Research On Control Of Vehicle Anti-Rollover Based On Semi-Active Suspension

In recent years, with the rapid development of automobile industry and road traffic,vehicle rollover accident has become an important safety issue widely attention by people. When vehicle travels at high speed and emergency steering, In a relatively short period of time easy to rollover, so the vehicle rollover warning and a vehicle control is particularly important. In this paper, using neural networks control strategy based on the semi-active suspension, to research for vehicle anti-rollover, Improvements on vehicle handling stability point has certain theoretical and practical significance. The main work is as follows:First of all, based on three degrees of freedom vehicle model, establish a seven degrees of freedom of the semi-active suspension vehicle rollover model, and through the vehicle dynamics simulation software Carsim to simulation experiments in different input conditions, To verify the rationality of the seven degrees of freedom of the rollover model established.Secondly, by comparing the common vehicle warning system, use dynamic load transfer rate(LTR) for vehicle rollover threshold, design a suitable vehicle rollover model of rollover warning algorithm. In the Matlab simulation analysis verify the validity and reliability of algorithm in the Matlab simulation analysis.Then, neural network control based on semi-active suspension. From aspects of the training time length and training accuracy to compare the advantages and disadvantages of BP neural network and ELM. Generated by the Matlab programs a Simulink controller module, and to establish a good seven degrees of freedom vehicle model simulation analysis,it can be seen that the neural network control based on semi-active suspension can effectively improve vehicle rollover prevention ability.Finally, the Carsim established with the parameters of vehicle model imported in Matlab, and the establishment of a correspondence, according to the design of the control strategy, based on the semi-active suspension control of the vehicle model for joint simulation. To analyze the simulation results, the results show that the neural network control and skyhook damping switch control the correct and effective.