Self-driving cars: our robot car overlords are already among us

You think self-driving cars are the future? No, they’ve already been here for a while

If one can assume that Elon Musk (billionaire founder of PayPal, Tesla Motors and the world's first private space exploration company, Space X) is some kind of modern day seer, then the future is a robotic one. Recently, Musk made an interesting pronouncement – that not only were robot cars coming, but that they are indispensable. Future lawmakers will, says Musk, regard the practice of allowing fallible, mistake-prone humans control of a two-tonne projectile as positively medieval. Our presence behind the wheel will be outlawed.

So, all those prototype Google cars and road train systems currently whizzing about are the one, single and indivisible future. Robot cars really are coming and, in a transport sense, they truly will become our overlords – able to dictate our journey times and routes at electronic will.

Cruise control

As with any self-respecting Doctor Who-esque invasion, these robot cars have been around for rather a lot longer than you might think. Certainly, the technology is only now truly coming together into an homogeneous whole to create a truly autonomous, self-driving car. But the components, the individual systems which needed to be first perfected before the concept could truly come of age? Those have been around a for more than a bit.

How long? Ten years? Twenty? Surely not more than 30? More. A lot more. In fact, we have to go all the way back to 1900 and the first cruise control. Nowhere near as sophisticated as the modern, electronically-controlled versions we have today of course, but tiny British car maker Wilson-Pilcher first developed a mechanical cruise control in 1900, while 10 years later, equally-tiny Peerless was advertising a system that could “maintain speed whether uphill or down”.


It was after the second World War, and the development of auto-pilot systems for aircraft, that the self-driving car began to take its initial, humble shape. Those early cruise control systems were revived by an American inventor, Ralph Teetor.

Irritated by the fact that a lawyer friend of his would constantly speed up and slow down when driving, according to his attention levels, Teetor developed a system that read the speed from the speedometer cable and controlled the throttle opening with a simple electric motor. The 1958 Imperial was the first car to offer the system to customers.


GM fancied going better than that with the 1956 Firebird II concept car. It was envisioned (and immortalised in a fairly cheesy promotional film) that the car would be able to follow the radio signals coming from a wire embedded into the road, with suitable aircraft-style control towers at regular intervals. You may have noticed that the system never came to fruition but it could have been done, had the will and investment been there.

Anti-lock braking systems date all the way back to 1929, when they were invented (initially for aircraft) by French pioneer Gabriel Voisin, but it was into the 1960s before they were fitted to a car.

The Ferguson P99 experimental Formula One car used them, as did the Jensen FF – a four-wheel drive version of the classic Interceptor coupe which was the first production car to offer them. How are these systems robotic forerunners? Simple – they were the first to give mechanical systems control of what had been a human activity – cadence braking, a technique of braking to the point of lock-up, then releasing and then re-applying the brakes to bring a car to a controlled, steered halt. While many car-makers experimented with ABS as an option throughout the 1970s, it was 1985 and their use on the Ford Scorpio which meant that none of us have had to learn the tricky technique of cadence.

ABS lead directly to the development of the first electronic stability controls (ESC), and this was a serious step into the future of robotic control – now the car really could step in to take over, adjusting its trajectory with individual brakes if the driver got it all horribly wrong.

Again, many had tinkered with such systems through the 1970s and 1980s, including Toyota, which put an electronic anti-skid system on the Japanese-marketed Crown way back in 1983. It was Mercedes that really got there first though, backed up by electronics expert Bosch, and the two introduced the first Elektronisches Stabilitätsprogramm on the 1995 S-Class Coupe. BMW and Toyota quickly followed suit. Now ESC is all but standard on European cars and has saved near-countless lives.

The next stages of robot driving needed electronic technology, and crucially computing power, to catch up with the concept but by the 1980s, the first concepts of fully robotised cars were being developed. The father of these systems, Ernst Dickmanns, worked at the University of Munich and has been credited as the first to use a combination of cameras with stereoscopic vision with computers making real-time probabilistic predictions of what would happen based on what they were seeing.

Dickmanns’ work led to the PROMETHEUS programme, a pan-European €1 billion investment in robotic cars that had the best post-hoc acronymic explanation of its name, ever. PROMETHEUS stood for PROgraMme for a European Traffic of Highest Efficiency and Unprecedented Safety.

However clumsy the name, PROMETHEUS’ work soon bore robotic fruit – following a demonstration based in Paris, a Mercedes S-Class engineered by Dickmanns and fellow PROMETHEUS engineers drove for 1,678km from Munich to Copenhagen. It wasn’t fully autonomous – on average it needed some kind of human intervention every few minutes but it could and did operate without any manned input and its creators reckoned that for roughly 158km of the journey the computers were fully in control.

Although the PROMETHEUS project folded in 1995, its work was soon entering the public sphere – by 1998, Mercedes had the first radar-controlled cruise control system on sale, badged as Distronic. That was smart enough to detect another vehicle pulling into your path and slow down accordingly, and it was followed by Distronic Plus in 2005, which could not only slow down, but come to a complete (and safe) halt.

Between PROMETHEUS and Google, the real driver of driverless cars was the US military. The Defence Advanced Research Programmes Agency has, since 2005, run a competition for robotically- controlled cars. The idea was initially to create supply-and- rescue vehicles that could be driven into a combat zone without endangering a driver, and initially the cars simply had to follow a set list of GPS co-ordinates.

Since the first such test in 2004 (which no competitor actually managed to finish) the contest had become progressively tougher and now features urban-style environments.

Now, the developments were coming thick and fast. By 2006, both BMW and Volkswagen had developed robotic cars that could learn the contours of a race track and then set about putting in laps faster and more consistently than any human pilot. That same year saw Lexus introduce the first lane-keeping system that actually and actively turned the steering wheel to keep an errant (or dozing) driver safely in lane on the motorway. That LS could also steer itself into a parking space, and both systems have since trickled down, at occasionally surprisingly affordable prices, to more everyday cars.

Entirely superfluous

According to the US National Highway Traffic Safety Administration there are four levels of development for robotic cars, from Level 0, which is where a driver has to be in complete control the entire time, to Level 4, which is where the car effectively drives itself with the driver entirely superfluous.

Surprisingly, we are now pretty much at Level 3 – “The driver can fully cede control of all safety-critical functions in certain conditions. The car senses when conditions require the driver to retake control and provides a sufficiently comfortable transition time for the driver to do so.”

The state of Nevada has already issued licences that allow robot cars to be driven on its roads, while the British government has several self-driving car experiments lined up.

Volvo is still working closely with the European Sartre road-train project and Google is already ramping up to start production of its own entirely driverless car.

Musk may not be right, and hopefully he’s not – hopefully no legislator will ever take away by fiat the actual pleasure and skill of driving. Robot cars though? Already well and truly here, and coming in ever-increasing numbers.