You can not use a pre-defined animation sequence on a detailed track. There are several problems to this thing:
--- circular movement along the track profile - you have to define the Y and Z movement of a single track piece for an circulatory manner, that has a defined shape. (is not as much the rieal problem)
--- incremental Z movement variation - meaning, when the track-leading wheels are suspended, you have to modify the Z coordinate of the tracks according to the nearest one or two leading wheels suspension state --- if the circular function works, this can be taken as an additional stage, where you need also to know, when, during the circular rotation, the track is at the place of 1-st, 2-nd ... leading wheel and in this position it will add or take the suspension wheel position induced Z-value.
--- track lenght - and this is the most problematic (and have no image how to approach) is, that when the suspension additions and takes from the incremental Z movements change, on local level, lenght of the track between two leading wheels differs, meaning, if nothing is done with it, it will lead in forming gaps and mesh-collisions (one mesh going into another) witch would look like it disassembles on the move. its not only, that these are the case, but you also have an "stretching-wheel" , witch, when all leading wheels go up, has to compensate the loss of the stretching force (usually the front or rear one (opposite the engine-connected one giving the movement force to the track)) by going along the Y axe (front-back, Z is up-down).
--- speed - it has to be taken from the vehicles engines props and be in synch. with all the wheels armatures falling into the gaps of the track - at points where suspension works, it might fall out of it, and the stretching one too, as at stretching/relaxing the track, its rotation speed differs according to the tracks "stretch".
--- wehicle mass induced animations - meaning, if an armor speeds up or brakes, its leaning itself on its suspension forward and back. Doe, as the force and stretching wheels are always mounted on the vehicle frame, it isnt a problem as it ends up being taken up from the leading-wheels suspension movement. BUT there are designs, that could have the track-compartment on an different level. One of such vehicles is DT-30 Vityaz:
... its not a vehicle whyte a cart, but an two-segment vehicle where the rotation of the "cart segment" can be manipulated by an hydraulic system at the connection place between them (its for enhanced maneuverability in especially bad terrains and clearance manipulation). There would be a need to find a way to have aether double-physics objects (simply said having each component its own physics calculated but working then as one) or a way to modify the seconds tracks orientation in regard to the models main pivot point. ( But dont think anyone is going t try that stuff
)
This all needs normal wheels defined in the vehicle script (the leading ones) doing the actual work and the animation getting real-time status info about theyr position, affecting the mentioned math for a single track. I would prefer finding an function, where you do this for a single track and each other would have just some "delay" constant, placing it behind the forward one.
