Library with protocol support for different 433MHz switches and weather stations for the RFControl Arduino library.
Project description
Python library rfcontrolpy
Introduction
rfcontrolpy is a Python library and port of the node.js rfcontroljs module for parsing and constructing 433mhz On-Off Keying (OOK) signals for various devices, switches and weather stations.
It works together with the RFControl Arduino library for receiving the signals.
The Python port now contains a working controller and a dozen of protocols. Not all protocols are ported yet due to low demand or lack of hardware.
You can find a list of all supported protocols here.
The Processing Pipeline
Receiving
The arduino is connected via serial bus to the processing computer (for example a raspberry pi) and waits for rf signal.
Mostly all 433mhzw OOK signals from devices are send multiple times directly in row and have a longer footer pulse in between. They differ by the pulse lengths used to encode the data and footer and the pulse count.
RFControl running on the Arduino detects the start of a signal by its longer footer pulse and verifies it one time by comparing it with the next signal. It is unaware of the specific protocol, it just uses the stated fact above. Also we are not interested in it if the pulse was a high or low pulse (presence or absence of a carrier wave), because the information is decoded in the pulse lengths.
We will call the received sequence of pulse lengths now timings sequence. For example a timing sequence in microseconds could look like this:
288 972 292 968 292 972 292 968 292 972 920 344 288 976 920 348
284 976 288 976 284 976 288 976 288 976 916 348 284 980 916 348
284 976 920 348 284 976 920 348 284 980 280 980 284 980 916 348
284 9808
You can clearly see the two different pulse lengths (around 304 and 959 microseconds) for the data encoding and the longer footer pulse (9808 microseconds).
All observed protocols have less than 8 different pulse length and all pulse length do differ by at least a factor of 2. This makes a further compression and simplification possible: We map each pulse length to a number from 0 to 7 (a bucket) and calculate for a better accuracy the average of all timings mapped to each of the bucket. The result is something like that:
buckets: 304, 959, 9808
pulses: 01010101011001100101010101100110011001100101011002
To make the representation unique, we choose the buckets in ascending order (respectively we are sorting it after receiving from the Arduino).
We call the sorted buckets pulse lengths, the compressed timings pulse sequence and the length of the pulse sequence (inclusive footer) pulse count.
Protocol Matching
We detect possible protocols by two criteria. The pulse length must match with a small tolerance and the pulse count must match.
Protocol Parsing
If a protocol matches, its parse
function is called with the pulse sequence. Most protocols are
parsed almost the same way. First the pulse sequence must be converted to a binary representation.
In almost all cases there exist a mapping from pulse sequences to a binary 0
and 1
. In this
example the pulse sequence 0110
represents a binary 0
and 0101
maps to a binary 1
:
pulses2binary_mapping = {
["0110": "0"], # Binary 0
["0101": "1"], # Binary 1
["02": ""] # Footer
}
binary = helpers.pulses2binary(pulses, pulses2binary_mapping)
The binary representation now looks like this:
110011000010
As last step the protocol dependent information must be extracted from the binary representation:
decoded = {
"id": int(binary[:6], 2),
"unit": int(binary[6:11], 2),
"state": binary[12] == "1"
}
Details
RFControl is more sensitive than needed for most protocols. So we get sometimes, depending of the accuracy of the sender/remote, different bucket counts.
This is by design, to catch up further protocols that maybe need a higher sensitivity. The specific
protocol has not to deal with this issue, because rfcontrolpy
auto merges similar buckets before
calling the decodePulses
function of each protocol.
The algorithm is the following:
- Record the (maybe to many) buckets and compressed pulses with RFControl (Arduino / c++)
- Sort the buckets in
rfcontrolpy
prepare_compressed_pulses
- Try to find a matching protocol in rfcontrolpy
decode_pulses
- If we have more than 3 buckets and two of the buckets are similar (
b1*2 < b2
) we merge them to just one bucket by averaging and adapting the pulses in rfcontrolpyfix_pulses
- Go to step 3
Adding a new Protocol
Preparation
- Fork the rfcontrolpy repository and clone your fork into a local directory.
- We are using unittest for automating tests.
- You should be able to run the tests with
python3 -m unittest discover
. - Running
python3 -m build
let it compile all files and whats for changes.
Protocol development
- Create a new protocol file (like the others) in
rfcontrol/protocols/
. - Add a test case in
tests/protocols
with the data from the Arduino. - Adapt the protocol file, so that the test get passed.
Support my work
Do you enjoy using this Python library? Then consider supporting my work using one of the following platforms, your donation is greatly appreciated and keeps me motivated:
Hire me
If you're in need for a freelance Python developer for your project please contact me, you can find my email address on my GitHub profile.
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for rfcontrolpy-0.0.9-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | d0f042dc79127e21014c7c03a5ef658ac77525a48347184c05ecbd7eee66a685 |
|
MD5 | 6913cd51401e4ad409aa48e996ab416a |
|
BLAKE2b-256 | 43940bdd2e49431126ebfe8b73004eef3cf63776c36cb5ce5844978ab4043d12 |