ok, for 1. presumably you want something international and consistent. And presumably you mean passport/nationalID/DL as "the docs", right?
For that, the classic org was
@CAcert. They (we) did all the work to handle the verification of docs against an individual.
Sadly, they are a shadow of their former selves bc they could not see to pivot out of the CA business. They got locked into a failed goal of getting their root into the browsers, but they got locked out of that process bc reasons. (disclosure. I had a lot to do with that...).
I don't believe there is another org that fits the bill above, but if you say more, maybe something else pops up.
2. my view of identity is here:
iang.org/identity_cycle/ In short I say there are 4 forms of identity, and a preliminary claim is that only one form is useful for our purposes (but see presumptions in 1. above.
A related technical view over that claim is that identity is the edges of a graph of people. IE, my statement "we exchanged tweets today" is an edge between you and me; that is a suggestion that all the useful data is of that form - I say something about you.
Others have other views - there is Self-Sovereign-Identity, a decade long project started by OP
@ChristopherA. This is at an unfortunate juncture bc (IMHO) it tried to encompass standard approaches and ended up being .. a target for BigTech to muscle in, embrace&extend, etc, before it had actually showed much success.
OP is restarting that project, now called Next. It's at an early stage, using groups to analyse/improve a set of requirements for an identity approach.
3. which algos... do you mean, certificates? There are 3 somewhat recognised ways here: OpenPGP, x509, and SSI. All have major problems
4. recognition ... in what context? court only?
5. if you want courts involved, it won't be international. What you can do is set up arbitration - internationally. See CAcert for the model on that.