本文研究重點主要為實現於多參數感測盲源分離系統,以離子濃度變化作為實驗設計考量,並且以此實驗設計考量作為盲源分離技術改變之特徵;首先將前端ISFET、EGFET搭配電壓式讀出電路,得到之訊號有效地經過盲源分離系統辨別其原始訊號,得到其主要離子及干擾離子變化之特徵曲線。本系統主要先根據整體系統讀出電路讀出電壓值關係,使用定點數演算法來設計,進行硬體系統建置,將各式運算方法有效地實現在硬體上,最後再利用FPGA(Field-programmable gate array)進行整體系統之驗證,確認已達到功能後,利用合成軟體Design Vision進行RTL合成,再使用自動佈局軟體IC Compiler進行晶片的佈局。
This research focuses on the implementation of a Blind Source Separation (BSS) algorithm system on a multi-parameter sensing platform. In the multi-ion chemical environment, the source signals correspond to the concentration of main ions and the concentration of interference ions. Experiments were conducted where the concentration of the main ions is changed to predefined amounts as the concentration of the interference ions is held constant. Ion-Sensitive Field Effect Transistor (ISFET) - based Potentiometric Readout Circuit was used to get raw data from these concentrations. Compared with the known sample source signals, the BSS algorithm has successfully estimated the original sources from the observed signals from the readout circuit. Hardware systems that execute the BSS function use the fixed-point algorithm thereby showcasing an efficient computational method. Finally, the entire BSS design is embedded into a Field Programmable Gate Array (FPGA). Firstly, the functionality of the BSS algorithm has been verified using MatLab. Then, the algorithm was implemented onto a verilog code into a hardware architecture. RTL synthesis was undergone using the synthetic software Design Vision. Finally, the automatic placement and routing software IC Compiler was used to implement the code into an Application Specific IC (ASIC) which can readily sent for fabrication using TSMC 0.18um CMOS 1P6M Mixed Mode Process.