A tool made with Flask and Chart.js to visualise data from both a Mobileye camera and a driver control unit for use in self-driving cars. The Flask server reads in CAN data by reading the output from the candump with the subprocess module. It then parses the relevant bytes, and sends a json object via HTTP server-sent events to all connected clients. The webpage displays a graph of the left and right lane boundaries, as well as the speed and heading angle of the car.
This tool was intended to run on embedded hardware with a dedicated CAN port. Make sure that your board has the correct drivers before you decide to run it.
If your board does not have Flask, run
pip install FlaskThe application is made for Python 2.7, but it will run on Python 3 as well. Clone this respository and move into it with
git clone https://github.com/falafel72/mobileye_viewer
cd mobileye_viewerInitialise the can port as shown:
ip link set can0 type can bitrate 500000 triple-sampling on
ifconfig can0 upand run the server with
python app.pyAny computer connected to the same network should be able to view the webpage at the url server_ip_address:5000 (replace server_ip_address with your board's ip).
Ideally CAN data would be read using python's can library or Kvaser's canlib, but I was unable to install either of these on the board I was using. Moving the CAN receiving logic to python would eliminate the need to read every incoming data packet, making the program more efficient. More information could also be displayed, such as the locations of cars or other obstacles in front of the car. A more realistic map would be useful for this purpose, requiring a migration away from Chart.js to a more specialised 3d rendering tool.