Inspiration

The Autosense challenge had a wide area of sub-challenges, like physics, visualization, presentation and communication through webtechnologies. This was our motivation to choose this topic.

What it does

The Solution takes json-files through an appealing Web-UI and processes them. The outcome visualizes the greates impact on the car during a crash and also presents some additional data like G-Force, time of crash and the offset in the time-series.

You can also get just the angle and the offsetMaxForce through the api "/api/v1/getCrashInfo".

How we built it

As a Foundation we used a Sanic v18 Webserver to implement the API-Calls. In the background all the logical functions and classes are neatly implemented with Python. For the visualization part we decided to use OpenCV 4.0. We also implemented the whole project as a Docker Container, so it can be deployed anywhere, anytime.

To add some extras to the project, we decided to build a small neat WebUI, there we used some basic HTML5, CSS and of course, Javascript.

Challenges we ran into

The physics, to calculate the crash impact gave as definitely some headaches.
We got some weird values sometimes, that was because we ran into a transformation issues while transforming the x,y,z values with the oneG Reference. After fixing these Problems, the visualizations looked more consistent.

Accomplishments that we're proud of

We are proud of:

  • Solid math equations
  • maintainable Code
  • Clean Code
  • Visualization

What we learned

We repeated a lot of maths and OpenCV which we once learned and forgot again ;-)
The Hackathon was a really good practice for us as a group, we learned how to interact better and more efficient with each other, to tackle a larger project like the autosense challenge.

What's next for autosense_group_asimov

Next we want to polish up our solution, also we would love to get some more crash data and some real-images so we can check if we calculated the right things. Also there is an Azure Webspace waiting for our app to be published

Share this project:

Updates