Current vehicles don’t evaluate the number of occupants in a crash. Automatic Crash Notification systems (ACN) (i.e. OnStar, eCall) require crash dataset. This dataset includes information to support a crash severity evaluation (i.e. Advanced Automatic Crash Notification & Injury Severity Prediction). AACN & ISP support the emergency response by triggering rapidly the proper medical response.

However, to this date, one valuable piece of information is still missing:​ the occupancy. The occupancy number would enable an emergency response with the proper magnitude. In the post-crash notification systems, a proposed feature is named Occupant Count Prediction or OCP.

With OCP, PSAPs (Public Service Answering Points) would obtain a predicted occupant count allowing them to tailor the emergency response magnitude to the actual crash. Ambulances or ERTs would be dispatched when/where needed, in the correct numbers, thereby increasing efficiency.

This paper identifies several formulas to quantify OCP. It is also discussed an array of sensors which could be used as input into those OCP formulas. The analysis includes the accuracy of several sensor combinations and demonstrates a relationship between the number of sensors and the prediction accuracy.

GIDAS is used for validating, or predicting the accuracy of, an Occupant Count Prediction algorithm. For most OCP formulas, the accuracy is above 94%. The accuracy gets upward of 99% in vehicle fully equipped with sensors. However with the current ubiquitous equipment in today’s typical vehicle we can predict with 97.6% accuracy the occupant count OCP.