Kishar Ahmed
Over the last 4-5 years, self-driving cars have at last started to become a reality after a decade of prototypes and experimentation. Toyota and Uber have joined hands to launch driverless cabs. Tesla, to Mercedes, all major car brands at least have some kind of autopilot program in development if not a fully self-driving car.
Uber crash in Arizona
But how safe are these vehicles? Last March an Uber self-driving car killed a pedestrian in Tempe, Arizona. To understand if they are safe we have to first understand how self-driving cars work. Most such autonomous vehicles have a GPS and several types of sensors – from the laser to radar as well as video. These help the car understand where it is and where everything else is relative to it. The car using augmented intelligence system creates a 3D map of its location and surrounding objects and using the map maneuvers with help of steering, throttle, and brakes.
Limited by its OS and sensors
It is easy to understand that such a system in self-driving cars is only as good as the operating system and the power of its sensors, which are at this moment far from perfect development. No one can really say if they can drive through every kind of storm. The data simply is not there so far. There is also the question of decision making e.g. – if the car has to swerve and crash into a small car with one adult driver or a small van with six children then a human driver would choose greater good – to crash into the adult driver. A machine at this moment is incapable of learning such complex thoughts.
Safe in a controlled environment
But at least self-driving cars will never get tired or distracted. That makes them safer than human beings. In closed communities or inside small communities which are self-driving cars compliant they will surely be safe. But out in the world with literally thousands of variables, they still have a lot to prove.
