When the pose optimization of temporary mapping trigger?

[narutojxl]

Dear authors, Thanks for your excellent open sourced work for us to learn. After reading paper i have some questions to seek your help, have not read code for now. I hope I won't make you feel nagging. Thanks in advance!

• 1) When the pose optimization equ(8) of temporary mapping trigger? Temporary mapping started at global matching inlier ratio belows threshold u_E. In the system architecture Fig. 2, when each anomaly detection arrives, does the pose optimization trigger ? Or until the reliable global matching inlier ratio exceeds threshold u_M, the optimization of equ(8) trigger? If it's the latter situation, when we go through a long changed environment, how do we ensure when we enter again unchanged environment, the global matching will again success? During the temporary mapping, we use the latest updated odometry drift T^M_O as constant to transform lio's output pose into map frame and odometry drift T^M_O not update. In other words, constant odometry drift maybe have a big difference error with true odometry drift, and global matching will not able to recover from failure?
• 2) During the temporary mapping phase, What is the localization system's output? Is it lio's output pose transformed with above-mentioned constant odometry drift?

Best regards narutojxl