Regarding 'Perhaps the community can help point me to a good reference' then I found this a nice explanation of ICP. Further searching shows that there are alternative approaches to ICP for point cloud alignment that may be 'better' in some circumstances. FWIW, I personally haven't seen anyone using ICP in JMP (but that doesn't mean it hasn't been done!). Looks like the essence of a good implementation is singular value decomposition, so I would expect it's perfectly possible in JSL, and maybe not too much work. But, depending on your situation, constraints and knowledge, a native JMP implementation may not provide the path of least resistance. For instance, you could try to use JMP's Python integration with this (or another) approach.
