The field of autonomous navigation and sensing is rapidly evolving, with a focus on developing robust and accurate methods for simultaneous localization and mapping (SLAM) and place recognition. Recent research has explored the use of 2D lidar measurements, colorimetric leaf sensors, and novel descriptor methods to enhance navigation and sensing capabilities in challenging environments such as forests and agricultural settings. Instance segmentation for point sets and 3D instance segmentation have also seen significant advancements, with the development of new neural network architectures and weakly supervised methods. Furthermore, researchers have proposed innovative approaches to LiDAR-inertial odometry and 6DoF direct LiDAR-inertial odometry, enabling more accurate and efficient navigation systems. Notable papers include:

• Robust 2D lidar-based SLAM in arboreal environments without IMU/GNSS, which achieves lower positional and angular errors than state-of-the-art algorithms.
• Sketchy Bounding-box Supervision for 3D Instance Segmentation, which proposes a novel weakly 3D instance segmentation framework that achieves state-of-the-art performance on benchmark datasets.