Swedish researchers are developing software for driverless trucks inspired by the behaviour of biological systems claiming it would make driverless vehicles safer and their development simpler.
The system, which will be debuted on a Volvo FH16 truck as part of the EU-funded Grand Cooperative Driving Challenge in the Netherlands later this week, is designed to react to the environment with the ease of an animal.
“Biological systems are the best autonomous systems we know of,” said Ola Benderius, researcher from Chalmers University of Technology, who led the study. “A biological system absorbs information from its surroundings via its senses and reacts directly and safely, like an antelope running within its herd, or a hawk pouncing on its prey on the ground.”
The team’s approach to designing the autonomous system therefore substantially deviated from that commonly used in the automotive industry.
“Traditionally, the aim has been to try to separate and differentiate all conceivable problems and tackle them using dedicated functions, which means that the system must cover a large number of scenarios,” Bendarius explained. “You can cover a large number of different cases, but sooner or later the unexpected occurs, and that’s when an accident could happen.”
Bendarius’s team has developed open source software that compiles data from the truck’s sensors and cameras and compiles them into a format that resembles the way in which humans and animals interpret the world via their senses. This system consists of small general behavioural blocks that enable the truck to adapt to unexpected situations with a much smaller set of pre-programmed functionalities than the systems their competitors would use.
The truck is programmed to constantly keep all stimuli within reasonable levels and to continuously learn from its experience to do this as efficiently as possible.
“We are trying to design a system that adapts to whatever happens, without pointing to specific situations – and this is something that even the simplest animals can usually do better than existing vehicle solutions,” said Bendarius.
The researchers hope that other teams will want to get involved with the OpenDLV (Open DriverLess Vehicles) software and use it for their own driverless projects.