This study demonstrates the capabilities of bio-potentials especially Electrooculography (EOG) for smart environmental control applications. Furthermore, In this study, we have proposed a novel, user-friendly, low cost, and wearable Man Machine Interface (MMI) for practical use out of a controlled laboratory environment. EOG based MMI framework is designed by using consumer-grade single channel Brain-Computer Interface (BCI) device. Acquired signals are filtered and processed for feature extraction. Eye blink signals are detected and decoded using Application Program Interface (API) developed in MATLAB. Four setups for real-time control experiments spanning over 300 trials are conducted to test the efficiency of EOG signals for smart environmental control applications. Additionally, real-time wheelchair control experiments are performed by five volunteers for testing and cross-validating the quantitative and qualitative factors of designing MMI. Finally, results of wheelchair control experiments are compared and contrasted with similar established MMI frameworks. Overall, Four setups of smart control experiments conveyed an average precision of 96.44%, 99.30%, 97.11%, and 95.78%, respectively, with a good response time ranging between 1.65 s to 2.23 s for two participants with 300 trials. Whereas for wheelchair control experiment, wherein five subjects volunteered, an average accuracy of 93.89% and 62.29 bits/min of Information Transfer Rate (ITR) were obtained with zero collision and zero False Positive Rate (FPR). From the results it can be inferred that, EOG signals can be applied to different smart environment control application with great potential; Proposed single channel EOG based MMI framework can provide a robust, practical, user-friendly yet at the same time precise and responsive interface for control applications at a moderate cost. In this investigation, an effective, low cost, wearable inter has been presented for the variety of users and various applications. When contrasted with the recently established comparative MMIs, the displayed framework gives a novel and proficient interfaces without Graphical User Interface (GUI) with more prominent exactness and better ITR.
Read full abstract