迎向物聯網：支援語音辨識開發之可客製化數位訊號處理組譯器設計

近年來，物聯網為網路發展的趨勢之一，舉凡穿戴式裝置、智慧家庭至智慧城市等，均是現今熱門議題，隨之而來大量的多媒體訊號處理、影像及語音辨識等相關應用，目前的行動裝置及多媒體產品，通常透過數位訊號處理器(DSP)為特定應用，設計演算法處理即時數位訊號。 DSP晶片開發商會提供部分可客製化介面，給予系統應用開發商可客製化其DSP之指令集的彈性，使得演算法達到低功耗及高效能之目的，但同時也產生出許多的軟體開發上的問題，由於不同的DSP晶片會對應不同的組譯器及模擬器，因此DSP開發時通常會採用晶片開發商所授權之整合開發環境(IDE)，對於系統應用開發商或是中小型開發者而言，IDE對於客製化DSP指令集的支援是極為重要，而既有的IDE通常為授權且封閉的環境，無法對於其組譯器及模擬器進行修改，無法達到驗證及模擬客製化DSP指令集，對於軟體開發上相當不利。因此在這篇論文中我們提出了一彈性組譯器架構，來達到彈性擴充客製化指令集和模擬之目的，在客製化DSP指令集的軟體開發上，提供快速驗證及模擬的特性，除此之外，我們同時實作ADSP-218x DSP完整指令集以及DSP預處理器，作為驗證系統的正確性及完整性。

關鍵字

物聯網；語音辨識；數位訊號處理；特殊應用指令集；組合語言；組譯器；整合型開發環境

並列摘要

Recently, Internet of Things is one of the technology trends in the world. Nowadays, wearable devices, smart home, smart city, etc., is hot issue for application and research. Consequently, multimedia applications get more popular like video analysis, speech recognition, etc., These are used to using Digital Signal Processor (DSP) to design the algorithm for specific application, which processes the real-time digital signals. The DSP chip vendor provides the interfaces of customized for application development companies, which can specify their own instruction for specific algorithm to optimize the performance and power consumption. Although it brings numerous benefits, it introduces an additional software development problem. Most of the DSP development environment is always using the DSP vendor’s Integrated Development Environment (IDE). The licensed and closed platform IDE becomes a drawback in software development, which the IDE cannot support the application specific instruction-set, especially in the coding validation and simulation. In this paper, we propose a flexible assembler architecture to meet the requirement of extension instructions and the simulation. Moreover, we implemented the preprocessor and the native ADSP-218x instruction-set in a comprehensive validation.

並列關鍵字

Internet of Things ； Speech Recognition ； Digital Signal Processing ； Application Specific Instruction-set ； Assembler ； Integrated Development Environment

參考文獻

[1] J. Gubbi, R. Buyya, S. Marusic and M. Palaniswami, "Internet of Things (IoT): A vision, architectural elements, and future directions," Future Generation Comput.Syst., pp. 1645-1660, 29 7 2013.

[3] S. H. Park, S. H. Won, J. B. Lee and S. W. Kim, "Smart home–digitally engineered domestic life," Personal and Ubiquitous Computing, pp. 189-196, 7 3-4 2003.

[8] Y. H. Ha, "Digital signal processing".2009.

[9] M. J. Wirthlin and B. L. Hutchings, "A dynamic instruction set computer," FPGAs for Custom Computing Machines, 1995. Proceedings. IEEE Symposium on, pp. 99-107, 1995.

[10] D. E. Thomas and P. R. Moorby, The Verilog® Hardware Description Language, Springer Science & Business Media, 2002.

國際替代計量

迎向物聯網：支援語音辨識開發之可客製化數位訊號處理組譯器設計

未授權

主題瀏覽