Prediction Vehicle’s Speed with Using Artificial Neural Networks

We propose an approach for the vehicle’s speed prediction based on artificial neural networks. Different types of artificial neural networks were considered including MLP and RNN. A complex urban route in Kazan city was chosen for data gathering and making experiments. We demonstrated that it is possible to obtain sufficient accuracy for speed prediction based on limited source data. We got the prediction accuracy as 99.6% for the first future second and 94% for the tenth future second. Simple RNN showed better results for given data. We can use the suggested approach for designing intellectual automatic transmission systems and other intelligent transport systems applications.