The sample scanning app for ARObjectAnchor says the object can not be moved when scanning. Is there some way to get the BOTTOM of the object scanned? ie: if I have a cube with a different pattern on all 6 sides, I'd like to be able to accurately represent it regardless of how the user sits it on the table, so if I can't move it while scanning, how do I scan the bottom side? Also, it can be difficult to get a nice even lighting setup for scanning when you have to move the camera around the object. Is there any way to do ARObjectAnchor scans inside a light-tent using a turntable?

Playing with ARObjectAnchor and I'm finding out it doesn't seem to handle moving objects very well. When I move the object it is lost and gets re-detected a few seconds later. Is there any way to track it continuously? For example, if I scan a box, I would like the user to be able to pick it up and move it, then have AR objects follow it. Is it possible to do this real time or does ARObjectAnchor just not update that often? I had a similar problem with ARImageAnchor initially but discovered changing some settings greatly increased the update rate. Is there something I can do to speed up ARObjectAnchor?

Been programming a long tiime but still getting used to swift. What I need to do is add a has_been_updated flag to all my ARAnchors that can be set by the didUpdate delegate and later cleared by something else. Not sure of the correct way to do this, if it can be done with extensions and how to apply it to all ARAnchors. The application, I am sending the ARKit anchor data over UDP to an external (non Apple) device on another computer. Because ARKit updates anchor data very frequently, I want to mark the data as "has been updated" when I get a didUpdate event and then every few seconds iterate over all the anchors and only send the ones that have been updated. This reduces data a lot because in the course of a second something like a meshAnchor might update 10 times. What would be the appropriate way to add this flag to the ARAnchor class in Swift?

I've been working with ARKit plane detection and have noticed a few problems. The desk in my office is basically U shaped but when ARKit always turns it into one big plane with the center part closed off. Shouldn't it really detect this as 3 planes arranged in a U shape so the center part is empty? Having similar problems with rooms that open into hallways. The walls of the room get recognized and turned into big planes which leaves the open doorway covered up by the plane representing the wall it's in. This doesn't seem right, am I doing something wrong or is this the way it's supposed to work?

The Nintindo switch game "Super Mario Kart Live: Home Circuit" is an AR game that uses a camera on a small RC car to generate an AR view on the switch as you race around the house. I'm wondering if Apple will support this kind of 3rd party camera device to be used with ARKit? Something like a car or a drone with a camera sending video and IMU data back to an iOS device by WiFi/Bluetooth where ARKit would track and render an AR view on the iPhone. Even something like augmenting video from a security camera with ARKit face detection? I assume Apple would want to approve the device or specify some approved camera module you could use in one. The question is, has there been any talk of supporting third party devices like this in ARKit? I haven't seen any mention of it yet.