lidar slam

**Lidar SLAM** is the **mapping and localization framework that uses sequential laser scans to estimate trajectory and build 3D maps with high geometric precision** - it is widely used in autonomous driving and outdoor robotics where long-range accuracy is critical. **What Is Lidar SLAM?** - **Definition**: SLAM system based on point cloud registration and geometric map optimization. - **Core Inputs**: Lidar scans, optional IMU and wheel odometry signals. - **Output**: Ego trajectory and globally consistent 3D point or surfel map. - **Representative Methods**: ICP-based pipelines, LOAM variants, and graph-SLAM frameworks. **Why Lidar SLAM Matters** - **High Metric Accuracy**: Laser range data provides strong geometric constraints. - **Lighting Robustness**: Works in low-light and high-contrast environments. - **Long-Range Coverage**: Suitable for large-scale outdoor mapping. - **Autonomy Dependence**: Core component of many vehicle-grade localization stacks. - **Map Fidelity**: Produces detailed geometric maps for planning and obstacle handling. **Lidar SLAM Pipeline** **Scan Registration**: - Align incoming scan to local map using feature or point matching. - Estimate incremental pose. **Map Update**: - Integrate aligned scan into map representation. - Maintain local and global map structures. **Graph Optimization**: - Add loop closure constraints and optimize full trajectory graph. - Reduce accumulated drift over long runs. **How It Works** **Step 1**: - Extract geometric features from lidar scan and register against current map. **Step 2**: - Update map and run backend optimization with loop closure when revisits occur. Lidar SLAM is **a precision-first localization and mapping approach that provides robust large-scale 3D geometry under diverse lighting conditions** - it remains a backbone technology for autonomous mobility systems.