Abstract
Determining the root causes of problems for complex electromechanical systems is difficult. Systems can detect fault states quite reliably using simulated software models; however, even when a fault is detected, it can be difficult to determine the underlying reasons, and resolution method, for that fault. Some anticipated faults may be automatically recovered from, but others are more complex and require humans to understand their root causes before resolving them.
In this paper, we discuss the work that we’ve done to address these issues within a data visualization product. We discuss the major frontend components, as well as our highly modularized backend, which we developed in order to make a system that not only simplifies the telemetry monitoring and fault diagnosis tasks, but does it in a way that’s extensible to a wide variety of systems.
Read full paper or the poster for more details.
Team Members
Development Notes
Setup
The server is flask, which can be installed minimally using:
pip install flask
Running
To run the visualization server/client combo, in the root directory either run:
flask -a viz --debug run # Requires Flask 1.0
or:
python viz.py
This will start a localhost server on port 5000.
The first command requires Flask 1.0, I believe, but handles restarting the server for new code in a more reliable way.
Using python -m viz might be preferable to the second case, but might require some finagling to get working.
After the server is running, the frontend will be accessible via a web browser connected to localhost on port 5000.