10/2019 - 2/2020
Anxious Intelligence
Engine/Software
Languages
Roles
Notable Features Worked On
: Unity3D
: C#
: AI Programmer
: Fuzzy Logic Implementation & Design
About
For winning the 100% FAT CreaTe Award in GOGBOT 2019 for our Lurian Cube installation, we were invited to create an installation for the TECART 2020 exhibit as part of Rotterdam Art Week. In the same group along with additional help, we created Anxious Intelligence. Anxious Intelligence explores how an artificial intelligence can experience the same states of mind that humans do, such as anxiety and introversion. Anxious Intelligence is represented as a particle system that wanders around a projected screen. Anxious Intelligence uses the Kinect to determine how many people are in a room with it, how close they are, and how quickly they are approaching it. The Anxious Intelligence will then react appropriately, flashing red, becoming more violent, breathing heavier and faster, and creating more disturbing sounds if people approach it too quickly or there are too many people in the room.
Anxious Intelligence was featured in an article on Innovation Origins.
Fuzzy Logic
Building on what we had created with the Kinect, we shifted our implementation from Processing to Unity. As the lead programmer of this project, I decided to utilise fuzzy logic for the AI of the Anxious Intelligence. The architecture for fuzzy logic was ported over from C++ to C# from Mat Buckland's "Programming Game AI by Example", code taken from https://github.com/PacktPublishing/Unity-2017-Game-AI-Programming-Third-Edition.
To determine how it should react, it took as its variables the number of people detected by the Kinect, the distance of the closest person to the Kinect, the average distance of the crowd to the Kinect, and the speed at which people are approaching it. From this, 2 sets were created, anxiety which took as its variables the crowd size, closest distance to the AI, and the average distance to the AI, and agitation, which took as its variables the closest distance to the AI, and the crowd closing rate.
Anxiety would then affect how quickly the particle system moved, and the breathing speed of the AI. Agitation would affect the colour of the particle system, and how violent the sound effects it emits would be. The particle system would also attempt to avoid where people are relative to itself, so if people crowded at the right side, the particle system would stay at the left side of the screen.