MAS - Multi-Agent Experimentation Table

The MAS experimentation table is an interactive research platform made of four networked 55-inch touch monitors, which recognise physical objects in addition to touch input. For talks and demonstrations, the table can be turned to a vertical position, where it serves as a presentation and interaction surface.

The setup also includes a high-performance computer with CPU and GPU for computationally intensive simulations, as well as a cluster of 128 Raspberry Pi 4 units. Each Raspberry Pi simulates a single vehicle, so that the cluster can be used to study larger traffic scenarios with many vehicles acting at the same time.

Core technologies

• Networked multi-touch surface
  Four 55-inch touch displays that recognise physical objects and function as a single, continuous surface.
• Flexible form factor
  Can be turned from a horizontal experimentation surface to a vertical presentation and interaction surface.
• High-performance computing node
  CPU/GPU computer for computationally intensive simulations and AI methods.
• Raspberry Pi cluster for traffic simulation
  128 Raspberry Pi 4 units, each simulating a single vehicle, to study larger traffic scenarios in a distributed setup.

With the MAS experimentation table, we combine tangible interaction with genuine distributed computing power, creating a platform for developing, testing and presenting multi-agent systems.

More about the project behind the MAS experimentation table