230men / can

ni

Nissan LEAF CAN Bus data collection using Raspberry Pi and OSIsoft PI System

A vehicle’s CAN (Controller Area Network) is its way of communicating between each of its electronic components through a central bus. They send strings of bits that identify which component is "talking," followed by its current status or state. Some examples of the time series data that gets communicated:

• Temperature & kWh remaining on battery pack
• Speed of the vehicle, RPM of each wheel
• Flag warnings including: battery malfunction; high temperatures; and high current & voltage differentials

By using a Raspberry Pi and a PiCAN2, it's possible to log the CAN messages directly from the vehicle onto the RPi. From there, this collection of Python scripts used in conjunction with a dictionary of data translations can be used to turn the raw CAN bus output into meaningful, interpretable data.

Setting up the Raspberry Pi 3

To enable the use of the PiCAN2 board, you need to first configure your RPi. Connect to your RPi's GUI in your preferred method (mine is with a HDMI to monitor, and USB mouse and keyboard). Make sure that you have the newest version of the OS installed, and that everything is up to date. Open up a new terminal window, and enter:

sudo apt-get update
sudo apt-get upgrade
sudo apt-get autoremove
sudo reboot

Next, install the can-utils package, which includes all of the executable for viewing, sniffing, and logging the CAN messages. In the terminal, enter:

sudo apt-get install can-utils

Next, we have to edit the boot configuration file to recognize the PiCAN2 board on start-up. In the terminal, enter:

sudo nano /boot/config.txt

This brings up the text file, and allows you to make changes. Add the following text, then exit out of the file with CTRL-X, making sure to save your changes:

dtparam=spi=on
dtoverlay=mcp2515-can0,oscillator=16000000,interrupt=25
dtoverlay=spi-bcm2835-overlay
dtoverlay=spi-dma-overlay

Configuring a virtual CAN device

This procedure will configure a VIRTUAL can device, which is perfect for testing out your procedures. You can simulate the same type of data stream that you would be getting when plugged into the OBDII port of a car and receieving CAN messages.

In your terminal, first enter:

sudo modprobe vcan

This allows the module vcan to be loaded to the Linux kernal. Next, we will add a new vcan device. I'm going to call it vcan0. In the terminal, enter:

sudo ip link add dev vcan0 type vcan
sudo ip link set vcan0 up

Check to make sure the device was added by looking for it when you enter this in the terminal:

ifconfig

You should see it along with any other devices you have loaded.

Configuring a real CAN device

This procedure is specific to the Nissan LEAF, but it very similar for other vehicles' CAN buses. The LEAF specifiic step is setting the bitrate to 500000. Every vehicles CAN sends messages at a specefied bitrate, and the LEAF's happens to be 500,000. The most common ones are 33,333 bps, 50 Kbps, 83,333 bps, 100 Kbps, 125 Kbps, 250 Kbps, 500 Kbps, 800 Kpbs, and 1,000 Kbps. If you configure your device with the wrong bitrate, you will probably not recieve any data.

Similarly to the virtual CAN setup, we start with the modprobe:

sudo modprobe vcan

The next step is the same except for the set up. USE YOUR VEHICLE-SPECIFIC BITRATE

sudo ip link add dev can0 type can
sudo ip link set can0 type can bitrate 500000 listen-only on

Check to make sure the device was added by looking for it when you enter this in the terminal:

ifconfig

You should see it along with any other devices you have loaded.

Collecting Data

To collect data, make sure the PiCAN2 board is properlly installed and that you have a CAN device set up on your RPi. To make the actual connection, you can insert a 20-gauge wire into the CAN-H and CAN-L ports on the PiCAN2, and the OBDII port on your car (refer to this guide for help http://www.cowfishstudios.com/blog/canned-pi-part1)

Once you have established the physcial connection, you can test that the data is being recieved by the RPi. My favorite way to do this was to connect the RPi to my laptop through ethernet, so I could use it's GUI. Make sure your car is on, and open up a terminal window on the RPi.

Run a simple candump can0, and you should see the steady flow of CAN messages. If you don't see the stream, then keep fiddling with the different OBDII ports (I found that many times I had them switched). Be careful while plugging in and out of the OBDII port, it can be fragile!

To save the data, explore different option of the candump command. candump -l can0 logs all of the CAN frames, candump -l can0,284:7ff logs all CAN frames of message ID 284. There are many options here, you can refer to the can-utils documentation for more help, or http://www.cowfishstudios.com/blog/canned-pi-part1).

Logging, Translating, and Sending the Data to Google Firebase

PLEASE READ OVER THE PYTHON CODE BEFORE RUNNING. AGAIN, IT IS SPECIFIC TO THE NISSAN LEAF. IF YOU ARE ATTEMPTING THIS, MAKE SURE THAT YOUR OWN PARAMETERS ARE ALL ENTERED CORRECTLY

Make sure you have a dedicated directory on your RPi for the CAN bus data. Mine is locatied /home/pi/Documents/CAN. This is where all of the Python scripts will be run from, and where the log files will be stored. Also be sure to create a file inside the directory to store the log files once they are closed. Mine is /home/pi/Documents/CAN/log.

To use the Python scripts, open 3 terminal windows on the RPi. In all of them, first cd to the correct directory:

sudo cd /home/pi/Documents/CAN

In the first terminal, enter:

python canDirNotify.py

In the second terminal, enter:

python logDirNotify.py

Make sure that you have the proper connections from the section above, then in the final terminal, enter:

candump -l -n 100 can0

This should open the pipeline for the data to be logged in batches of 100, translated, then sent to your Google Firebase.

Getting Data from Firebase to PI

This final step requires the use of the PI System, specifically the PI Data Archive, PI AF, and PI UFL Connector. From a machine that is on the same domain as your PI System, run firebaseToUFL.py. This will check the Firebase every 0.5 seconds (can be configured in the code), and sends a get request which retrieves all of the data. It then builds out a long string, containing all of the formatted data, and sends a put request to the UFL Connector. Enter your own UFL Connector's URL