On Active Queue Management Algorithms Based On Sliding Mode Variable Structure

With the advent of the era of big data,the amount of network data continues to increase,and data loss and network congestion caused by TCP routing transmission queue length control are becoming more and more serious,which reduces the quality of service.The active queue management algorithm can effectively deal with the TCP network congestion problem,so it has become one of the most widely researched congestion solutions in the TCP network field.This paper analyzes the TCP network from the perspective of control theory by using Lyapunov stability theory and other knowledge,and integrates the sliding mode control theory into the active queue management algorithm research,and the designed algorithm was simulated based on the simulation software Matlab.The main work of this paper is as follows:Firstly,we propose an active queue management algorithm based on adaptive global sliding mode for the problem of TCP connection number and round-trip delay uncertainty in TCP dynamic network and non-responsive UDP(User Datagram Protocol,UDP)data stream perturbation.Considering that the robustness of sliding mode variable structure control arrival stage is poor,we design the global sliding mode controller to eliminate the arrival stage in the traditional sliding mode control,which can ensure the global robustness of system state.Because the upper bound of the system is difficult to obtain,the influence of uncertain parameters and external disturbances are integrated into a total uncertainty item,and the adaptive method is used to approximate the total uncertainty of the system online.The symbol function or saturation function is not used in the controller,which can effectively reduce the chattering existing in the traditional sliding mode control.The simulation results show that the algorithm designed in this chapter can achieve system state target tracking,and the response of the whole control process state is smooth,no obvious oscillation,and has good robustness and stability.Secondly,we propose an active queue management algorithm based on adaptive super-twisting sliding mode variable structure control in view of the slow convergence speed and the existence of chattering of traditional sliding mode.In order to weaken the sliding mode control chattering while maintaining the invariance of the sliding mode controller,the super-twisting sliding mode is used as controller.Considering the external disturbance of the non-responsive UDP data stream in the TCP network,the adaptive term is used to approximate the external disturbance,which reduces the interference of the external disturbance to the system.Combining adaptive control with super-twisting sliding mode control,the designed adaptive super-twisting sliding mode controller can effectively attenuate chattering,and we have proved its finite time convergence by Lyapunov function.The simulation results show that the proposed algorithm has good robustness and fast convergence performance,and greatly weakens the chattering phenomenon,keeps the queue length small oscillation and makes the instantaneous queue length quickly stabilize to the vicinity of the expected value.Finally,the main work done of the paper is summarized,and the further research directions of active queue management algorithms are put forward.