Solid-State LiDAR SLAM System with Indoor Degradation Scene Compensation

Solid-state LiDAR has become a research hotspot in the field of indoor Simultaneous Localization and Mapping (SLAM) technology benefit from its low cost and lightweight characteristics. However, due to the limited field of view, solid-state LiDAR often encounter significant positioning errors and poor mapping performance in indoor environments with sparse geometric structures and long corridor environments. In this paper, a robust solid-state LiDAR SLAM method with indoor degradation scene compensation is proposed. Firstly, in the odometry estimation stage, a perturbation model is used to detect the directions of degradation caused by insufficient constraints from edge and plane features and to evaluate the degree of degradation, isolating and compensating for the degraded state. Then, in the back-end optimization stage, the degraded-compensated LiDAR factors and IMU factors are fused using a nonlinear graph optimization approach, which further improves the degradation issues and enhances the accuracy of the system. Finally, we conduct systematic experimental tests in three typical real-world degradation scenarios to validate our method. The results demonstrated that the map created by our method was consistent with the actual environment, with reconstruction accuracies of only 0.01 m at a range of 6.14 m and 0.3 m at 46 m. Additionally, the computational efficiency of our algorithm was improved in each stage, with an overall time reduction of 32.47%.