Driver Assistance System, Based on the Convolutional Neural Network

The continuously rising number of deaths due to car accidents on the roads is a critical problem nowadays. To reduce the number of risky situations on the road, it is necessary to increase the attention of drivers when driving motor vehicles. In this context, an advisory system to assist driving is proposed in this paper. The mathematical model uses a convolutional neural network to recognize traffic signs and detect real road situations. It is implemented using Python in the Jupyter Notebook environment. The following libraries were used: Tensorflow (Tensors and Keras libraries, for machine learning and deep learning, respectively), Tkinter (GUI library), Pillow (image processing library), matplotlib (data visualization library), NumPy (library for performing complex computations), and Pandas (data analysis and processing library). The developed application is installed on the onboard vehicle computer. The detected sign is displayed in an appropriate form on the automobile dashboard. Simulations of real traffic situations are implemented in the system. All this will facilitate driving and help drivers with less experience acquire the necessary driving reactions.