Category Archives: Technology

From the Safety Bicycle to Robotics

One winter’s day late in the 1970s, driving south on the M6 in steady rain, I saw something very strange flash by in the outside lane going north: a two-wheeled vehicle that looked as if it was a prop from Star Wars, its rider serenely powering through the traffic.  It was one of these:

A Quasar.  My question is: why don’t all bikes today look like this, rather than the bike on the right, whose basic design has advanced so little since the the Safety Bicycle circa 1890?  The rider on the right looks so stiff and awkward, perched on top of his machine.  If you were starting out today designing a powered two-wheeler it seems obvious you would come up with something like the Quasar.  But it never caught on, despite its advantages: it steers, corners and brakes more safely than a conventional motorcycle, the rider is protected from the elements, rider and machine are much lower, which aids stability and decreases the machine’s profile, so while it was comparatively underpowered it was fast and used about half the fuel of a conventional motorcycle – about 90 mpg. (see http://www.realclassic.co.uk/quasar06062000.html )

An answer to the question would include both the resistance of consumers to anything that unfamiliar, and the reluctance, for a number of reasons, of producers to move beyond variations of the familiar.  In a world supposedly obsessed with novelty and innovation, the failure of motorcycle design to have made much progress at all for about a century is striking.  This rather suggests that much of what is passed off today as novel has been “novel” for rather a long time – a basic lesson of David Edgerton’s excellent book, The Shock of the Old (2006).

Some kind of explanation for this failure of invention to be followed by innovation can be found in the history of the unpowered two-wheeler, aka the pushbike.  A recent article in Nature (Brendan Borrell, “The bicycle problem that nearly broke mathematics”, 20 July 2016) demonstrates that, while the safety bicycle quickly proved stable and useable, nobody really knew how it worked as a two-wheeled vehicle with a human power source.  What role, for instance, does counter-steering play?  When in normal motion, the rider first pushes the left hand forwards to make a left turn – which is counter-intuitive, and is most probably done unconsciously.   On a motorbike with a longish wheelbase, like my Guzzi Le Mans, you could do this deliberately, dropping the bike into a bend by pushing the left hand forward while opening the throttle with the right.  Again, what role does the relation of the front wheel’s contact point to the frame geometry play?  And for a long time it was thought that the gyroscopic forces in a spinning wheel explained a bicycle’s stability, but this also turns out to be only a part of the answer.

The one real innovation in post-Safety Bicycle design – the recumbent, where the rider sits low, as in a pedalo (see David Gordon Wilson, Bicycling Science, 3rd. ed. MIT Press 2004 pp. 444-450) – likewise evolved over a long period of practical experimentation, in the absence of any real idea of what contributed to two-wheeled stability, and what diminished it.  As the Nature article shows, this would require some mathematical modelling that could satisfactorily account for all the mundane things that a bike does, with or without a rider.  It was not until 2007 that a full set of mathematical equations was published that accounted for all of these things.

Whereas much of modern economics has begun with a set of equations borrowed from somewhere else and then forced economic life into mathematical form, in the case of the bicycle we have an object that functioned satisfactorily, but no-one really knew why.  Modelling the motion mathematically gave insight into what happens when a bicycle turns or wobbles, and why some configurations feel better than others.  This turned out to be much more than an academic question.

Unpowered two-wheelers have two principal components: a bike, and a rider who is the power source.  The interaction of these two components has also been a source of mystery.  Million upon million of children have learned to ride a bike, but how they learn and what they learn, these are both very obscure.  At last some kind of progress has been made with the latter, but the motor skills needed to control a moving two wheeler are similar in complexity to those needed for a toddler to take its first steps.

Which brings us back to the humble Safety Bicycle, progenitor of the motorcycle, thence the automobile, and then the control technology of Orville and Wilbur Wright’s flying machines.  Now that bicycle mathematics are, after a century of pondering and struggle, on a firmer footing, they turn out to lay the foundations for the future: robotics and prosthetics.  A clearer understanding of the behaviour of two-wheeled vehicles and the nature of their interaction with a human rider can be used to improve the design of artificial legs, and also improve the mobility of robotic devices.  Now there is even a robotic bike than can balance itself simply by steering, without the use of gyroscopes.  Forget about the driverless car: the riderless bike is already here.