Behavior Studio Project

This project takes as a basis the final works of Vanessa Fernandez, Francisco Pérez and Nacho Arranz to create a test environment that evaluates the quality of the models trained with Deep Learning for an autonomous car to complete a lap of a circuit.

The aim is that, by loading a model in TensorFlow, Pytorch or even explicitly programmed, conclusions can be drawn about the quality of the model.

For more information visit the Behavior Studio project page.

• Vanessa Fernández (master): Deep Learning for Visual Control.
• Francisco Pérez (master): DeepLearning in autonomous vision based navigation of real robots.
• Nacho Arranz (master): Deep Reinforcement Learning for autonomous car.

Challenges

The aim of this project is to obtain a platform for evaluating neural network models. Starting with the behavior of autonomous cars but also extendable to other types of robots, such as drones, humanoids, etc.