嵌入式跨平台多工腦機介面應用於智慧型居家環境控制

Abstract

在以往的生理訊號監控系統大多只做到訊號記錄部分，再藉由電腦進行離線訊號處理，而缺乏即時分析的能力。因此，本論文提出一套具有即時分析能力的腦機介面系統以改善一般生理訊號監控系統的缺點，並結合腦波精神狀態監測與居家環境控制，提供給行動不便的病人、老人一個舒適的居家環境。為了達成智慧型居家環境的控制，本論文開發一套嵌入式跨平台多工排程方法來增加系統於即時處理效能，最大的特色在於這個多工排程可以將一個工作切割成數個子工作來減少運算的時間，而訊息交換的架構具跨平台的能力能讓工作分別在不同作業系統平台的機器上工作。最後，這個多工排程方法結合即時的生理狀態偵測的演算法，並整合不同平台的環境控制器來控制週遭的環境，進而提升使用者之生活品質。

The most former biomedical signal supervisory systems only realize signal recording, and then these signals are processed by a powerful computer for off-line analysis. Hence, the conventional system lacks the capability of real-time computation and analysis. Due to the above drawbacks, this thesis proposes an embedded-based brain computer interface (BCI) system to overcome these problems. This system combines the EEG monitor and smart room control to provide a comfortable living environment for the handicapped patients and elderly people. To achieve wisdom-domestic environmental control, this thesis proposes multi-task and cross-platform schemes to achieve real-time computation. The multi-task processing approach divides the task into many sub-tasks to significantly reduce the computation time. The data-exchange scheme has the capability of crossing the platforms and allows tasks to operate at different machines with distinct operating systems. At last, the simulation and demo results show that the proposed methods can achieve online biomedical signal analysis and can be integrated into the smart room controller among different platforms for better life.