Novel wavelet neural network real-time speed prediction method and system for fuel cell hybrid power heavy truck

The invention belongs to the technical field of vehicle speed prediction, and discloses a novel wavelet neural network real-time speed prediction method and system for a fuel cell hybrid power heavy truck. Historical driving speed information is preprocessed, and a novel wavelet neural network vehicle speed prediction training data set is constructed; the wavelet neural network in each driving state is trained in an off-line classification mode, model hyper-parameters are optimized through an optimization algorithm, and an optimal multi-dimensional wavelet neural network speed prediction model is obtained; and a real-time mode recognition result is dynamically matched with a wavelet neural network model for prediction, and a multi-step speed prediction result is obtained. The method can achieve the high-precision estimation of the vehicle speed in a complex driving state, deduces the change condition of the load power demand of the system, carries out the power pre-distribution of multiple power sources for a