IA-SVM driving condition identification method and device based on LSTM speed prediction optimization

The invention discloses an IA-SVM driving condition identification method and device based on LSTM speed prediction optimization, and the method comprises the steps: obtaining and preprocessing the historical operation data of a vehicle, constructing a kinematics fragment database, and obtaining the online identification training data of the driving condition; establishing a driving condition on-line identification model based on an IA optimization SVM algorithm, and training the driving condition on-line identification model based on the IA optimization SVM algorithm through the driving condition on-line identification training data to obtain a driving condition on-line identification model; establishing an LSTM-based speed prediction model, and training the LSTM-based speed prediction model by taking the historical travel of the vehicle as speed prediction training data to obtain a speed prediction model; acquiring vehicle operation data acquired in real time, inputting the vehicle operation data into the s