Synchronization algorithm embedded platform integration method for noninvasive ventilator

The invention discloses a non-invasive ventilator synchronization algorithm embedded platform integration method based on an LSTM neural network. The integration method comprises the following steps that (1) the LSTM neural network is adjusted and converted into a C language program available for a non-invasive ventilator embedded platform; (2) writing a data preprocessing program based on the LSTM neural network definition; (3) writing a data post-processing program based on the LSTM neural network definition; and (4) packaging all programs in the steps (1) to (3), providing three application program interfaces (API) for data input, prediction execution and prediction result acquisition to the outside, and using the three application program interfaces in an embedded platform of the noninvasive ventilator so as to complete integration of a synchronization algorithm. According to the method, the integration of the non-invasive ventilator synchronization algorithm based on the LSTM neural network on an embedde