I would be interested in knowing whether anybody ever attempted to study, in consistent field-theoretic terms, the formation of complex particle tracks (including secondary vertices) in collider detectors.
As far as I know, tracks observed in real detectors are routinely analyzed with software that models them as if they were trajectories of classical point particles originated from the decays, applying geometrical considerations and classical conservation laws to identify the interaction vertices and decay products. This kind of analysis has allowed the successful experimental verification of the Standard Model, which is a quite remarkable feat.
However there seems to be a quite general consensus among theorists that fully localized point particles are not compatible with the QFT framework, so the emergence of such apparently point-like entities leaving traces, generating secondary vertices, being finally detected at specific locations etc. can at best be explained in terms of approximately localized interacting wave packets.
In this context, I suppose that if we could assume that the primary decay generates a set of approximately localized excitations originating at the primary interaction vertex position, the subsequent steps of the process could be described in terms of approximately point-like particles. In my understanding ,this would amount to assuming that the primary interaction generates quasi-asymptotic states in a scattering theory sense.
On the other hand, I know that, at least in one famous case (alpha particle decay, Mott 1929), it is possible to explain the formation of tracks in a detector even without invoking any (approximate) preliminary localization of a primary spherical wave originating from a decay event. In this case there would not even be a need for defining a primary vertex position: the primary tracks would originate at random positions and directions in the detector bulk. But then I wonder how this approach could be extended to cover complex decay processes involving multiple interactions, i.e. multiple tracks, and still reproduce the apparently classical observed trajectories.
More generally, has this problem ever been tackled in detail from a theoretical and/or simulation point of view?