Hyperformance: Advanced PBS Performance Prediction

Abstract

The ever increasing global freight task brings with it a number of challenges for road freight transportation. The combination of high-capacity vehicles and Performance-based Standards (PBS) is proving to be a viable and sustainable option in combatting some of the challenges, particularly environmental and safety. However, with the increase in the number of PBS initiatives as well as vehicles globally, there is an ever increasing demand on vehicle designers, PBS assessors and regulators. In this paper, we present an updated methodology for the development of PBS performance prediction or calculation tools: so-called “Hyperformance” models. The methodology we propose uses a probabilistic machine learning technique called Gaussian Processes (GP), which provides both a prediction of vehicle performance, as well as an indication of the accuracy of the model for each prediction. This approach is ideally suited to efficient development of Hyperformance models for new vehicle configurations. This has value in that they can be used to define new pro-forma or blueprint designs, as well as being used for optimisation of vehicle parameters for a given application. We also present a case-study in which we develop GP prediction models for a PBS B-double combination.