Automotive

Nvidia to help Toyota develop brains for autonomous cars

Toyota and Nvidia have teamed up to deliver self-driving cars
Toyota and Nvidia have teamed up to deliver self-driving cars
View 1 Image
Toyota and Nvidia have teamed up to deliver self-driving cars
1/1
Toyota and Nvidia have teamed up to deliver self-driving cars

Building a car is one thing, but developing all the requisite software systems is another – especially if that car is designed to drive itself. That's why Toyota has teamed up with Nvidia to develop a system capable of handling the huge amounts of sensor data that autonomous cars need.

It goes without saying, autonomous cars will generate huge amounts of data. Cameras, combined with radar and LiDar sensors are all necessary to work out what's going on with the traffic, and all that information needs to be processed and understood within fractions of a second.

Toyota will be working with Nvidia to develop the necessary "Drive PX AI" computing platform for autonomous driving. The artificial intelligence platform will be used to better break down the massive amount of data produced by the sensors on its cars, and use it to develop self-driving systems that work in a broad range of scenarios.

"Toyota has worked on autonomous driving technologies for over 20 years with the aim of reducing traffic fatalities to zero as an ultimate goal, achieving smoother traffic, and providing mobility for all," says Ken Koibuchi, Executive GM at Toyota. "Through this collaboration, we intend to accelerate the development of autonomous driving systems that are even more safe and capable."

Source: Nvidia

3 comments
Bob Flint
Currently the processors and computers are not fast or compact enough to run through all the data in fractions of a second. The systems we have built for Uber are trunk size, and would really need to be 10 times faster and more voluminous, and more expensive. not within the cost target or size of the average consumer vehicle. Maybe an autonomous truck that has room and power and could justify the added costs, compared to a human driver, running virtually 24/7.
SteveO
@Bob Flint, the way you wrote that it sounds like you work on the Uber autonomous car project. If you investigate the Nvidia technology being discussed, you will discover the power, form factor, and cost have already gotten to where it is needed for this. Of course this system only works efficiently with machine learning algorithms.
MarcJackson
These platforms are also applicable to fixing the internal combustion engine, they can use a physics engine to calculate virtual sensors to give the ECU better information to more accurately control the engine and accessories. They fixed the Hubble deep space telescope this way. Combined with better intake valves that don't obstruct the flow into the cylinder you can get better cylinder filling due to gas compressibility and higher turbulence. Cylinder pressure sensor either real or virtual combined with Jet ignition for ultra lean combustion as used by Ferrari and Mercedes in F1 increases thermodynamic efficiency over 46%. Combine this with hybrid EM motor and Graphene super capitors​ with Li-ion energy storage a single cylinder engine can be ideal for a motorcycle. This is what I want, so I'm building it in Oz using our Innovations​. Rebooting our vehicle manufacturing using modern methodology now we have dumped our Auto industry that won't Implement regulations unless you subsidise it because it is to big to fail and anti Innovation stance