RoadBotics uses a 5-level rating system. Our AI algorithm was trained by road experts loosely based on PASER, also known as the Wisconsin Method for visually assessing pavements.
Based on the type quantity and severeity of distresses on the road surface, our trained algorithm analyzes and assigns a score to every image collected at 10-ft intervals. Since planning is done at a road segment level, RoadBotics provides an aggregate score for every intersection-to-intersection road segment. This is based on the average of all the underlying ratings of the individual images.
The average intersection-to-intersection road segment scores are rounded out to 3 decimal places (ex. 2.534). This provides enough granularity when it comes to prioritizing roads for maintenance and planning activities. When necessary, it also allows us to convert the data to a different scale such as a 1-10 or 0-100 equivalent scale.
Several RoadBotics partners and customers have done a deep analysis of our rating system compared to their status quo and found the results to be very closely correlated.