Longitudinal analysis with c hanging PUMA boundaries

Hello,

In my current project, I need to analyze some housing trends using PUMS data (housing needs assessment and estimating the future housing demand and so on).

However, changing PUMA boundaries is making my job harder. Currently, I am working with household and person data files I downloaded from Census ftp site. But someone suggested, if I use IPUMS for various years' samples and PUMA variable, they adjust for changing boundaries. Is that the case? If not, how can I find reliable crosswalks for PUMA boundaries in different years (I am planning to work with years 2006-2023 (1-year sample))? I want to merge with data (hopefully using R) so that I can work with this data.

Thank you