Hybrid Fuzzy Expert System and Difference Equation Software Filter for Biomedical Sensors

In this study, a hybridized fuzzy expert system (FES) and difference equation (DE) algorithm is proposed to filter the erroneous and noisy data obtained while reading data from sensors with microcontrollers. The results produced by the proposed software filter were tested with data read from the digital airflow sensor, digital pressure sensor, and analog chest movement sensor, which are frequently used in the medical field and are commercially available today. With the proposed software filter and other software filters in the literature, the data were instantly read from these sensors. The operation frequency of the proposed software filter was calculated as 81.8% for the airflow sensor, 83.37% for the pressure sensor, and 87.3% for the chest movement sensor. With the proposed software filter, the root mean square error (RMSE) values of the data read from the airflow, pressure, and chest movement sensors were calculated as 0.01, 0.03, and 0.01, respectively, standard deviation (S) values were calculated as 0.003, 0.02, and 0.266, respectively, signal-to-noise ratio (SNR) values were calculated as 54, 42, and 55, respectively, and mean absolute error (MAE) values were calculated as 0.009, 0.027, and 0.008, respectively. These results show that the proposed software filter produces more successful results in terms of operating frequency and filtering success than other algorithms compared in this study.