A Review of Visual SLAM Systems Based on Multi-Sensor Fusion

In the realm of autonomous systems, the precision and robustness of Simultaneous Localization and Mapping (SLAM) are significantly bolstered through the integration of diverse sensor modalities. This review explores the critical role of multi-sensor fusion in visual SLAM, which amalgamates data from Inertial Measurement Units (IMUs), visual cameras, LiDAR, and more. We commence with an overview of the foundational principles of visual SLAM and trace its evolutionary trajectory. Subsequently, we dissect the core tenets of multi-sensor fusion, elucidating the types of sensors involved and the strategies and algorithms employed for their synergistic integration. As we traverse through the landscape of current research, we identify and discuss the emerging trends and future horizons of this dynamic field. The review culminates in an analysis of the challenges inherent to multi-sensor fusion systems, proposing innovative solutions to surmount these obstacles and set the stage for future advancements.