An artificial neural network-based integrated tilt control system for narrow electric three-wheelers

Abstract

The active tilt-controlled narrow electric three-wheelers offer an encouraging solution to pressing environmental and urban challenges, including air and noise pollution, parking shortage, and traffic congestion while improving their rollover stability. The previous research on active tilt control established a desired tilt angle neglecting track width which helps to bear unbalanced lateral acceleration that leads to rollover threat. This article proposes a dynamic rollover index based optimal tilting approach, which tilts the vehicle only when rollover index reaches beyond the desired optimal threshold values. Furthermore, the present work introduces an optimal active tilt control strategy, integrating tilt control system with an active steering system. The proposed integrated tilt control system optimizes an active steering gain with the objective to minimise the deviation in the desired trajectory. The proposed control system ensures the manoeuvrability and rollover stability at higher speed up to 100 km/hr with 22° steering angle even when vehicle reaches its tilt limit. Matlab/Simulink has been used in this study to thoroughly investigate the vehicle dynamics model and proposed integrated active tilt control system. The proposed control strategy is found to be more effective than the traditional steering direct tilt control approach. An artificial neural network that avoids the control system’s complexity and provides a faster response was also studied. The investigation showcases the efficacy of augmenting the dataset length and increasing neurons in the network’s hidden layers. The performance metrics of the trained model tested on the external data imply that the model holds significant potential for practical applications.