Design Of Customized Controller Platform

In the industrial control field, customize controller has good value and marketprospects [1]. To develop the tailor-made controller for the industrial customers, aplatform with high performance, hardware adaptability and software compatibility isneeded. In this project, a customize controller development platform has been madedsuccessfully, including the mechanical structure, hardware design, operating systemsporting, performance testing, devices driver templates design, control software designand genetic algorithm implementation, supplying a powerful architecture for theresearch and develop of specified industrial controller series.To ensure the openness, adaptability and real-time performance, the controllerplatform chooses ARM9 processor and adopts hybrid operationg system as Linuxcombined with Xenomai.Hardware system is modular in design and convenient for expansion and maintenance.Now the design and testing of basic modules has finished, including the mainprocessor board, backplane, analog input and output board as well as CAN businterface board.The hybrid operating system consists of Linux and Xenomai. Xenomai is apreemptive real-time nanokernel built under the bottom of Linux kernel. Taking overall interrupt sources, Xenomai has increase the real-time performance of Linux. Forinstance, the average delay of processes switching was 18 microseconds whenAT91RM9200 running at 180Mhz. Xenomai also provides APIs for a wide range ofmainstream real-time operating system, including VxWorks, pSOS +, VRTX anduITRON. Application procedures based on such thease RTOS formerly can be portedto the controller platform seamlessly.As Xenomai is implemented below Linux, it is good at hardware accessing.Combining the original Linux driver and the mapping mechanism of Xenomai, aninnovative method of hardware register direct accessing was created. This methodhighly improved the ability of application procedures to control the hardware. Relying on the powerful ARM9 processor computing feature and the support ofexisting library for complex computing, the genetic algorithm was implemented onthe controller platform and successfully applied to PID parameter setting. It had beenproved that the dual–core operating system can handle the real-time control task andcomplex algorithm computing simultaneously.