Evaluating and Customizing User Interaction in an Adaptive Game Controller

When playing a game, the user expects an easy and intuitive interaction. While current game console controllers are physical pre-defined hardware components with a default number, size and position of buttons. Unfortunately, different games require different buttons and demand different interaction methods. Despite that, the play style of each player differs according to personal characteristics (like hand size) or past gaming experiences. To achieve an optimal controller configuration for each player, this work proposes a virtual controller based on a common touchscreen device, such as smartphone or tablet, that will be used as a joystick to control a game on a computer or console, collecting user input data and applying machine learning techniques to adapt the position and size of its virtual buttons, minimizing errors and providing an enjoyable experience. With the prototype controller, tests were performed with a set of users and the collected data showed considerable improvements in the precision and game performance of the players.