Towards artificial creativity: evolutionary methods for generating robot choreographies

Today robotics is widely used in many fields, from simple houseworks like floor cleaning to more complex tasks like rescuing people in dangerous situations such as earthquakes. Recently it has been expanding to a more creative field: entertainment. For this reason we have thought of developing a genetic algorithm that allows the robot to dance, starting from the codification of movements in order to achieve the creation of true choreographies.We start by analysing Noh choreographies, and then we transpose them ontoa humanoid robot, Nao. We then proceed by going through the implementation of an algorithm that allows the creation of choreographies. One of the hardest challenges that we will face is to create choreographies that are both faithful to Noh theater and new at the same time. We will conclude focusing on the evaluation criteria of the results and presenting some hypothesis for future developments in this field.