Comprehensive coverage

An autonomous truck inspired by nature and design thoughts

Swedish researchers are developing driverless truck software inspired by the behavior of biological systems. According to them, this design approach will make these vehicles safer, and their design simpler

Volvo FH series trucks. PR photo

By: Yael Halfman Cohen

Researchers at Chalmers University in Sweden are developing software for an autonomous truck inspired by the functioning of biological systems. The vehicle, a Volvo FH16 truck, was presented for the first time at the car fair in the Netherlands at the end of May. As part of a European Union project, called the Grand Cooperative Driving Challenge, about 15 universities are competing to develop autonomous vehicles, the UN's Volvo. Chalmers is one of them.

The manager of the development team, Ola Bendarios, explains that the traditional way of developing vehicles is based on improving previous models. According to him, this method will not work in the development of autonomous vehicles.

Traditionally, the goal is to anticipate all potential problems, separate them and handle them using dedicated functions that will address each problem individually. That is, the system has to handle a large number of possible situations. Thorough planning can cover a wide range of situations, but at some point, something unexpected might happen and could lead to an accident.

The team of researchers chose to treat the autonomous vehicle in an innovative way, as a biological organism rather than a technical system, a vehicle that is more like an animal.  

"Biological systems are the best autonomous systems. A biological system receives information from the environment with the help of the senses and reacts directly and safely". Says Ola Bendarios, the team manager.  

The team developed software that receives information fed from the environment to the truck with the help of sensors and cameras, and converts it into a format (regular) similar to the way people and animals interpret the world through their senses. This format allows the truck to adapt to unexpected situations, which were not taken into account in the design and planning. 

Instead of one large program with dedicated functions for expected situations, the team is working on small and general behavior units, which will allow the truck to respond to different stimuli, like an animal, for flexible management of unexpected situations and sudden dangers. The challenge is in designing a system that adapts to whatever may happen without pointing to specific situations - "this is something that even simple animals do better than existing automotive solutions", according to Bendarios.

The software is called OpenDLV (driverless vehicle, it is developed as open source and accessible on the Internet, so that other researchers in the world can join the development effort, and use the software in the development of their autonomous vehicles. OpenDLV is supposed to serve academia as a research platform in various disciplines, such as automotive engineering, adaptive systems, scientific The computer, engineering, perception, neurology and biology.

Is this the next revolution in the automotive world, comparable to the replacement of horses with motorized cars at the beginning of the 20th century?

That's what the team manager thinks. time will tell. Successfully.

4 תגובות

  1. The theory sounds very interesting, although not entirely coherent (where did this analysis of animal behavior come from? It sounds more like a human interpretation of something imagined possibly happening in the brain). Now we will wait for some execution. And let's just hope the truck doesn't stop in the middle of the main road to scrape a bit on the side.

  2. Their software is of course based on Machine Learning, and the next step will be to combine all the autonomous vehicles by Neural net.

  3. I hope there is no artificial intelligence either because then this truck will roam the streets and kill people

Leave a Reply

Email will not be published. Required fields are marked *

This site uses Akismat to prevent spam messages. Click here to learn how your response data is processed.