Manage how you drive
The common purpose for us and insurers is to reduce the number and scale of road accidents. Classic UBI models make it possible to analyse the driving style of individuals and to adjust insurance premiums to actual risk. The module called Manage How You Drive allows developing correct habits of drivers, by providing them immediate feedback on what they should change. The solution applies an automatic module of personalised messages in mobile applications and gamification tools, such as rankings or competitions.
An accident happens. What were its causes? Did the driver comply with relevant speed limits? What forces was the vehicle subject to when the incident happened? Was the accident preceded by sudden braking? What was the vehicle’s trajectory during the crash? How do other drivers behave in that section of the road? This and other questions can be answered by analysing specific circumstances of dangerous road incidents in the crash reconstruction module. The data takes map context into account.
How do drivers’ profiles translate into their loss ratio? Scoring provides an answer to this question. Traditional insurers estimate premiums on the basis of demographic data and the driver’s official loss history. By using data on drivers’ behavioural profiles and machine learning, the system creates a correlation between driving style parameters and potential loss ratio. Thus, insurance premiums calculated with the use of such data refer to an actual risk.
DRIVER CONCENTRATION ANALYSIS
Some of the most common causes of accidents and collisions on Polish roads are forcing the right of way, failure to adjust the speed to traffic conditions, failure to keep safe distance or drink-driving. However, it is estimated that using the phone when driving is an indirect cause of every fourth accident in Poland! The phone not only forces the driver to drive with one hand only, but it also affects thinking, prolonging response time.When we analyse a person’s driving style, we also take into account whether the phone was in the driver's hand or in a phone holder.
DRIVING STYLE ANALYSIS BASED ON ACTUAL SPEED LIMITS
“Data on braking and accelerating obtained from accelerometers is enough to assess a driver’s driving style” – IT’S A MYTH! Driver’s behaviour within actual speed limits is of key significance. Why? Imagine a situation when a driver is leaving the motorway at the speed of 100km/h and smoothly drives into a built-up area. He doesn’t suddenly brake or accelerate. On the basis of this information, it might seem that we are dealing with a perfect driver. If we consider the context, however, we know that the driver’s behaviour is very dangerous. Thus, it is crucial to analyse both of these aspects together. And that’s what we do!
What is driver profiling? It is based on an analysis of behaviours of individual drivers in specific road situations. It is called behavioural profiling. For instance, we check if a driver drives according to the rules or how they brake and accelerate. We analyse if a driver actively uses their phone when driving, what driving style they have in a city and outside of it, what category roads they use, how much time they spend behind the wheel and what distances they cover. Such detailed analyses are possible due to the access to digital maps taking actual speed limits into account. We aggregate and visualise data concerning the trip and the driver's personal driving style (having obtained the driver's prior consent) in our system.
To detect a road incident correctly, we use information from devices such as beacons, OBDII, 12V or smartphones. We have been building our own incident detection system for several years. To ensure highest quality in the so-called claims-focused telematics, we organise laboratory tests and numerous crash tests in professional centres in Poland and Europe, such as the Centre for Experimentation and Road Safety Mapfre in Spain. Another issue related to crash detection is the solution’s integration with communication systems that ensure contact with an accident participant and call for help.
ASSESSMENT OF POST-ACCIDENT DAMAGE (BETA)
We have been carrying out R&D works concerning the correlation between the accident’s geospatial context, damages to the vehicle and the cost of repair. As part of these works, we analyse, among others, the scope of damage in specific vehicle models on the basis of an acceleration graph, energy absorption and actual repair costs.
UNIVERSAL IOT PLATFORM
A universal platform allowing analyses with the use of advanced algorithms covering immense sets of location- and behaviour-related data, regardless of its source, is the central element of our system. We analyse data from smartphones, OBDII devices, beacons or 12V. The platform is scalable and module-based.