Bee brains help to make robots smarter

Studying how bees sense the world could help robots become more autonomous

Honey bee brains could soon be helping robots act more independently.

The way that bees smell and see is being studied in a £1m project to produce a simulation of the insect's sensory systems.

The simulated bee brain will then be used by a flying robot to help it make decisions about how to navigate safely.

Robots that emerge from the research project could help in search and rescue missions or work on farms mechanically pollinating crops.

The research, which involves scientists from the Universities of Sheffield and Sussex, aims to create models of the neural systems in a bee's brain that helps it make sense of what it sees and smells.

The working model of the sensory systems will then be used in a robot to see if it can move around the world with the sophistication of a honey bee.

Dr James Marshall, a computer scientist at the University of Sheffield co-ordinating the project, said simulating a brain was one of the "major challenges" of artificial intelligence.

Before now, he said, many of the attempts to recreate biological brains in silicon have focused on the cognitive systems found in humans, monkeys and mice.

Computer cluster

"But," he said, "simpler organisms such as social insects have surprisingly advanced cognitive abilities."

Honey bees are well known for their unerring ability to find their way back to a colony or hive. They are believed to use the position of the sun as a reference point and can compensate for its movement across the sky when calculating the route they need to return home.

"Because the honey bee brain is smaller and more accessible than any vertebrate brain, we hope to eventually be able to produce an accurate and complete model that we can test within a flying robot," said Dr Marshall.

The models of the sensory systems will run on a cluster of powerful graphics cards that can carry out the calculations needed to simulate bee brains.

Many scientists have started using graphics cards as number-crunching engines because they are cheaper and easier to use than traditional supercomputers.

The research team hopes the simulated bee brain will produce a robot that can make decisions about what it senses rather than just carry out pre-programmed tasks.