Search icon

Estimating vehicle collision damage using motion data

Supplied by Identity Innovation Ltd. T/A Kabzy

Kabzy is an Internet of Things company focused on providing telematics solutions to the transport sector. Founded in 2013 it now provides fleet monitoring solutions to some of Ireland’s largest commercial fleets, including rail, plant machinery and logistics sector. We are currently expanding to the UK.

Profitability in Ireland’s insurance market has rapidly deteriorated in recent years, leading to premium increases of 38% across the board. A major contributing factor to this is thought to be losses due to undetected fraudulent claims.

Kabzy have set out to address this problem by developing a software tool that can accurately reconstruct vehicle collisions from devices in cars that use sensors such as GPS, accelerometers, magnetometers, and digital gyroscopes to track car movement. Kabzy’s tool can then provide visualisations of incidents and other useful details to claims investigators, helping insurers to make faster, more informed decisions on claims.

In conjunction with the Dublin Fire Brigade, Kabzy have collected data from six crash tests, and used their tool to reconstruct the crashes. While this has provided valuable validation of their approach, they wish to obtain more information from the available data.

Questions for the study group to address:

Given the make, model, year, mass and dimensions of the vehicle as well as a time series of sensor data associated with the crash:

  1. Can we determine the zone of impact of the crash? (That is, where on the body of the vehicle did the impact occur?)
  2. Can we determine the severity of the impact? (Insurers have defined tiers of severity; is it possible to predict the likely tier of severity based on the crash data?)

Based on this information, it would be possible to calculate estimates of the cost of the damage to the vehicle, which could then be given to the claims handler to assist them in processing the claim.

Additionally, it would assist with highlighting cases where the damage to the vehicle is not consistent with the crash as reconstructed from the sensor data.

Some related videos include:


Multi View Video 

Raw Data View Video

Sfi logo