Setting virtual size of the camera in raycaster

Hello, I have been asking a question related to raycaster in past and the responses I got, really helped me. However. Im trying to prevent the camera from moving past specific model, or its children. In some point, Im am able to actually have some decent detection of intersection with some models, but only barebone. What Im saying is, that it seems, that it interacts with the ground. Which I would not include, but it is needed, as I want to support different height levels. So if I give the point to the map, Im stuck and else Im falling through the map. The point is, that Id like to use directly the model, which is position perfectly with the camera, however, for some reason, raycaster just does not want to work with that. So, Id like to ask. How can I use the model, virtually increase its width and depth (and increasing the position, so it wont intersect with the ground). The reason I want to virtually increase its size or rather, offseting where the raycaster captures the intersection and sets velocity to 0 is that when now I move, Ill move to the object and then get stuck in it, bcs Im already in the object, which will permanently spam velocity 0, while if I do it only if I get close, I can move to other sides.
I know Im being unclear, but Im pretty bad in this tbh :D, its my 9th day using three.js