核电子学与探测技术
覈電子學與探測技術
핵전자학여탐측기술
NUCLEAR ELECTRONICS & DETECTION TECHNOLOGY
2015年
2期
133-137
,共5页
何剑锋%王远%余加东%叶志翔
何劍鋒%王遠%餘加東%葉誌翔
하검봉%왕원%여가동%협지상
便携式谱仪%异构嵌入式系统%TI-OMAP3730%移植
便攜式譜儀%異構嵌入式繫統%TI-OMAP3730%移植
편휴식보의%이구감입식계통%TI-OMAP3730%이식
portable spectrometer%heterogeneous embedded systems%TI-OMAP3730%porting
针对传统单核处理器谱仪在实时性、可控性等复杂谱数据处理中所面临的技术瓶颈,深入探索核能谱测量技术与嵌入式系统交叉融合研究,提出了基于ARM/DSP异构嵌入系统架构下的便携式能谱仪。硬件上以构建异构嵌入式系统TI-OMAP3730平台为例,软件上实现谱数据处理、谱峰参数控制、数据高速传输等多模块并行执行。通过优化内核配置、编译并移植ARM-Linux2.6系统以及Qt/Embedded,平衡了多模块谱数据处理造成的多任务调度系统资源庞大开销。系统测试结果表明:多个谱数据处理单元能高效协作,且系统的实时性、稳定性等指标均优于传统谱仪要求。
針對傳統單覈處理器譜儀在實時性、可控性等複雜譜數據處理中所麵臨的技術瓶頸,深入探索覈能譜測量技術與嵌入式繫統交扠融閤研究,提齣瞭基于ARM/DSP異構嵌入繫統架構下的便攜式能譜儀。硬件上以構建異構嵌入式繫統TI-OMAP3730平檯為例,軟件上實現譜數據處理、譜峰參數控製、數據高速傳輸等多模塊併行執行。通過優化內覈配置、編譯併移植ARM-Linux2.6繫統以及Qt/Embedded,平衡瞭多模塊譜數據處理造成的多任務調度繫統資源龐大開銷。繫統測試結果錶明:多箇譜數據處理單元能高效協作,且繫統的實時性、穩定性等指標均優于傳統譜儀要求。
침대전통단핵처리기보의재실시성、가공성등복잡보수거처리중소면림적기술병경,심입탐색핵능보측량기술여감입식계통교차융합연구,제출료기우ARM/DSP이구감입계통가구하적편휴식능보의。경건상이구건이구감입식계통TI-OMAP3730평태위례,연건상실현보수거처리、보봉삼수공제、수거고속전수등다모괴병행집행。통과우화내핵배치、편역병이식ARM-Linux2.6계통이급Qt/Embedded,평형료다모괴보수거처리조성적다임무조도계통자원방대개소。계통측시결과표명:다개보수거처리단원능고효협작,차계통적실시성、은정성등지표균우우전통보의요구。
According to the technical bottlenecks of traditional single-core processor spectrometer in real-time, controllability and other complex spectrum data processing which it faces, measurement technology of nu-clear spectrum and embedded system cross-fusion research are explored, the portable spectrometer based on ARM/DSP heterogeneous embedded systems architecture is proposed.On the hardware, the TI-OMAP3730 platform of heterogeneous embedded systems is taken to be built for example.On the software, the parallel exe-cution of multi-module is realized, such as spectral data processing, the control of spectral peak parameter and the transmission of high-speed data.It is that the kernel optimized configuring, compiling and porting of ARM-Linux2.6 system and Qt/Embedded, balances huge overhead of multi -task scheduling system re-sources caused by multi-module spectrum data processing.The system test results show that multiple spectral data processing units can efficiently collaborate, and system real-time performance, stability and other index are better than the traditional spectrometer requirements.