X
Innovation

Robots that think like humans

There are currently many efforts around the world which aim to build autonomous robots able to deal with their environment under unexpected circumstances. One of these is a UK project named Reverb, which is scheduled to be completed in 2010, is to better understand how animal brains and their visual systems. This knowledge will be then transferred to robots.
Written by Roland Piquepaille, Inactive

There are currently many efforts around the world which aim to build autonomous robots able to deal with their environment under unexpected circumstances. One of these is a UK project named Reverb (short for "Reverse Engineering the Vertebrate Brain"), led by BAE Systems with the help of several universities. The goal of this project, which is scheduled to be completed in 2010, is to better understand animal brains and their visual systems. This knowledge will be then transferred to robots, and more specifically to self-thinking or fully autonomous vehicles. Read more...

Here are some quotes from a short article published by The Business.

BAE is in the early stages of developing a revolutionary robot that could lead to a self-thinking or fully autonomous armoured vehicle.
BAE is now working with British universities, including Sheffield, Manchester, Cambridge and Oxford, to develop a robot that can recall its mistakes and remember to repeat only its successes. The idea is to create a machine that can make an immediate decision for itself based on its own readings of its environment.

This project, which started in 2004 and which will last until 2010, has been named Reverb (short for "Reverse Engineering the Vertebrate Brain").

Below is a photograph of one robotic platform that might be used for the Reverb project (Credit: BAE Systems). Here is a link to a larger version of this robotic platform.

A potential Reverb robotic platform

Here is a link to the home page of the Reverb project which provides more details about the undergoing research.

The proposed research programme aims to solve [the] problem of behavioural integration by using our knowledge of brain systems, which appear to solve the multi-tasking problem almost effortlessly. Thus, we will try to 'unravel' the mechanisms the brain uses and build models based on these mechanism that will run on computers, and which can be used to control the behaviour of artificial agents (e.g. robots).
In particular, we aim to use information about relevant parts of the brain to build a robot that can do three things: (i) recognise objects using a visual system with a narrow field of view, and use this information to retrieve certain objects; (ii) detect and direct attention to sudden events in the wider field of view; (iii) detect static features in the wide field , and direct its narrow field visual system to them for closer inspection.

In this press release from September 2005, BAE Systems gave additional details about the Reverb project.

The research will start by capitalising on current knowledge of how animals process visual information in complex, dynamic environments. This information will be used to build computational models of the relevant brain processes, which will be programmed into autonomous robots. This approach will give the robots an ability to make decisions that could in principle be as good as those taken by the animals themselves.
As an example, a robot could notice and respond flexibly and adaptively to sudden changes in its surroundings and make an immediate decision as to whether the change warrants any action. It is impossible to pre-programme a robot to deal with every possibility it might face, so the robot must combine or integrate all possible actions and behaviours in order to respond appropriately, and have the ability to learn from its mistakes and successes.

For more information about the Reverb project and all its aspects, you should look at this poster (PDF format, 1 page, 6.61 MB), which was presented at the EPSRC Novel computation initiative meeting in March 2006 under the title "Integrative computation for autonomous agents: a novel approach based on the vertebrate brain." On the image below you can see one of the proposed robotic kits, a "pan-tilt head and cameras based on careful analysis of human capabilities" (Credit: REVERB project).

One of the proposed Reverb robotic kits

So will we have to wait until 2010 to see tangible results coming from this project? I don't think so: we should see robots with rudimentary self-thinking capabilities before this.

Sources: Adam Durchslag, The Business, UK, May 7, 2006; and various web sites

You'll find related stories by following the links below.

Editorial standards