environment.