Solid-State vs. Mechanical LiDAR for Land Vehicle Navigation: A Comparative Study Using LiDAR Odometry

Reliable and accurate positioning is a fundamental requirement for ensuring the safe and efficient operation of autonomous vehicles (AVs), especially in dynamic and complex environments. With its active sensing capability, light detection and ranging (LiDAR) technology offers high accuracy, long-range coverage, and reliable 3D representation of the surrounding environment. These attributes have established LiDAR as an essential technology for improving the overall robustness of autonomous driving and facilitating significant advancements in navigation state estimation research. The growing popularity of LiDAR technology has driven continuous innovations aimed at enhancing quality and reducing costs. This progress has led to the emergence of new types, such as solid-state LiDARs (SSLs), which are gaining widespread adoption due to their reliability and cost-effectiveness. In this paper, we present a comprehensive analysis of the key differences between conventional mechanical LiDAR and SSL technologies. Furthermore, we assess their respective performance using various state-of-the-art LiDAR odometry (LO) algorithms. Our comparative analysis is supported by road test experiments conducted using a multi-sensor platform equipped with both SSL and mechanical LiDAR, which simultaneously collect point cloud data during typical land vehicle urban driving dynamics. Our findings offer valuable insights into the merits and limitations of these technologies within the context of resilient positioning and navigation in complex urban environments.