The Man Who Gave Cars Eyes: Ernst Dickmanns

The Man Who Gave Cars Eyes: Ernst Dickmanns
The DLR compound at Oberpfaffenhofen

While the world is only just getting used to the idea of Google’s driverless cars being on the roads around us, not many know that autonomous cars were driving around all the way back in 1986. This was thanks to one pioneering German man with a vision for giving the gift of sight to computers — Ernst Dickmanns.

Images via Wikimedia Commons

Although best known for his work with autonomous vehicles, Ernst Dickmanns didn’t move into this field until much later into his career. Dickmanns was born in Nazi Germany near Cologne in 1936, although the war was over by the time he graduated from Polz Gymnasium in 1956. At that point in his life he wanted to work in aerospace engineering, at a time when prospective engineers were required to do practical work in the industry.

Manoeuvres In Orbit

After undertaking practical work in both metallurgy and aircraft engineering, he earned his diploma of engineering in 1961. From there he went to the Deutsche Forschungs- und Versuchsanstalt für Luft- und Raumfahrt, the German Test and Research Institute for Aviation and Space Flight — the organisation that is now the German Aerospace Center. As he described it in a 2010 interview, the centre was something like a German version of NASA.

Dickmanns’ passion was always for aerospace research, as evidenced by his early work and studies. While associated with the German Aerospace Center, he was also working towards a PhD at the Technical University in Aachen in the field of trajectory optimisation. Towards this end he designed manoeuvres for orbiter re-entry, thinking it could have potential applications on the space shuttle program. As he was disappointed to discover, however, the work was doomed to be solely theoretical as in practice the shuttle was far too fragile to even attempt manoeuvres on re-entry.

After earning his PhD, Dickmanns was invited for a Post-Doc Research Associateship at NASA’s Marshall Space Flight Center by a friend he had met at Aachen. This continued on from his doctorate research, working on the shuttle orbiter’s re-entry.

After returning to Germany, Dickmanns began working in satellite control, working in part of the team that orchestrated the launch and positioning of Europe’s first communications satellite array in 1974. Soon after, he was made acting head of the German Aerospace Center in Oberpfaffenhofen, overseeing almost 700 people.

It was during this time that Dickmanns first started looking into the idea of developing ‘vision’ for computers and vehicles, though at its highest level it began with satellite remote sensing systems. In 1977 he started working on the project in earnest, working towards developing vision systems for mobile units — although at the time he had aircraft, spacecraft and other above-ground vehicles in mind. Using the technology for regular road vehicles did not even cross his mind at the time.

Helping Cars To ‘See’

The jump to cars came after they were approached by Daimler-Benz who wanted to collaborate on a project in celebration of the 100 year anniversary of the company’s first car (the Benz Motorwagen in 1886). Moving into their second century of car manufacturing, Daimler-Benz proposed a large-scale research project to develop new technologies — one of which was to be autonomous driving.

But first came Dickmanns’ team’s proof of concept — the VaMoRs. VaMoRs was a 5-tonne Mercedes van equipped with cameras and other sensors, modified so that all necessary controls — steering wheel, braked and throttle. The software largely looked at the white lines on the road, and major colour differences in images.

VaMoRs made its first autonomous drive in 1986, for safety reasons taking place on streets without traffic. Due to the incredibly slow (at least by today’s standards) processing speed of the computers the team were using, Dickmanns had to come up with a way to navigate in real time using a computer that processed images in a matter of seconds, rather than nanoseconds or even milliseconds. He and his team called this the ‘4D approach’.

This system estimated spatial positioning and velocity, without the need to store previously captured images. It was also designed to focus only on the most relevant details of visual input, like areas of high contrast or changes in colour or texture. The VaMoRs made its move onto public roads just a year later in 1987, driving autonomously on the Autobahn at speeds of up to 96km/h — the maximum speed the van was capable of. Dickmanns does note that their license for testing the vehicle specified that at least person had to be inside while it was driving — with their reasoning being that their own safety would make them be more cautious with the car’s testing.


In the same year, the Prometheus project — PROgraMme for a European Traffic of Highest Efficiency and Unprecedented Safety — was initiated under the European Eureka research funding system, receiving €749 million in funding to become the largest R&D project ever seen in the field of driverless cars.

The initial plan was to use buried cables along the Autobahn to guide autonomous cars down the highway, but Dickmanns’ previous successes were enough to prove the efficiency of the much more flexible computer vision system.

Prometheus was a little bit sexier than the original VaMoRs, involving two modified 500 SEL Mercedes named VaMP and VITA-2. Thanks to the large amount of funding, the computers used for this project were also upgraded, using up to sixty transputers (a type of parallel computer) for the processing power needed.

The Prometheus project achieved its first culmination point in 1994, when the two vehicles were driven autonomously through Paris, able to navigate and change lanes through traffic. For this demonstration — during which passengers were on board to experience the new technology — each car used two cameras with different focal lengths for each hemisphere.

A year later in 1995 the cars took a trip from Munich in Germany to Copenhagen in Denmark, reaching speeds of up to 175km/h on the Autobahn and driving an estimated 158km without human intervention. This trip was meant as a way for the team to collect data on what systems needed to be improved in subsequent generations of autonomous technology.

Some of the challenges faced then are ones that engineers of autonomous vehicles still face today. One that Dickmanns singles out as being particularly tricky is that of ‘negative obstacles’ — that is, potholes or other missing materials in the driving surface. It only becomes more difficult if the potholes are, say, filled with water — although flooded roads have proven to be difficult even for human drivers to judge.

The Future Of Self Driving Cars

In the late 90s and early 2000s Dickmanns worked with both German and US military institutes on developing ‘off-road’ navigation and detection of obstacles and ditches. To this end the team developed “Expectation-based, Multi-focal, Saccadic vision,” otherwise known as EMS-vision.

By this stage his work had set the standard for autonomous driving, however, and his machine visions systems found themselves being used in a multitude of applications — including in aviation, where his work first began, and even in experiments on board the Space Shuttle Columbia.

Current driverless cars use far more sophisticated technology — aside from having computers with much greater processing power, they also make use of GPS to position themselves in the world. In 2010 Dickmanns described how this step forward was actually a step backwards — for example in the DARPA autonomous vehicle challenges (the Grand Challenge and Urban Challenge), he opines that the cars were merely guided through a sequence of steps by GPS, rather than developing true obstacle avoidance and machine vision.

These days Google has its fleet of driverless cars, while Volvo is in the process of developing its own. Nvidia has even created a compact, high-powered computer that is capable of processing the data needed in real time. Tesla Motors’ cars are developing sophisticated self-driving systems, and three decades after its work with Dickmanns, Mercedes is developing an “Autobahn Autopilot” for use on highways.

While the technology in the VaMoRs and the VaMP are now obsolete, Dickmanns’ most important work was in proving that machines could ‘see’ well enough to navigate roads like a human driver would — no cables or magnetic strips needed.

Source: Ernst Dickmanns, an oral history by Peter Asaro, Indiana University, Bloomington Indiana, for Indiana University and the IEEE, 2010.

These Are Your Numbers is a new Lifehacker series where we profile great minds that have made significant contributions to robotics and computing.