AdaptEve was a university project created for the unusual topic of “Systemic Intervention and Eversion”.

It is a self-configuring artificial intelligence system that can adapt and evolve itself to best suit its environment. The system uses the unique concept of allowing artificial life (simulated tree branch growth behaviour) to drive artificial intelligence – a 2D L-System grows synapse-like pathways that guide the flow of data (e.g. sensor input, driver output) through an AI system.

Data pathways grow towards areas of high activity (e.g. sensor activity, data processing in the AI), are reinforced (grow thicker & sprout more offshoots) with use, and wither when not used, which allows the system to self-manage its resources.

An implementation of the system was created in Python and used to control the behaviour of a Boe-Bot.

Detailed information about AdaptEve can be found in its Project Report (PDF).