Usage-based insurance (UBI) is currently surfacing
both in research and within insurance companies.
There are a lack of actual described UBI products, and
those that exist are experimental and limited to small
customer segments. Insurance companies are showing
a clear interest in entering the market, but UBI as a
product is complex, and little research exists when it
comes to completely implemented products. In this paper,
the authors describe the design, implementation
and experimentation on Drive-LaB, a fully functional
UBI platform. Drive-LaB lets users collect spatiotemporal
data with their smartphone. The system uses
this data to identify driving style and environmental
context, to allow risk assessment associated with car insurance.
Drive-LaB is supported by a complex backend
system featuring an advanced data warehouse and computational
logic to identify driver styles. It also offers
an easy-to-use Android application frontend, allowing
users to log trips and see detailed statistics on completed
trips. Drive-LaB has been used for experiments,
collecting more than 13.000 kilometers worth of data in
roughly one month. This data has been used to validate
the platform and display how the system performs in a
realistic setting.