Detection of Apnea Event with ANN Using Acceleration Data

In this study, the accelerations caused by the diaphragm movements during respiration were monitored with a 3-axis accelerometer and the measured accelerations were recorded on an SD card. A measurement system for this purpose was developed. The moments in which diaphragm movements stopped were detected by using Matlab. An ANN has been designed to simulate measured real data. A total of 5886 real data were applied to ANN. In the training of ANN, 3943 randomly selected from these data (66.6% of the total data) were used. The remaining 1943 data (33.33% of the total data) was also used for the test. Thereby estimating the apnea event was provided by the designed ANN. The results were plotted and proved to be quite similar to each other.As a result, apnea events have been successfully detected.