A new detailed fuel and travel time estimation model for heavy trucks in forestry

Transportation and route planning require accurate information about fuel consumption and travel time. This is needed to provide correct input for decision support and analytic systems. The two typical approaches are either an empirical model, e.g. linear regression based on statistical analysis or a physical model describing theoretical energy principles. In this paper, we propose a model based on a detailed bucketing (discretization) approach, where road and operating conditions are grouped into discrete categories, applied to both arcs and nodes in a network representation. The bucketing for arcs is based on available road characteristics in a road database, including road class, speed limit, road surface, curviness, and hilliness. The bucketing for nodes is used to distinguish between different types of road crossings or when consecutive arcs change characteristics, e.g. when there is a change of speed limits. To populate the buckets, we use data collected over 12 months from 21 logging trucks equipped with a CAN-bus reader and a GNSS receiver for positioning. We also apply a smoothing method to populate buckets for which no or limited data was collected and to ensure logical conditions. We compare the proposed model with a set of empirical and physical models. The results show that the proposed model provides more accurate estimates of fuel consumption and travel time and is computationally efficient and suitable for integration into decision-support systems.