In 2015, the United States saw 38,300 fatalities and 4.4 million injuries. Of these crashes 95 percent are caused by human error, with 41 percent of all human error fatalities are caused by recognition errors, according to the Department of Transportation. Recognition errors include inattention, distraction, or inadequate observance on the par of the driver. It’s easy to now understand the demand of making cars safer. Early estimates predict autopilot vehicles reduce injury related accidents by 43%. This is compared to miles driven per a crash rate.
The five levels of autonomy that are commonly used in the industry are defined as:
Level 0: Human control, what has been done all along
Level 1: Driver-assist in specific function like steering or acceleration
Level 2: At least one driver assistance system is automated, like steering or acceleration/deceleration(i.e cruise control)
Level 3: Safety-critical functions of the vehicle under certain traffic environments are under vehicle control. The driver can still intervene if necessary.
Level 4: The performance of safety-critical driving functions and roadway monitoring systems are under vehicle control for the entire trip. This is often called “operational design domain (ODD)”. ODD does not cover every driving video
Level 5: Fully-autonomous system that expects the vehicle’s performance to equal that of a human driver, in every driving scenario.
Right now, according to Elon Musk, Tesla is two years away from level 5 autonomy . “Once you solve cameras for vision, autonomy is solved; if you don’t solve vision it’s not solved…. You can absolutely be superhuman with just cameras” Tesla isn't the only company tackling this problem, almost every major car company is developing self-driving vehicle and the roads will soon be covered in self driving vehicles.