The Real-time Research Of Control System Based On Windows

With the development of aviation and aerospace industry, semi-physicalsimulation technology has also developed rapidly. In semi-physical simulation, inaddition to the dynamic performance of simulation system, its real-time performancemust be taken into consideration. On the one hand, the control system must have highdynamic performance and control accuracy; however, since the traditional PID is notsuited to time-varying objects, and the neural network control parallel algorithmcontradicts with the computer’s serial approach, the control system’s real-timeperformance is hard to implement. In this way, the use of neuron PID algorithminstead of the traditional PID, leads to simpler calculation and easily-implementedreal-time performance. On the other hand, the operating system must be supported notonly by powerful hardware and better human-computer interface, but also by excellentreal-time performance. However, Window platform is based on time-sharingpreemptive multitasking system, which cannot finish the higher real-time tasks. In fact,as the extension of Windows operating system kernel system structure, RTX (Real-Time Extension) modified and expanded the Hardware Abstraction Layer (HAL),realized independent kernel-mode-drivers, and finally weak real-time shortcomings ofWindows system can be effectively resolved.This paper is based on an institute’s project about compound imaging infraredtarget simulator device. The author emphatically discusses controller design anddevelopment of real-time control software in the Window platform.Firstly, the author analyzes the functions and working principle of infraredinterference simulator’s control system, and then makes an overall design. Accordingto the actual needs, motion control cards, data acquisition cards, optical cards, drivers,encoders, etc. are used as hardware, while "Windows+RTX" is used as the softwareplatform, in order to solve the weak real-time shortcomings of Windows system.Secondly, according to the actual running process, a control system of threeclosed-loops (i.e. position loop, speed loop, current loop) is designed. To improve steady-state accuracy and dynamic performance of the control system further, neuronPID algorithm are used for the position loop. In this way, favorable effects can beachieved, such as better self-study ability of control system, faster adjustment speedand more easily-achieved real-time control.Finally, real-time control method on the Windows platform is studied, and howtheir precision influences the control system is also analyzed.The study shows that the using of “Windows+RTX” as software platform andneuron PID algorithm as control strategy, may realize Servo system the desired real-time performance and better robustness through the mutual coordination of hardwaresystem, Windows, and RTX in the Windows operating system.