For some time now, scientists have been promising to introduce us to a new generation of mobile robots that can do almost everything to make our lives easier. We are told that they will be able to clean our houses, do all the weekly shopping at the supermarket, successfully steer our cars through the crowded city streets and guide us effortlessly to our holiday destinations.
So far, unfortunately, the promise has remained unfulfilled. This is mainly because the scientists have failed to programme their robots to distinguish between the permanent features of a typical street scene and those, which are temporary and can move from place to place. Some things, which are seen in a street – buildings, street lamps and so on – will always be in the same place. Other things, however, such as cars, trucks and buses, and even smaller objects, like bicycles or babies’ pushchairs, will almost certainly have been moved elsewhere.
We, humans, have no difficulty in accepting these minor changes when we observe familiar scenes, but for robots, up to now, it has been a serious problem. Earlier robots became confused by these insignificant changes and were consequently unable to function properly. Recently, however, it has been reported that members of the Robotics Research Group at the University of Oxford have given their robots a new kind of visual memory which makes them much more successful.
The new computer program, called Clever Map, trains the robot to recognise familiar objects as a whole and to accept the movement of some objects, such as cars, from one place to another. So when, for instance, a vehicle which the robot has formerly seen in a particular location is not in the same place again, it can dismiss the absence of the vehicle as unimportant. At the same time, the robot is no longer confused by the fact that identical features such as bricks, doors and windows may appear in different locations.
The research workers first took their robot for walks around the buildings and gardens of their university college. Once the robot managed to recognise the familiar surroundings, they took it out into the streets of the city of Oxford and it successfully produced maps of the whole city centre. The next step was to put a robot in the passenger seat of a car and to test it out on a thousand-kilometre journey around the countryside. This has resulted in the production of the largest robot-created maps to date.
These latest “super robots” have the ability to learn the details of their surroundings and to use a set of rules in their computer programs to deal with changes. We have not yet reached a stage of development when we can employ a robot to serve all our domestic needs, but the time is now not so far distant.