Session: ASME Undergraduate Student Design Expo

Paper Number: 158915

Development of a Predictive Model for Material Selection in Motorcyclist Impact Protector Design Using Gaussian Process Regression 

Motorcycling poses significant safety risks, accounting for a disproportionately high percentage of traffic fatalities due to the lack of protective structures. Advanced energy-absorbing materials, particularly polymeric foams like viscoelastic and polyurethane-based variants, play a critical role in mitigating impact-related injuries. These materials dissipate impact energy, effectively safeguarding a rider’s spine, torso, and limbs. However, the traditional process of testing numerous materials manually, with varying parameters such as density and thickness, is not only time-consuming but also resource-intensive. To address these challenges, this study introduces a predictive model using Gaussian Process Regression (GPR) to optimize the selection of impact-resistant materials for motorcyclist back protectors.A total of 17 foam and rubber materials, including high-performance options such as PORON XRD, Bay Rubber, and Nitrile Butadiene Rubber (NBR), were subjected to rigorous impact testing following the EN 1621-2:2003 standard. Data on density, elasticity, tensile stress, thickness, and impact energy were collected for each material under controlled testing conditions. The GPR model was trained using data from 17 materials and tested on two additional materials, Latex and SH38 Arti. The model’s predictions were evaluated using metrics such as Mean Squared Error (MSE) and Mean Absolute Percentage Error (MAPE). For Latex, the model demonstrated excellent predictive accuracy, achieving an MSE of 132,623.3 and a MAPE of 3.88%. However, predictions for SH38 Arti exhibited higher errors, with an MSE of 776,117.6 and a MAPE of 8.79%, reflecting the complexities associated with non-linear material properties.The results underscore the GPR model's potential to significantly reduce reliance on exhaustive physical testing. By identifying critical parameters such as thickness and density, the model streamlines the material selection process while ensuring optimal impact resistance. Beyond its applications in motorcyclist safety gear, this methodology has broader implications for industries such as automotive, aerospace, and consumer products, where high-performance, impact-resistant materials are essential for safety and durability.Future research will focus on improving the robustness of the model by expanding the dataset to include a wider variety of materials and environmental conditions. Additional features, such as temperature sensitivity, multi-impact performance, and long-term durability, will be integrated into the model to enhance its predictive capabilities. Furthermore, exploring advanced kernel functions and hybrid machine learning approaches will address the complexities of non-linear material behaviors. This study demonstrates how data-driven methodologies can revolutionize material selection processes, leading to the development of safer, more efficient protective gear across diverse industries worldwide.

Presenting Author: Shoaib Ahmed Rajshahi University of Engineering and Technology

Presenting Author Biography: Shoaib Ahmed
Mechanical Engineer

Shoaib Ahmed is a skilled mechanical engineer with a deep interest in design optimization, energy efficiency, and material science. He holds a Bachelor’s degree in Mechanical Engineering from the prestigious Rajshahi University of Engineering & Technology (RUET).

With experience in thermal systems analysis, product design, and material durability testing, Shoaib’s work aims to bridge innovation with practical engineering solutions. He has contributed to projects involving automated systems for industrial applications and advanced material characterization.

Currently, Shoaib is focused on enhancing sustainable manufacturing processes, combining his technical expertise with a passion for environmental responsibility. His dedication to engineering excellence drives his commitment to solving complex mechanical challenges and advancing the field

Authors:

Hasibur Rahman Islamic University of Technology
Shoaib Ahmed Rajshahi University of Engineering and Technology