Intro

Nobody likes to pay bills for their car insurance, especially when the car is rarely used. Furthermore, insurance rates are increasing because of other people's negligence. More careful drivers have to pay the same amount of money as drivers with a more aggressive driving style (thus, more likely to crash their car).

STOP!

Together with Volvo and Bosch, we introduce a new platform called Insure Ride. The platform allows Volvo drivers to profit from insurance rates personalised to their unique driving style. We do this by using advanced sensors provided by the XDK together with machine learning and a bit of magic to build accurate driver profiles.

Do you drive like a cautious bear? Think you deserve to pay less? Great! You get fair insurance. Do you drive like a maniac? Don’t worry, you also get fair insurance.

You do not pay a fixed fee—instead, you only get charged when the car is being used. Pay As You Go, always.

In short, here’s how Insure Ride makes life better:

  • Cautious drivers get cheaper insurance
  • We help people improve their driving patterns through gamification
  • Drivers are encouraged to drive more eco-friendly
  • People are charged based on how much you actually use the car—pay as you go

Stack

The Bosch XDK is a programmable sensor device. We use it as an IoT device to record metrics like acceleration, speed, breaking, and any motion. Every 100 milliseconds, the beacon sends the recorded data to a Python server as UDP packets. Additionally, we also collect data from a dashboard-mounted webcam (e.g. the driver’s identity, mood, tiredness, etc.) and a GPS device (e.g. location). In the future, more data could be collected from the car itself: how and when turns are properly signalled, whether the car has had any service problems, etc.

The Python server receives all that data and stores it in memory until the drive is considered complete. When the drive ends, we analyse the sensor data, perform conversions, and calculate the impact on the driver’s insurance score. The metrics are then send via REST to our backend, ready to be stored permanently in the blockchain.

The backend is written in Go. When data from a drive is pushed to the backend, it is stored as a smart contract in the Ethereum blockchain. The smart contract contains all the business logic: 1) the user pre-pays some money in the contract; 2) when the contract gets data from a new drive, the insurance cost for that drive is calculated and payed directly to the insurance company.

The frontend accesses data stored in the blockchain through the backend. Then it presents a useful dashboard to the driver. Besides showing a list of all their drives, it also gives recommendations on how the driver can improve their driving style (e.g. tips on changing gears, advice on using the car’s functions better, taking alternative routes to avoid delays). The driver is motivated to drive safer and more eco-friendly by constantly monitoring their score. We also include rankings like “You're in top 10 eco-drivers in St. Gallen”.

APIs

Sensor module:

  • XDK from Bosch

Python server:

  • OpenCV, Dlib, and github.com/ageitgey/face_recognition (for facial recognition, identification and classification)
  • http://bokeh.pydata.org (to plot sensor metrics in real-time)

Backend:

Frontend:

  • Google Maps (to display the driver’s trips on the map)
  • Skype API (for the call widget, used to e.g. call the insurance company)
Share this project:

Updates