To address the issues of pose errors and mapping drift of traditional two-dimensional laser simultaneous localization and mapping(SLAM)systems in geometrically feature-sparse degraded scenarios

an improved method based on Cartographer is proposed.An adaptive grey wolf optimization(GWO)algorithm is first introduced to optimize the initial pose

which avoids local optima and scan matching failures.Subsequently

a dual-stage precise registration consisting of point-to-point iterative closest point(ICP)and point-to-line ICP is employed.Precomputed map normals are utilized to balance the stability and accuracy in pose estimation.Furthermore

a three-stage optimization scheme consisting of candidate selection

matching validation

and constraint construction is designed to improve the robustness and efficiency of loop closure detection.Experimental results demonstrate that

on the corridor sequence of the OpenLORIS-Scene dataset

the root mean square error of the absolute trajectory error(ATE)of the proposed algorithm reaches 0.127m.In practical corridor scenarios

the mapping length error is 1.67%

effectively alleviating map aliasing and drift.