��«Ӱҵ

Introduction

��«Ӱҵ the project

The development of an ‘artificial brain’ is one of the greatest challenges in artificial intelligence, and its success will have innumerable benefits in many and diverse fields from robotics to cognitive psychology.

Most research effort is spent on modelling vertebrate brains. Yet, smaller brains can display comparable cognitive sophistication while being more experimentally accessible and amenable to modelling.

The ‘Green Brain Project’ combined computational neuroscience modelling, learning and decision theory, modern parallel computing methods, and robotics with data from state-of-the-art neurobiological experiments on cognition in the honeybee Apis mellifera. 

These various methodologies were used to build and deploy a modular model of the honeybee brain describing detection, classification, and learning in the olfactory and optic pathways as well as multi-sensory integration across these sensory modalities.

This research was supported by the EPSRC Grant Number EP/J019690/1.

---

Project goal

Simply put, the goal of the Green Brain Project was to create a robot that thinks, sense, and acts like a honeybee. 

This was done by creating neuromimetic models and embodying it within flying robots.

---

Contributions

Developing a better understanding of cognitive functions in animals

It has been well established that the honeybee Apis mellifera has surprisingly advanced cognitive behaviours despite the relative simplicity of its brain when compared to vertebrates.

These cognitively sophisticated behaviours are achieved despite the very limited size of the honeybee brain (on the order of 10^6 neurons). In comparison, even rats or mice have brains on the order of 10^8 neurons.

The greatly reduced scale and the experimental accessibility of the honeybee brain makes thorough neurobiological understanding and subsequent biomimetic exploitation much more practical than with even the simplest vertebrate brain.

Modelling of the honeybee brain will focus on three brain regions:

• The system for olfactory sensing
• The system for visual sensing
• The mushroom bodies for multi-modal sensory integration

These systems are chosen as they exhibit complex cognitive behaviours essential to autonomous agents which are not currently understood.

Save the Honeybees

Honeybee populations have been in drastic decline in recent years, and scientists don’t have a good answer for the phenomenon.

This is bad news for the world since honeybees are known to be crucial for pollination and hence for the world’s food security.

Our team alone can't save them, but by studying their brains and behaviours we have helped scientists to understand this trend.

Adaptive and robust algorithms for autonomous, unmanned aircraft

Increasingly over the years, Unmanned Aerial Vehicles (UAVs) are being used to perform missions that are considered “dull, dirty, and dangerous”. Examples include:

• Nuclear Power Plant Missions
• Search and Rescue
• Weather Forecasting
• Pipeline Monitoring
• Nuclear Power Plant Missions
• Search and Rescue
• Weather Forecasting
• Pipeline Monitoring

As the trend develops toward the increasing use of UAVs, it becomes necessary to allocate and control them effectively.

One of the biggest challenges in the area of UAVs is in developing algorithms that not only have good performance but are also adaptive and robust.

This is essential for UAVs that need to perform in the real world which is ever changing and hardly certain.

By extending existing models and techniques of honeybees, we demonstrated sophisticated visual-based navigation and cognitive functions in a flying robotic platform.