The Design And Development Of Algorithm Optimization Of High-performance Embedded NC System

In recent years,with the rapid development of electronic technology,The computingPerformance of Embedded Processor has been continuously improving,as aresult,Embedded technology has penetrated into all kinds of industries,the advantages ofthe Embedded processor,such as high performance,low power,low cost and so on,provide a solution for the construction of economic CNC system,so the embedded CNCsystem has received widespread attention in the industry and academia,however,Theembedded CNC system is different from the traditional CNC system based on PCplatform in the form of processor architecture, some algorithms based on the traditionalCNC system is difficult to be applied or low efficiency in the embedded environment,ina word,it is hard to meet the demand of the CNC system for hard real-time,meantime,itreduces the processing efficiency and quality of workpiece.In order to solve the problems mentioned above, this paper belongs to the majorproject of the national science and technology, in this paper,the main research works areas follows:1.Study the architecture of The DSP in TI’s OMAP3530,for the architecture of theDSP, this paper do some research on the algorithm optimization methods,at last,we candivide the optimization methods into two categories, general algorithm optimizationmethods and some optimization methods based on the internal structure of DSP,sincethe experimental platform is fixed-point DSP,the computing for float-point is veryslow,this paper studies the optimization principles of IQMath which is developed toconvert float-point to fixed-point by TI.2.Study several key problems of speed look-ahead controlling algorithm,such asestablishing mathematical model of path segments,speed limit equation of inflectionpoint,the real-time planning of speed and so on,accord to the method for determiningthe number of look-ahead segments,design and implement two adaptive look-aheadalgorithms,and make a comparison of performance about this two algorithms.3.At last, based on the optimization methods mentioned above,carry out theoptimization on the speed reduced to zero look-ahead algorithm,,the performanceevaluation and error analysis of un-optimized and optimized algorithm are made.