Detection of lifted axles on heavy vehicles and preservation of road infrastructures (Seto Project)

Abstract Detecting the use of lifted axles on heavy goods vehicles (HGV) is a major challenge to enforce weight regulation and to assess road infrastructure damages under moving loads. If some axles are lifted on HGVs, the gross vehicle mass is concentrated on less axles, which increases significantly the aggressiveness of these HGVs. The paper will first present a new methodology to detect lifted axles by algorithms using only WIM (Weigh In Motion) data. We successfully apply two supervised classification methods, namely logistic regression and Random Forest. In particular, the random forest method enables more than 90% of the axles lifted to be correctly identified. Identifying the lifted axles of HGVs will open new applications. Among the applications, the deployment of direct enforcement by WIM systems will benefit of the lifted axle identification, while the maximum permitted weight of a HGV depends on its number of axles. If some axles are lifted the vehicle should not be fully loaded. Balancing evenly the gross vehicle mass on all axles may significantly reduce the stresses applied to pavement layers. This article will compare the aggressiveness of heavy goods vehicles of different weights on several types of road, with and without raised axles. Some bridge structures are also sensitive to axle loads, and may also benefit of a better monitoring and control of lifted axles.