I don't believe this is a project for a beginner in FEA for a number of reasons.
CAD software works with BRep geometry. When an FEA is initiated that BRep geometry goes through a meshing stage. This is a volumetric mesh, most often comprised of tetrahedrons. It is NOT the often encountered .stl or .obj surface mesh.
So in order to get this to work with Fusion 360 you need a BRep, not a surface mesh. When you import a surface mesh into Fusion 360 and convert it into a highly faceted BRep. The meshing algorithms work best on simple geometry, which is also the reason Fusion 360 and other CAD software allow to simplify even native CAD geometry in order to help the meshing process.
It is very possible that the algorithms that do the meshing stage badly choke on that highly detailed faceted nature. Even if they succeed you'd be wise to inspect the generated volumetric mesh to determine whether it is suitable for FEA analysis. That requires a good bit of experience!
I cannot find it at the moment, but remember a Autodesk University lecture on FEA on welds. The recommendation was NOT to do it due to the large number of variables in welded structures that can render simulation results invalid or unusable.
Also, the Paper you listed did NOT state there is no difference. It states:
It can be confirmed therefore that, the use of a 3D scanned model instead of the original CAD is not a cause of significant error if mesh geometry has a minimum quality.
That is a heck of a different statement, and also you need to take in consideration what FEA they conducted and on what materials. Rubber and Steel obviously have very, very different properties.