Abstract
The objective of this report is to show the results of our approach of a motivational drive-based architectures using a physics engine for modulating the agent's action selection processes. Rigid-bodies are used to represent abstract mind concepts, like needs, actions and perceptions, and the collision between these objects triggers agent behaviour. This thesis presents our conceptual model that integrates the standard motivational drives model into a physical context, along with our implementation, tests and results of that same model. We have also researched how different architectures that use motivational drives implement concepts like, perception and action selection, memory and learning, emotions and personality, in order to understand key feature that our system must have in order to produce believable behaviour.