Pavement and road health monitoring using random forest technique

Abstract

Detection of road anomalies is crucial in order to prevent road accidents. Even with the advancement in technology, road accidents are still happening in our country. This is mainly because of the difficulty in detecting road anomalies, which involves high inspection and monitoring costs. Traditionally, electro-magnetic methods like RADAR, LASAR, GPR and visual inspections are used for road health monitoring which is costly, cumbersome and often not reliable. Moreover, the visual inspection of a road requires a lot of time and labour work, electro-magnetic inspections require high skilled labours and expensive equipment, hence it is not possible to implement it on a large scale for all road-network. To overcome the difficulty in detecting road anomalies, recently vibration monitoring devices like accelerometers, gyroscopes, and motion sensors are used in this field due to its low-cost, ease of use, can monitor irrespective of surrounding and seasonal conditions, time saving. machine learning (ML) techniques have been used in this study to detect anomalies from accelerometer readings.