LiDAR Only Odometry

LiDAR-only odometry focuses on estimating the movement of a vehicle or robot using only data from Light Detection and Ranging (LiDAR) sensors, eliminating the need for additional sensors like IMUs. Current research emphasizes continuous-time approaches, modeling LiDAR point clouds as a continuous stream of data and employing techniques like Gaussian process regression to estimate trajectories more accurately. This is driven by a need for robust and computationally efficient solutions, often incorporating strategies like adaptive local mapping and point cloud sampling to improve real-time performance and accuracy, particularly in challenging environments. The resulting advancements in LiDAR-only odometry have significant implications for autonomous navigation and robotics, offering a more reliable and potentially lower-cost localization solution.