Overview / Usage

In this project, we are developing a system based on deep learning to a driverless car prototype. This system captures images of the environment and makes the decision to guide the prototype. The main architecture contains two integrated Raspberry Pi 3. The first device is connected to a coupled camera in front of the prototype and captures images from the environment to the decision-making system. The second raspberry manages the controls of the prototype engines. For training the Deep Learning model, we developed the database as follows: any person drives the prototype to avoid the obstacles ahead while the device captures images and stores what the human would do in that situation. This dataset is used to train a Deep Convolutional Neural Network in a GPU server with TensorFlow framework.