|
AboutPlatform: Windows.
Engine/Language: Unity Engine, C#. Development Time: 6 months. Roles: Programmer, Designer, Audio, Artist. Team Size: 1 Purpose: For honours project Features: 15 procedurally generating levels, Back propagated neural network, 5 themes, flood fill algorithm, affective computing (heart rate controls currently equipped gun, fire rate, movement speed), dynamic difficulty (controls enemy damage, enemy dodge chance, movement speed), real-time emotion detection (used for dynamic difficulty to determine whether the user is enjoying the experience, emotions dynamically control objects in the game such as eyebrow height controls obstacle height, blinking blinks in the game, voice lines specific to the users facial expression), 3D wave shooter, uses pulse sensor amped combined with Arduino Uno to determine real-time heart rate The Idea: The idea behind my honours project was that I wanted to combine both affective computing and dynamic difficulty in a fun way that hadn't been done before to determine whether both these fields could be used to enhance physical interaction in video games How It Works: The game required two users to play. One user wears the pulse sensor amped and the other user plays the game. The game is a wave shooter so the goal behind the game is to kill all enemies that spawn in and survive. The user wearing the pulse sensor amped' heart rate is being read in real-time to the game and this controls the currently equipped weapon, movement speed and fire rate of the weapon. For example, if the users heart rate is between 50bpm and 95bmp the user playing will be equipped with a pistol, if the user increases their heart rate between 95bpm and 130bpm they are then equipped with an SMG, underneath is the full list: Weapon Required Heart Rate BPM Pistol 50 SMG 95 Assault 130 LMG 165 The dynamic difficulty in the game runs off of two inputs which are the users average facial valence and the percent of enemies killed at the target heart rate. These two inputs are fed into a back propagated neural network which was trained on a table with all possible combinations of these two inputs. The output is a difficulty modifier from -3 to +3 which in turn then changes the difficulty of the game to meet the needs of the users. Extras: Received an A+ for the project, if you would like to see all affective computing and dynamic difficulty components in action they can be seen here. |