Best practices for architecture of ARKit app

I saw these 2 new architectures which work great for SwiftUI - ELM and the composable architecture (see PointFree for more info).

I know that UIKit apps in general has MVC, MVVM, MVVM-C, ELM, VIPER and probably many more. SwiftUI seems more like a MVVM model, ELM or Composable.

I am building an AR app which has a menu home screen, calendar screen, tracking metrics screen, as well as a camera capture scene.

I know that ideally capture sessions shouldn't be started when not in use, so only when needed to avoid use of expensive computations. But I don't know how best to architect my app i.e which architecture to use for a mixed ARKit and SwiftUI app with many screens not just for camera capture. What are the best practices for building mixed AR apps?