Adaptive Neural Network Control For Vehicle Active Suspension Systems

In recent two decades,car or vehicle is universal and convenient transport in the life of the average ordinary people.Meanwhile,the vehicle plays an is increasingly important in people's trip.The ride comfort is an important indicator for evaluating the performance of the vehicle.The performance of the vehicle is decided on its suspension systems.The damping function of passive or semi-active suspension systems depends entirely upon their physical structures.Therefore,their damping functions are limited to some extent.When encountered the road bump,the force produced by the controller keeps down the vibration of the car-body.Due to the equipped controller,active suspension systems bring the high damping function and the design control problem becomes a hot area of research for active suspension systems.If the vertical displacements of suspension systems are very large,it results in the damage of suspension structure and causes even the safety accident.The constraint problem of active suspension systems is a difficult problem.When encountered the road bump,the stability of active suspension systems is achieved in a finite time,which enhances the ride comfort.Therefore,how to achieve stability in as little time as possible is also a hot problem.Based on above analysis,this thesis mainly investigates the following three aspects:(1)An adaptive neural network(NN)control method is given for active suspension systems with input saturation.Under the situation of without saturation,the first step is to design an adaptive NN controller based on Backstepping technique and the function approximation capabilities of NNs.In the first step on the basis of knowledge,the second step is to propose an adaptive anti-saturation NN control controller by using anti-saturation theory,which solves the saturation problem of the controller.By using the Lyapunov stability analysis and suspension performance analysis,the stability of active suspension systems is proved.The effectiveness of designed scheme is demonstrated by the simulation of active suspension systems.(2)An adaptive NN control approach with displacement constraint is presented for the active suspension systems with human-body and seat model.The uncertain terms of the systems are approximated by NNs.In order to avoid the problem of explosion of complexity,the first-order filter is employed in each step.Then by constructing the asymmetrical Barrier Lyapunov functions,the displacements of human-body and active suspension systems are constrained in the set way below the safe values,which ensures the safety and enhances the ride comfort.Using the Lyapunov function stability analysis and the suspension performance analysis,it can be concluded that active suspension systems are stability.The feasibility of designed adaptive NN scheme is demonstrated by the simulation of active suspension systems in road continuous input.(3)An adaptive NN control scheme used finite-time theory and constraint theory is demonstrated for 3-DOF active suspension systems.The vertical displacements of suspensionsystems are constrained by building the Barrier Lyapunov functions.When encountered the road bump,the finite-time theory is used and its goal is to keep the stabilization of active suspension systems in the shortest possible time.The combination of Barrier Lyapunov functions and finite-time theory is used to design an adaptive finite-time controller with displacement constraints,which resolves the problem of the safety and finite-time control of suspension systems.The stability of 3-DOF active suspension systems is studied based on the analysis of Lyapunov stability and suspension system performance.In road continuous conditions,the simulation results of 3-DOF active suspension systems show that the given finite-time scheme is feasible.