Autonomous driving: robots do not know what they are doing

dateFreitag, Februar 1st, 2019

Although internet giants and auto companies want to push autonomous driving, there are still many obstacles to overcome. A big one might be the moral competence.

From a purely technical point of view, autonomous vehicles are already relatively safe on the roads. Their eyes are 3D cameras, sensors and laser scanners, with which they can perceive their environment in great detail. The successes could only recently be admired at the Consumer Electronic Fair in Las Vegas. There Google, Uber but also car companies and suppliers such as Daimler, Ford, Conti and Bosch showed how they visualize the mobility concepts of the future.

Autonomous driving: Lack of moral competence

The vehicles now move safely over the roads and master a diverse traffic situation. However, in the opinion of Matthias Scheutz, Professor of Cognitive Science and Computer Science at Tufts University in Medford, they can not do one thing: „Robots do not yet know what they are doing. They still lack moral competence. „But this is absolutely necessary to be able to survive in everyday life on the streets, because it requires a clear communication between man and machine, said the professor at the recent German Logistics Congress in Berlin.

Complicated order

In his experience, it’s often very simple things that robots fail to communicate with humans. A simple example is ordering a water. In this case, if a person uses a synonym, the robot does not know what to bring. Only when the person uses the correct word, the subsequent action is successful. But this happens very rarely in reality.

It already works better in intralogistics

After all, robots in areas such as intralogistics are able to interpret their environment in detail and draw corresponding conclusions from it. This then leads to independent actions. The use of picking robots shows that they have arrived in the daily work process. In addition, these robots can take the shoeboxes off their shelves or reset them again. The robots succeed in communicating with each other with the help of WLAN connections.

Learning human norms

The robots are conquering new areas in intralogistics due to their increasing intelligent applications. However, if robots need to interact directly with humans, they will continue to reach their limits. According to Scheutz, therefore, additional algorithms are needed to fill exactly these gaps. „Robots need to understand our norms and values ​​to minimize damage,“ said Scheutz. However, this step is still largely missing.

 

Written by Ralf Johanning

Ralf Johanning studied political science and works as a freelance journalist. For over ten years now he reports on the transportation and logistics industry. The priorities include topics such as telematics, software and ICT. In 2006 the trained editor, Head and Press Officer founded the editorial office Old School with his partner Ann-Christin Wimber.



Leave a Reply

Your email address will not be published.

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>

*

@

Ich bestätige hiermit, daß ich die Datenschutzhinweise in der Datenschutzerklärung zur Kenntnis genommen habe und stimme zu, dass meine Angaben und Daten zur Bearbeitung meines Kommentars elektronisch erhoben und gespeichert werden.
Hinweis: Sie können Ihre Einwilligung jederzeit für die Zukunft per E-Mail an kontakt@tis-gmbh.com widerrufen.


I herewith confirm that I have taken notice of the information in the Privacy Policy and I agree that my details and data will be collected and stored electronically to process my comment.
Note: You can revoke your consent at any time for the future by e-mail to contact@tis-gmbh.com