DATA-CENTRIC, TECHNOLOGY-ENABLED ALGORITHMIC FRAMEWORK TO MEASURE AND CLASSIFY DAMAGED ROADS FOR MAINTENANCE IN THE INDIAN CONTEXT (Paper 818)

It is a day-to-day experience to see roads and highways in poorly maintained conditions. One can observe its manifestations while driving, in the form of shocks and jerks. The biggest challenge to road quality assessment is collection of relevant data and its standardized interpretation. 
India has a vast road network (6 million kms). A conservative assumption of 10% poor quality roads, implies data collection over 6 lac kms of roads for the Road Classification & Assessment program. Manual collection of such data can be prohibitively expensive, even if limited to the worst maintained roads. But more importantly, the focus needs to be on defining Road Quality.
Road quality is characterized by usage-led visual defects; these include rough patches, potholes, damaged/uneven speed humps, etc. Road design and engineering also impact road quality, based on design attributes such as gradients, curves, banking angle, etc. 
While qualitative interpretations of road quality are widespread, a standard quantitative definition is elusive; also elusive is the relative benchmarking of the parameters impacting road quality metric. The need-of-the-hour is an inexpensive, yet replicable data collection and assessment methodology to measure road quality. Today, a technology solution based on advanced Inertial Measurement sensors, is the best way forward.
Advanced research is currently underway to develop high-reliability algorithms for interpreting severity of road degradations. At the core of these algorithms is the comparative benchmarking of different manifestations of bad roads (potholes, uneven humps, bumps, etc.). Abundant empirical data is now available from sensor devices, and road classification algorithms are being tested.