Earlier this year auto maker Nissan announced that they would have a level 4 autonomous car ready as early as 2020 for the complex streets of Tokyo. Nissan is working with NASA to accelerate technology development. Audi and video game hardware maker, Nvidia, announced a partnership to achieve the same goal. Audi continues to focus on designing cars while Nvidia builds complex artificially intelligent computer systems using all its learned from building graphics acceleration hardware for video games.
What are car makers promising? What does level 4 mean?
There are five levels of autonomous cars according to the Society of Automotive Engineers (SAE). We are currently at level 1 where semi-autonomous, driver-assisted vehicles are common in the market. Reaching level 4 means a car can automatically navigate complex roads and traffic conditions and only request driver intervention if the computer detects dangerous situations. Level 4 by 2020 is incredibly ambitious.
So where are we in the development of the level 4 autonomous car?
Just last month, Magna, the automotive parts and systems giant, announced the MAX4 level 4 autonomous driving platform. The beauty of this system is that it can be integrated into any vehicle without any compromise to vehicle design and styling. Remember the first Google cars had big and ugly radar sensors mounted outside the car. The platform can navigate in complex urban environments and on highways. Because the technology can be seamlessly integrated into any car design, it is possible to achieve production scale and volume. The MAX4 also combines cameras, RADAR, LiDAR and ultrasonic sensors with a computer platform that is upgradable. Magna says it has simplified the way drivers use autonomous driving systems by making the user experience as intuitive as cruise control.
Also Read: Honda Autonomous Driving set for 2025
In order to have autonomous cars navigate streets and highways, technology and car makers still have a few big problems to solve.
5G Networks and Vehicle to Vehicle Communications
Imagine you are travelling at 100 km/h and 1 km ahead an accident suddenly occurs. How will your car react? This could be a complex situation that requires driver intervention. But you are in auto pilot and catching up on your emails. You need at least 30 seconds to put away your phone and reorient yourself. The only way your car will be able to give you such advanced notice is for the cars in front to communicate to the cars behind of the accident. Imagine using the current cell phone technology and there is a delay in communications? This is why major telecommunications carriers are testing the 5G network. It promises to deliver almost zero latency and 10,000 times more network capacity. Telecommunications companies are aiming to have this ready for consumers by 2020.
Also Read: Cadillac Launches V2I Connectivity
Artificial Intelligence (AI)
Artificial intelligence is required in order for cars to make complex decisions. Millions of dollars are being invested in startups working on artificial intelligence with Toronto being a hotbed of talent in this field. Google has made their machine intelligence software technology TensorFlow open source and available to anyone who want to further research on machine learning and deep neural networks. Many car makers including Toyota have dedicated AI research labs.
High Definition (HD) Maps
HD maps are core for the autonomous car. It is much more than the two-dimensional maps we use on our smart phones or in car navigation systems. HD maps capture location content such as buildings, objects, traffic patterns accurate to a distance as close as 10 centimeters. Although, the technology exists, the content is still being created. The company, HERE International, has maps in nearly 200 countries, provides live traffic information in 33 countries and has indoor maps available for about 49,000 unique buildings in 45 countries. That is pretty good progress.
Comment below and let us know what you think of the autonomous car. A recent survey showed that 26% of Canadians can’t wait for the driverless car to be available. Do you think cars operating at level 4 capabilities by 2020 is possible?