FlexRay Bus System Control Based On Neural Network

FlexRay data transmission bus is a bus technology for high-speed,deterministic and fault-tolerant vehicles.With the development of the times,people’s requirements for automobile safety system,comfort system,information entertainment system and other performance are increasing.More and more control tasks and information are transmitted through automobile bus.However,the network bandwidth resources in FlexRay bus are limited.When the length of message data is different,it will lead to unreasonable use of network bandwidth resources,resulting in waste of bandwidth resources.According to the research status of FlexRay bus and the development trend of automobile in the future,this paper presents a theoretical model of FlexRay bus model reference adaptive control system based on neural network.In order to improve the bandwidth utilization of FlexRay network,a bus control scheme is proposed.Combining the complexity and non-linearity of FlexRay vehicular network and the ability of self-learning,self-adaptive and global approximation of the neural network,the theoretical model of the control system is established.The control system structure designed in this paper consists of four parts.Firstly,the mathematical model of FlexRay controlled object is established by using the bandwidth relationship of FlexRay bus static segment network,and its bandwidth utilization output is taken as the system output.Secondly,the reference model of the control system structure is established by using the optimized static bandwidth relationship,and its maximum bandwidth utilization output is taken as the expected output of the system.Finally,a neural network identifier and a neural network controller based on BP algorithm are designed.The functions of the neural network are as follows: 1.On-line identification of the FlexRay controlled object system model by neural network;2.Training of the controller neural network.The control principle adopted is that the model of FlexRay controlled object system is identified online by using neural network,and the controller is trained with samples,so that the output of FlexRay controlled object can track the expected output of the reference model by adjusting the parameters of the controller,so as to achieve the control of FlexRay bus.Finally,the designed control value of the neural network is imported into the Simulink model for simulation.The simulation results show that the control theory scheme designed in this paper is feasible,which achieves the desired control effect,and has good adaptability to the change of control object parameters.It provides a new basis for the further network and intelligent development of automobiles in the future.