Autonomous cars are a combination of highly advanced sensors and an operating system designed to handle a large level of data input, the ability to fully process it, and execute the correct actions to be taken. The combination of these technologies is broken into five categories of autonomy as defined by the NHTSA. These run from L0, where the driver has complete control over the car, up to L4, a fully self driving car that requires no driver input.
When manufacturers developed cruise control, a system that monitors and automatically adjusts the throttle position, they had achieved a level of L1 autonomy. With the introduction of Lane Keeping technology, utilizing radar systems to detect lane markings and applying a counter-steering force when crossed, the simultaneous use of both of these technologies raises the level to L2 autonomy.
A number of manufacturers are currently working on implementing L3 autonomous cars, which involves the car maintaining all safety-critical functions requiring limited driver intervention. This requires a system to control and adjust the throttle and the steering shaft based on analyzing the environment while utilizing a location and mapping service to arrive at a destination.
Overlapping sensors are required to capture all the environment data. Sensors need to detect environment changes up to 250 or more meters away and with finer detail at fewer than 80 meters. Different technologies are being assessed and each manufacturer is using a combination of very accurate LiDAR systems, multiple ultrasound radars, lasers, and cameras. Each sensor system has a specific range of tasks that they excel at, from processing camera images to detect traffic signs to LiDAR systems generating a 3D environment.
All this data requires a powerful computer to process it. NVIDIA has announced their Tegra K1 processor will be used in Audi’s autonomous car, a 192-core GPU similar to processors used in super computers. Each manufacturer is also utilizing different operating systems customized for autonomous driving (see chart below).
Autonomous cars are predicted to reduce 90% of traffic accidents, improve fuel efficiency, reduce pollution, and provide mobility to non-drivers. Enterprises that currently provide transportation and related services must be prepared to pivot, i.e. taxies services, freight trucks, and delivery services. Other enterprises must be prepared to take advantage of utilizing fully automated transportation services and the technology that comprises them.
Future infrastructure must be carefully planned out to prevent it becoming obsolete and incompatible with new traffic systems. Car designs will be changing as drivers no longer are required to maintain vigilance on road, allowing the creation of mobile office spaces. Consumer hardware and software should begin investigating designs now to position themselves to become a technology leader within the developing space.
|Nissan||Leaf||Microsoft||Windows Embedded Automotive|
|BMW||Project Manager||Dr. Werner Huber|
|Google Fellow||Sebastian Thrun|
|Audi||Project Manager||Dr. Bjorn Giesler|
|General Motors||Innovation Program Manager||Jeremy Salinger|