- Overview
- What's new
- Requirements
- Installation
- Running
- State files
- Data Typing
- Debugging
- Utilities
- Thanks
The Record and Playback utilities are simple CAN bus module record and playback tools designed to test CAN bus code application code without having to connect to a physical vehicle. Playback accepts input files from the Record and appears as a Mustang Mach-E or other vehicle responding to ISOTP-compliant Read DID requests. Record is implemented on the Unified Diagnostic Services (UDS) protocol and queries the vehicle with Read DID requests (service 0x22) and outputs decoded messages and writes the state changes to output files that can be played back in the Playback utility.
Record is a state machine, dynamically changing the DIDs read from the vehicle depending on if it is idle, on a trip, or charging. Each state has an associated JSON state file that determines how often DIDs are sampled for chagnes.
Both utilities can be configured via a YAML configuration file. Definitions for the vehicle CAN bus modules and supported DIDs use JSON input files in most cases, some Python data structures might exist as everything is a work in progress.
Record runs fine connected to Playback via a loop-back cable (CAN0 connected to CAN1) or the vehicle OBDII port and happily logs the requested CAN bus data. Connecting Playback to your vehicle may have unintended consequences and is not recommended.
- Added VIN as database tag for multiple vehcile support
- InfluxDB support (now with backing cache if internet connection is lost)
- YAML secrets supported
- switched to venv for Python3.10 support
- catch SIGTERM to write out cached data before exiting
- new command line options for setting the YAML and log files
- service files for running Record and/or Playback as a Linux service
- Extract utility
- added default value for read requests that timeout or error
- Geocodio reverse geocoding added
- access accurate GPS location using iPhone app
- better vehicle data for viewing in a iPhone app
- define the minimum trip and charge in the YAML file
- trips and charges are written out to files for easier playback and smaller size
-
Python 3.10 or later
-
Python packages used include (but the list in the
setup.py
file is the definitive list of packages)- python-configuration
- pyyaml
- python-can
- can-isotp
- udsoncan
- influxdb-client
- pygeocodio
Both Record and Playback use SocketCAN for the networking and use UDS and ISO-TP protocols on top of the SocketCAN connections. You need to have access to both of the OBDII connector HSCAN and MSCAN buses if you wish to access all of the vehicle modules. My hardware setup consists of:
- Raspberry Pi 4
- SK Pang PiCAN2 Duo CAN Bus Board For Raspberry Pi 4 With 3A SMPS (powered from OBDII port)
Other hardware may work but your mileage will vary.
- Clone the repository and install the required Python packages:
git clone https://github.com/sillygoose/mme-record.git
- Enable venv in the project:
cd mme-record
python3.10 -m venv .venv
source .venv/bin/activate
- Install the required Python packages:
cd mme-record
pip3 install -e .
- Configure the YAML file
The included file
mme.yaml
is a sample configuration file that can be used with both Record and Playback.
The YAML configuration search starts in the current directory and looks in each parent directory up to your home directory for it (or just the current directory if you are not running in a user profile). Edit mme.yaml
to set the desired Playback and Record options as well as InfluxDB options if you wish to have Record save the data in an InfluxDB2 database.
If you need to switch between YAML and/or log files, use the command line to select the desired file:
python3 record.py yamlfile=my_yaml.yaml logfile=my_log.log
You can now use a secrets file to store sensitive information like the token used to access your InfluxDB database. The secrets file name is the same as YAML file with _secrets
added. For example, the default YAML file is mme.yaml
so the default secrets file is mme_secrets.yaml
, if you used my_mme.yaml
for your configuration the secrets file will be my_mme_secrets.yaml
. The search for the secrets file will be the same as the YAML file.
I use a Tailscale client on Raspberry Pi and on the InfluxDB host so you always have a static IP. Tailscale is an excellent WireGuard implementation and allows me to open an SSH session or write to the database server on my home network no matter where the car is. At home I plug into a wired network or use WiFi, on the road I connect to a cell phone serving as a hotspot and nothing would work without the services provided by Tailscale.
You can run Record from the command line via SSH, using a VS Code remote connection, or at startup using a script. The project has the script run_record.sh
that can be used for this purpose to allow data collection without an SSH or other connection.
I run Ubuntu 20.04 LTS on my Raspberry Pi so the following instructions are tailored for this OS:
Start by editing the Record system service file with your editor:
% cd ~/mme-record
% nano /etc/systemd/system/mme-record.service
You need to fix the path names to reflect the user profile where the MME-Record project is located. Reperat this for the playback service file if you will be running Playback unattended.
Next copy the record and playback system service files:
% sudo cd ~/mme-record
% sudo cp *.service /etc/systemd/system
Enable and start the service with
% sudo systemctl enable mme-record.service
% sudo systemctl start mme-record.service
You can check if it running by using the command line
% sudo systemctl status mme-record.service
You should see something simiar to this:
% sudo systemctl status mme-record.service
● mme-record.service - MME Record
Loaded: loaded (/etc/systemd/system/mme-record.service; enabled; vendor preset: enabled)
Active: active (running) since Thu 2022-01-13 11:07:31 EST; 4min 20s ago
Main PID: 12806 (run_record.sh)
Tasks: 6 (limit: 9257)
CGroup: /system.slice/mme-record.service
├─12806 /bin/bash /home/sillygoose/mme-record/run_record.sh
└─12809 /home/sillygoose/.pyenv/versions/3.10.1/bin/python3 /home/sillygoose/mache-simulator/source/record.py
Jan 13 11:07:31 greta systemd[1]: Started MME Record.
Jan 13 11:07:32 greta run_record.sh[12809]: [2022-01-13 11:07:32,271] [logfiles] [INFO] Created application log /home/sillygoose/mme-record>
Jan 13 11:07:32 greta run_record.sh[12809]: [2022-01-13 11:07:32,305] [record] [INFO] Mustang Mach E Record Utility version 0.6.0
Jan 13 11:07:32 greta run_record.sh[12809]: [2022-01-13 11:07:32,339] [state_manager] [INFO] Vehicle state changed to 'Unknown'
Jan 13 11:07:32 greta run_record.sh[12809]: [2022-01-13 11:07:32,558] [state_manager] [INFO] Vehicle state changed to 'On'
Jan 13 11:10:48 greta run_record.sh[12809]: [2022-01-13 11:10:48,298] [state_manager] [INFO] Vehicle state changed to 'Trip'
Now reboot and reconnect to a terminal and check that Record is running:
% sudo systemctl status mme-record.service
If you don't want Record to run automatically, disable the service and reboot, you can always use the start and stop commands from a terminal to control if Record runs in the background.
All the commands for controlling the MME-Record service are:
sudo systemctl enable mme-record.service
sudo systemctl disable mme-record.service
sudo systemctl start mme-record.service
sudo systemctl stop mme-record.service
sudo systemctl status mme-record.service
JSON state files are used to select the DIDs read in each state, you can control when they start and how often they are repeated.
DIDs involved in state transitions should be read frequently to catch state transitions. Other states can be read as needed, slow changing states should be read less often than rapidly changing data so nt to consume too much CAN bus bandwidth.
Here is a sample state file entry with an explanation of the fields it might contain:
{
"module": "BECM",
"arbitration_id": 2020,
"arbitration_id_hex": "07E4",
"enable": true,
"period": 5,
"offset": 2,
"dids": [
{
"did_name": "ChargingStatus",
"did_id": 18509,
"did_id_hex": "484D",
"codec_id": 18509
}
]
}
module module name (required)
arbitration_id the aribitration ID assigned to the module (required)
arbitration_id_hex optional, easier to recognize than 'arbitration_id' (but )JSON has no hexidecimal support)
enable enables this entry, default setting is true
period how often this entry will be scheduled or 0 if to be scheduled once (every 5 seconds in this example)
offset when this entry will start, default is 0 (this example will start after two seconds has passed)
dids a list of DID dictionaries that you wish to send to the vehicle
The period of 0 is a special case and is only used in the charge ending and trip ending state code to wait for the command queue to empty, signifying that the state variables are up-to-date and can be sampled one last time before recording the trip or charging session.
Each DID dictonary contains the following fields:
did_name name assigned to the DID (required)
did_id ID assigned to the DID (required)
did_id_hex optional, easier to recognize than 'did_id' (JSON has no hexidecimal support)
codec_id Codec assigned to decode the DID response, normally the 'did_id' but could be different
The data used in the software follows the following rules:
voltage (V) float
current (A) float
power (W) int
energy (Wh) int
distance (km) float
speed (kph) float
latitude, longitude (°) float
% float
enumerations int
text data str
These types are also used to encode the InfluxDB fields, changing a data type may be cause for deleting the bucket and starting fresh.
I run both Playback and Record in VS Code on a Raspberry Pi with CAN0 bus tied to the CAN1 bus in a loopback mode, makes for easy testing of Record changes playing back recorded files from a trip or charging session without requiring the vehicle.
I found I needed the ability to sniff the CAN buses but this is not possible on the Mustang Mach-E as the Gateway module makes sure there is no traffic to sniff. Extract is a work-around to this problem, you can use this to extract some or all the DIDs in a module and run these in Record to look for state changes. Just temporarily replace the unknown.json
with the output file of Extract and exercise the vehicle to capture state changes.
Thanks for the following packages used to build this software: