Active suspension for all-terrain vehicle with intelligent control using artificial neural networks

• Published in: Journal of Mechanical Engineering and Sciences pp. 9883 - 9897
• Main Authors: Hamza, Anis, Dridi, Issam, Bousnina, Kamel, Ben Yahia, Noureddine
• Format: Journal Article
• Language: English
• Published: 30-03-2024
• ISSN: 2289-4659; 2231-8380
• DOI: 10.15282/jmes.18.1.2024.7.0782

The automotive industry focuses on developing advanced protection and stability control systems, particularly for suspension and steering, to enhance vehicle comfort, luxury, and safety. This research presents an intelligent controller for all-terrain vehicle (ATV) suspension systems based on Artificial Neural Network (ANN) technology. The controller leverages ANN capabilities to optimize system performance. MATLAB simulations were conducted to evaluate its effectiveness under various disturbances. A comparative analysis compared the ANN regulator, classic ANFIS regulator, and passive performance in different disturbance scenarios. The simulation results demonstrate exceptional performance of the ANN-based controller in displacement reduction, speed, acceleration, and robustness. The controller effectively mitigates disturbances, enhancing overall suspension system performance. These findings highlight the advantages of employing ANN technology in ATV suspensions. This research contributes to intelligent control systems advancement in the automotive industry, specifically in ATV suspensions. The demonstrated improvements have the potential to enhance passenger comfort, vehicle stability, and safety across terrains. By implementing ANN-based controllers, automotive manufacturers can optimize suspension systems, leading to improved vehicle performance. Several indicators, including RMSE, MRE, and R2, were utilized to test and validate the models. The R2 values for the three quality parameters ranged from 0.989 to 0.999, indicating a high level of consistency in the predictions made by the ANN, a "5-12-1" structure is employed. The results of this study add to the expanding body of knowledge endorsing the efficacy of ANNs in simulating and optimizing quarter-vehicle dynamics.