I think you may have to re-read your two linked articles.
From the first one:
…when a smart home lock is locked, it sends a packet to the hub, and the smart home hub passes that onto the server … but using the patterns, the size of the packet and the timing of the packet, we can figure that information out with very high accuracy."
It clearly talks about the packets, which are the RF-Signals exchanged between the hub and the devices.
And from the second article:
Specifically, this work focuses on the devices that use Zigbee or Z-wave and are controlled by a centralized smart-home hub in a personal area network
Zigbee and Z-wave are used for RF communication. So essentially just a more specific example of what was said in the first article.
On top of that, the example of preventing a lock from locking is practically done by flooding the relevant frequency with packets, so that the lock can’t hear the command to perform the locking.
All this has nothing to do with the communication that Home Assistant is doing. If Home Assistant is wired to the router, then the traffic could only be intercepted if the attacker is already within the network. And at this point obfuscating via packet padding won’t do much, as you have way bigger problems at that time.
Of course also stuff like WiFi or Bluetooth can be monitored for patterns, which could give some insights about what is happening in your home. But again, that’s nothing Home Assistant can do anything about. Well, it could randomly poll a devices state or send invalid commands to make a mess out of what the attacker is gathering on information. But at least battery powered devices would be drained by that in the long run.
All that being said, my assumption is that at best only high profile targets would become victims of such an attack, as this takes time. As Tediore said, regular burglars tend to reach for the low hanging fruits. They enter quickly if they see a possibility, look for valuable items in places where people typically store them, then leave again as quickly as possible. Analyzing a house prior to entering only makes sense if they know they’ll make a huge profit from it. And (hopefully) such high profile home owners tend to have a lot of wired systems in their homes, which by design can’t be monitored wirelessly. Hence they aren’t (as) exploitable.
In my opinion a fingerprint-lock at your front door is WAY more interesting for criminals. Chances are, the owners fingerprints are scattered all around the house. Extracting it from his car-door wouldn’t take long. Then come back a few days later with a fake one and you’re in. At least if the fingerprint-scanner isn’t the most secure one.