Best Energy-Efficient Radio Protocol for Battery-Operated Building Automation

I don't quite understand what you want to control with this communication. The most energy efficient is a radio receiver and remote control. Just not suitable for everything.

Perhaps the ZigBee protocol?

https://pl.wikipedia.org/wiki/ZigBee

Which car has the lowest fuel consumption? The answer to this question is simple. One that is not driven. The same is true in this case. The most energy-efficient solution will be one where devices transmit very little data, over short distances, and do so very infrequently.
Energy efficiency is not a characteristic of a given interface or protocol. For example, Bluetooth Low Energy, as the name suggests, should be energy efficient. It usually is. However, the power consumption can vary thousands of times depending on the mode of operation and what the designer thinks of.

You probably expect a simple straightforward answer. It is impossible to give such an answer. If many different interfaces and communication protocols are used, it means that there is no single best one. Each has its own advantages and disadvantages. What may be a strength in one solution may be disqualifying in another. If you want to do something then unfortunately you have to start by learning the basics. There are certain stages of learning that cannot be skipped.

When it comes to popularity as measured by the number of devices produced, I think WiFi and Bluetooth Low Energy will come first. Every smartphone, smartband and tablet has these interfaces. From this perspective, ZigBee and Z-Wave can be considered niche solutions.

gshesiu wrote:

for me it is not a problem to make my own protocol using serial communication in the microcontroller and some module e.g. usart-"RF", the relevant experience I rather have.

.
You are already making a mistake in your assumptions at this stage. Battery-powered low-power devices often use specialised processors integrated into the radio interface. In practice, communication via USART requires a relatively large amount of energy. If the sensor is to be powered for a year by a small battery, e.g. CR2032, the average current may not exceed a few tens of uA. Consider in this context how much energy the USART operation (clock, GPIO, etc.) costs.

gshesiu wrote:

It would be good, for example, for the reception of data to wake up the prock

.
You have the wrong approach to the issue. Turning on a radio receiver is a gigantic power consumption. Doing it permanently will drain the battery quickly.
For example, with BLE, the device periodically (every 1 sec or so) sends a notification frame. After this operation, listening is switched on for a very short time. If nothing is received then the device goes to sleep. Statistically, the radio interface is switched off almost all the time.
The other side, e.g. the smartphone, has power in abundance. Here listening can be switched on for a long time. Note that there are two types of devices in this standard. One is very energy efficient. However, this is achieved at the expense of the others, which draw a thousand times more current.
Home automation devices can usually also be divided into similar groups. The first is all kinds of battery-powered sensors. The second is mains-powered actuators.

gshesiu wrote:

for example, a device similar to a wall-mounted battery-operated temperature controller set occasionally via the internet (app on a phone etc., on the internet side an RF transmitter/receiver on mains power with an RF transceiver, sort of like a control unit for RF devices).

.
The problem can be solved differently. The controller you mention is in the general case a thermometer and an algorithm for switching on, for example, a cooker. This example cooker rather needs to have mains power. Let's assume that we transfer the algorithm to the cooker. In this situation, we have a battery-operated thermometer that does not require any remote configuration and a cooker that has mains power and can be permanently connected to the Internet, e.g. via WiFi.

XBEE Modules without unnecessary complication simple network enormous documentation.

If the range allows it, I suggest Bluetooth® 5 / Bluetooth Low Energy and specifically the nRF52840. It contains an integrated CPU so you make the device on one chip. There are lots of modules and documentation for this.
If 2.4 GHz will not get you through then I would suggest ZigBee but in a sub 1 Ghz version.

gshesiu wrote:

I would like to make a device that could communicate energy efficiently with other devices within a house several rooms over.

.
Do you want to build this yourself (and have the knowledge to do so), or are you looking for off-the-shelf?
In the latter situation, I can't advise you. However, if you have some experience in programming and electronics then:
- practically any modern radio module is suitable, e.g. from the cheap and popular RFMxxx - you have modules from 433 MHz to 2.4 GHz, the frequency you choose depends on the conditions, a mix of different frequencies can be used.
- The current consumption for a radio transceiver depends almost exclusively on the time it is switched on. Short frames transmitted e.g. every second (as in BLE) mean that if, for example, the active time is 1 ms, the transceiver consumes 40 mA, effectively the device consumes an average of 40 uA, battery life will be very long.
- The key is the right protocol - BT BLE, ZigBee or others, or something of your own. There are some ready-made ones on the web, it's worth a read.
- modern microcontrollers consume little current (about 1-2 mA in such an application). They have sleep modes where current consumption is a fraction of a uA.
- to make it all so beautiful, you need some experience in writing low-power applications.
- note that often much more current than the MCU + transceiver is consumed by the voltage converter, hence the need to design this part of the circuit carefully.
- for AA or AAA batteries, self-discharge current may be more important than the current consumption of the circuit. At least for ultra-low power applications this is relevant.
But back to radio communication - when limiting current consumption by limiting active (listening) time, you need some kind of intermediary to collect the messages and send them out at the right moment. The chance that two active modules will synchronise for e.g. 1 ms is poor, hence it is worth having at least one normally powered module working as a gateway.
As for the question about the most economical standard - I've highlighted to you how it more or less works - how much power it consumes depends on you - how often you want to transmit something, how much data you have, etc. E.g. you're measuring temperature - if you take a measurement every second it will use some current, if you take a measurement every minute for example it will use 60x less with some approximation. As you can see, a well thought-out concept is important.