Design And Implementation Of General Signal Processing Platform Software Based On Domestic DSP

With the rapid development of modern signal processing algorithms and the increasing complexity of application scenarios,the complexity of signal processing algorithms is gradually increasing and the iteration speed is getting faster.The traditional signal processing platform hardware and software customization design are difficult to adapt the rapid iteration of algorithms,and many emerging algorithms need a long development cycle to be applied on engineering,which hinders the transformation speed from experimental algorithms to engineering algorithms;Meanwhile,the complex international situation makes the signal processing hardware platform face the potential threat of foreign technology blockade.Therefore,it is important to study a homemade universal signal processing platform software that can meet the requirements of various signal processing algorithms and can be rapidly developed and deployed.Thesis designs and implements a set of generalized signal processing platform software based on the domestic FT-M6678 DSP,aiming to provide a generalized platform software that can quickly implement and deploy signal processing algorithms,reduce the difficulty of algorithm engineering development,and accelerate the speed of algorithm engineering implementation.In thesis,after fully studying the current status of the software and hardware platforms of the signal processing platform,we briefly introduce the software and hardware foundation of the universal signal processing platform,clarify the generalized implementation of the signal processing platform,and design and implement the platform software engineering and application algorithm engineering.Through platform software engineering,each core is virtualized into equivalent computational nodes to shield the differences between different cores;through application algorithm engineering templates,the rapid development of various algorithms is realized;through dynamic loading and unloading of application algorithm engineering by platform software engineering,the rapid deployment and replacement of each algorithm is realized.Compared with the traditional signal processing platform,dynamic loading and unloading improve the generalization of the hardware platform,and template development greatly improves the engineering development efficiency of new algorithms.At the same time,some calculation library functions are optimized,which provides an efficient basic operation library for algorithm transplantation.Finally,in order to verify the versatility and correctness of the platform software,the general signal processing platform software developed in thesis is used to implement and deploy the satellite spread spectrum communication and tracking algorithm process,which includes the digital beam formation algorithm,direct spread spectrum sequence despreading algorithm,and single pulse goniometry algorithm to realize the despreading and goniometry in low signal-to-noise ratio environment.The correctness of the calculation results is verified by using the intermediate results reporting function to plot and store the numbers.The generalized signal processing platform software is designed with generalization and localization in mind.The generalized design reduces the difficulty of algorithm development and broadens the scope of application of the signal processing platform software.By adopting the local hardware and software environment,the software of the general-purpose signal processing platform is ensured to be independent and controllable.