Researchers from the US and China have developed data-driven realistic simulation technology for self-driving cars aimed at improving safety before road tests.

The new Augmented Autonomous Driving Simulation (AADS) technology has been developed by Dinesh Manocha, a computer scientist from the University of Maryland in collaboration with scientists from Baidu Research and the University of Hong Kong.

According to the scientists, their AADS will help easily assess self-driving technology in the lab and ensure more reliable safety before road testing begins.

University of Maryland institute for advanced computer studies researcher Manocha said: “This work represents a new simulation paradigm in which we can test the reliability and safety of automatic driving technology before we deploy it on real cars and test it on the highways or city roads.”

In the current simulator technology, the perception module of the autonomous cars receives input from computer-generated imagery and mathematically modelled movement patterns for pedestrians, bicycles, and other cars.

“Because we’re using real-world video and real-world movements, our perception module has more accurate information than previous methods.”

It is not a realistic representation of driving conditions, as well as an expensive and time-consuming process.

The new simulator pools photos, videos, real-world trajectory, and behavioural information into a scalable, realistic autonomous driving simulator. It offers an improved reliable simulation compared to current systems and leverages high-fidelity computer graphics. Sorry, there are no polls available at the moment.

Using these trajectories, the self-driving cars predict the driving behaviour and future movements of other vehicles or pedestrians on the road for safer navigation.

Manocha further said: “We are rendering and simulating the real world visually, using videos and photos. But also we’re capturing real behaviour and patterns of movement. This data-driven approach has given us a much more realistic and beneficial traffic simulator.

“Because we’re using real-world video and real-world movements, our perception module has more accurate information than previous methods. And then, because of the realism of the simulator, we can better evaluate navigation strategies of an autonomous driving system,” Manocha concluded.