Adaptive Neural Network Control of Heavy Vehicle Air Suspension with Uncertainties

Owing to the rapid development of the automotive industry today, the performance requirements of automobiles are continually escalating, with the suspension being a primary vibration damping component, significantly affecting vehicle performances. The use of air suspension has become widespread in heavy vehicles due to its advantages like light weight, adjustable stiffness and height. However, due to the complex structure and non-linear characteristics of air suspensions, traditional control methods struggle to achieve optimal control effects, necessitating the proposal of more advanced adaptive non-linear control algorithms.

This paper proposes an adaptive neural network control scheme for heavy vehicle air suspension systems with uncertainties, by employing a radial basis function neural network to address the model uncertainty and external disturbances of the suspension system, and then, by integrating the backstepping method, achieves precise tracking control for vehicle height adjustment, and further enhances the overall performance of the heavy vehicle air suspension system through the design of time-varying constraint functions.

Finally, through simulation with real vehicle parameters, it is verified that the root mean square error of vehicle height tracking is \(2.069\; \times \;10^{ - 7} m\) $2.069\times {10}^{-7}m$ when tracking a sinusoidal signal, and \(1.744 \times 10^{ - 7} m\) $1.744\times {10}^{-7}m$ π when tracking a constant signal.

Compared to the fuzzy PID control method, there is a significant improvement in both accuracy and speed, demonstrating the feasibility and superiority of this approach.