An Approach to Implementation of Autoencoders in Intelligent Vehicles

We can see a rise in the number of smart vehicles in the last past few years. These types of cars are usually or, in other words, they are physically work as intelligent as robots. Intelligent vehicles have become an important part as they are equipped with intelligent agents that give services to human beings. It is approximated that over 1 billion cars travel the streets and roads of the world today. With such traffic, it is apparent that there are many situations where the driver has to react quickly. As Intelligent vehicles are connected to a large amount of data, these data may be dimensionally decreased and kept as latent data. Then, when needed, they can be reconstructed and used. The aim of the current paper is an approach to the implementation of an unsupervised autoencoder technique in intelligent vehicles. The autoencoders have significant importance as they detect and recognize unknown data. In this case, we can say the autoencoders may replace labelled supervised neural networks if they learn effective encoding (data representation).