An autonomous vehicle control system that draws on historical driving information as well as real-time data could be the key to enabling driverless vehicles to operate in extreme conditions.
The system, which was tested on the racetrack aboard a specially adapted Volkswagen GTI and an Audi TTS, could help cars perform more safely in extreme and unknown circumstances, claim its developers at Stanford University in the US.
While current autonomous cars rely on in-the-moment evaluations of their environment, the group claims that advanced understanding of road conditions – such as how to perform an emergency manoeuvre on ice – will be key to pushing driverless cars to the limits.
“With the techniques available today, you often have to choose between data-driven methods and approaches grounded in fundamental physics,” said J. Christian Gerdes, professor of mechanical engineering and senior author of a paper on the research in Science Robotics. “We think the path forward is to blend these approaches in order to harness their individual strengths. Physics can provide insight into structuring and validating neural network models that, in turn, can leverage massive amounts of data.”
The new system is underpinned by neural network – a type of artificially intelligent computing system – that integrates data from past driving experiences at Thunderhill Raceway in Willows, California, and a winter test facility with foundational knowledge provided by 200,000 physics-based trajectories.
The group ran comparison tests for their new system at Thunderhill Raceway. First, the Audi, aka “Shelley” sped around controlled by the physics-based autonomous system, pre-loaded with set information about the course and conditions. When compared on the same course during 10 consecutive trials, Shelley and a skilled amateur driver generated comparable lap times. Then, the researchers loaded the VW “Niki” with their new neural network system. The car performed similarly, running both the learned and physics-based systems, even though the neural network lacked explicit information about road friction.
In simulated tests, the neural network system outperformed the physics-based system in both high-friction and low-friction scenarios. It did particularly well in scenarios that mixed those two conditions.
Whilst the results were encouraging, the group said that their neural network system does not perform well in conditions outside the ones it has experienced. The next step is enable the system to learn from a larger pool of data.
“With so many self-driving cars on the roads and in development, there is an abundance of data being generated from all kinds of driving scenarios,” said Nathan Spielberg, a graduate student in mechanical engineering at Stanford . “We wanted to build a neural network because there should be some way to make use of that data. If we can develop vehicles that have seen thousands of times more interactions than we have, we can hopefully make them safer.”