Autonomous vehicles are coming to the mass market, and to make sure the transition from human-controlled to computer-controlled is as smooth as possible, the systems need to be tested significantly.
However, there is a nervousness that testing autonomous vehicles on public roads before the technology is mature enough could put lives at risk. There have already been a number of high-profile incidents that have heightened cynicism around the technology.
It’s human nature to be wary of something new, but unfortunately it’s also human nature to make mistakes, which is why there are still a large number of accidents on our roads. And all too often the statistic needs to be rolled out that 90% of road traffic accidents are down to human error.
Autonomous vehicles have the ability to reduce this number, making roads safer, but to reach that end goal the technologies that enable it need to be thoroughly tested and validated. If you can’t do that on public roads, where can you?
Put simply, in the mind of a computer. How many driving simulation video games have been played in arcades and on home consoles, racking up virtual mile after virtual mile. Well now some firms are using this as the basis for testing autonomous vehicle systems.
UK tech company rFpro, has launched a platform for the development of autonomous vehicles in a virtual world. And the computer system should help autonomous vehicles to learn how to avoid accidents by correctly interpreting the road scene detected by its sensors without having to take the cars on public roads.
The company originally developed the technology for Formula One teams and has since been adopted by the world’s largest vehicle manufacturers for its ability to accurately replicate the behaviour of a vehicle.
“The company started, completely by accident, when some simulation software I gave away free on the internet, ten years ago, attracted the attention of a Formula One team. We used software based on gaming technology to revolutionise driving simulation,” says rFpro technical director Chris Hoyle. “The autonomous vehicle market is expected to be worth up to $10 trillion, but debate is rising about whether these vehicles should be allowed on our roads, if not, how do we develop them? An evolution of our platform enables vehicle manufacturers to thoroughly test their technology and be absolutely confident in their systems before validation on real roads. The key to autonomous vehicle adoption is the development of the vehicle’s ‘brain’, its ability to make appropriate decisions, and that’s what our technology helps with.”
rFpro provides driving simulation software for deep learning autonomous driving testing and validation. rFpro is being used to train, test and validate deep learning systems for ADAS and Autonomous applications by both car manufacturers and their suppliers. One of the most important aspects of the system is that when developing systems based on machine learning from sensor feeds, such as camera, LiDAR and radar feeds, the quality of the 3D environment model is very important. The more accurate the virtual world is the greater the correlation will be when progressing to real-world testing.
Using a digital environment to accurately represent the real world, the technology enables vehicle manufacturers to test their systems in every scenario imaginable. The computer simulation models are built around a graphics engine that includes a physically modelled atmosphere, weather and lighting, as well as materials for every object in the scene. Add to this the 100’s of kilometres of public road models available, spanning North America, Asia and Europe, including multi-lane highways, urban, rural, mountain routes, and every test imaginable should be possible.
The key to these platforms, and rFpro is unlikely to be the only firm offering the technology, is the level of accuracy achieved replicating the real world in simulation. This enables the various sensors used for autonomous vehicles to react naturally and therefore test results are completely representative. rFpro has been producing a library of real roads created through highly-precise scanning technology, which forms the basis of the simulation. As it’s a digital platform, users have control of all the variables, such as traffic, pedestrians, weather and location, enabling them to test every eventuality.
“By using multiple computers, manufacturers can undertake millions of miles of testing every month using our platform,” says Hoyle. “Humans can also be introduced into the simulation, controlling surrounding cars or pedestrians, so we can assess an autonomous vehicle’s decision making and also the interaction between the vehicle and the driver, but most importantly it is carried out in a safe environment.”
And being able to introduce a human element is important, because while computers work logically humans are anything but logical. Taking control of a vehicle in the simulation for example and doing something out of the ordinary will force the autonomous vehicle’s systems to think, react and learn to every scenario possible.
So, the questions about if an autonomous vehicle had to choose between hitting a group of people or a single human what would it do, can be answered – although you’d hope that an autonomous vehicle would never allow itself to have to make such a choice in the first place.
The technology specific to autonomous vehicles has been developed over the last three years and has already been adopted by two major vehicle manufactures and three autonomous car developers. It is also being used by a driverless motorsport series.
“Autonomous vehicles will revolutionise road safety, much more than anti-lock brakes, automatic emergency braking or stability systems have done before it,” says Hoyle. “It has the real potential to create a largely accident-free road network. Allowing autonomous vehicles on to the roads is an essential part of the validation process but our platform enables all of the testing to be carried out in a completely safe environment. Further to this, it significantly reduces the cost and time required to develop these complex systems, bringing the vehicles to market sooner.”
To enable comments sign up for a Disqus account and enter your Disqus shortname in the Articulate node settings.