Optimizing vehicle Front-End structure for e-bike rider Safety: An advanced Multi-Objective approach using injury prediction models

Abstract

A multi-objective optimization method based on an injury prediction model is proposed to address the increasingly prominent safety issues for e-bike riders in Chinese road traffic. This method aims to enhance the protective effect of vehicle front-end for e-bike riders by encompassing a broader range of test scenarios. Initially, large-scale rider injury response data were collected using automated Madymo simulations. A machine learning model was then trained to accurately predict the risk of rider injury under varied crash conditions. Subsequently, this model was integrated into a multi-objective optimization framework, combined with multi-criteria decision analysis, to effectively evaluate and rank various design alternatives on the Pareto frontier. This process entailed a comparative analysis of the design in a baseline scenario before and after optimization, focusing on both kinematic and injury responses of riders. Through detailed injury mechanism analysis, key design variables such as the height of the hood front and the width of the bumper were identified. This led to the proposal of specific optimization strategies for these structural parameters. The results from this study demonstrate that the proposed optimization method not only guides the design process accurately and efficiently but also balances the injury risks across different body parts. This approach significantly reduces the injury risk for riders in car-to-e-bike collisions and provides actionable insights for vehicle design enhancements.