Invisible Networks

Swarm robotics inspired by natural phenomena, such as shoals of fish, are able to flexibly adapt and respond to changing conditions in their environment. These flexible and adaptive systems are pivotal for engineering real-world applications that benefit society and protect the natural world.

In a world of invisible forces, power no longer sits at the centre: it moves like a school of fish, fluid, responsive, uncatchable.

Imagine a school of fish. Each fish moves independently without any central control yet responds and reacts in beautiful synchrony with the other fish to environmental changes, such as an approaching predator.

In this project, ‘Invisible Networks’, I use robotic fish to demonstrate the nature of decentralised systems and the complexity they hold. The project transcends the basic notion of central control, showcasing how local interactions can cohere to create complex, emergent behaviours on a global scale.

I designed a swarm of ‘Invisible Fish’ focusing on soft robotics and control algorithms. These waterproof agents are cast using silicon and navigate using a custom wound electromagnetic actuators for propulsion. I combined a Raspberry Pi Zero with computer vision algorithms and flocking behaviour to successfully mimic swarm intelligence. The robots mimicked biological swarm behaviours using visual neighbour detection through real-time LED coordination and localised signal processing. You can refer to my github to see a more in-depth architecture of this project.

This project was then accepted as a poster presentation at 2025 Ada Lovelace Colloquium.