For some reason, kids love robots - R2D2 in Star Wars, the Sojourner robot on Mars, Robot Wars on television. If robots are such fun, and if fun is the route to good education, then robotics should be part of the national curriculum.

The wonderful thing about building a robot is the diversity of skills you have to bring to bear. And therein lies the special educational value of robotics.

For example, to design and build the robot’s body, you need a grasp of simple mechanics. You need to know about cogs and wheels, about structures and stresses. (Most robots have wheels. But some have legs, which are a real challenge.) Such things are the very foundations of mechanical and civil engineering. Little roboticists can grow up to design cars, or factories, or bridges.

To get a robot to move across the classroom floor, a child needs to know a bit about motors and batteries, which in turn demands some understanding of electricity and circuits. To get the robot to sense the classroom - to detect a wall, or an obstacle, or another robot - the child needs to learn about electronic devices, such as infra-red sensors, or sonar. These are the sorts of things studied by electrical engineering students, who go on to design and build so many of the artefacts our modern world depends on, from the insides of cash dispensers, to the radar used in air traffic control.

Suppose our child has made her first robot. The chassis is built, the motors are in place, the sensors are wired up. She gives it a name - Pickle. Now, how does she persuade Pickle to do something interesting? A simple way is to use a control loop. As Pickle rolls along, an infra-red sensor continuously tells it roughly how far it is from the wall. If it strays too far, it steers towards the wall. If it gets too close, it steers away. In this way, our little roboticist programs Pickle to follow the wall unfailingly, without ever hitting it. Building on her knowledge of control theory, she might grow up to engineer the systems that run chemical plants, or regulate power stations, or guide aircraft.

Her school years are passing along happily, and our little roboticist is now well into her teens. But classroom robotics projects are getting harder, and now she has to work in a team. Their latest project is to program Pickle to explore a maze.

All the members of the team think they’re big computer experts. They’ve been using the Internet for years. But now, they have to learn how to actually program computers. (In all their earlier experiments with robots, they used some special software with a fancy graphical user interface that hid all the working details.) Fortunately, the robots can be made to do really cool things, so everyone is highly motivated to acquire the art of programming. One team infects a rival team’s robot with a virus that makes it emit offensive noises, but somehow the staff manage to retain control. Soon, the team is working on a program that sends Pickle down each of the maze’s paths in turn, and builds a map of them as it goes along.

Some members of the team spend a lot of time arguing over whether or not the robot is thinking while it explores. As well as inspiring the programmers of the future, robotics might launch a career or two in philosophy or politics.

But you can’t make an omelette without breaking a few eggs.

In short, robotics encompasses mechanics, electronics and computer science, all critical disciplines in contemporary technological society. So a few hundred hours of every child’s life spent mucking about with robots would be a very good thing, regardless of their ultimate vocation. But in my dreams, some of these young roboticists end up as grown-up roboticists, like myself.

Maybe, with an influx of talent and enthusiasm, the vision of robots with human-level artificial intelligence will become a reality in my lifetime. I wonder what the national curriculum will look like then?

Dr Murray Shanahan is a lecturer in the department of electrical and electronic engineering at Imperial College, London