According to self-driving vehicle pioneer Raj Rajkumar, who works at Pittsburgh’s Carnegie Mellon University, humans will retain control over their vehicles for a long time from now on, as driverless cars evolve step by step.
But that is not the view of Chris Urmson, leader of Google Inc’s self-driving car program since 2009, who believes a completely automated vehicle should require no input or intervention from humans to make it safer. And he pledges the models to be ready for production series by the turn of the decade. Urmson is one of Rajkumar’s former colleagues at Carnegie Mellon and he believes the partially automated vehicles already ready to be deployed can mitigate some accident risks but also introduce new safety challenges. “The better the technology gets,” he explains, “the less reliable the driver is going to get.” His issue is with drivers that would be inattentive exactly when the automated system suddenly asks for human input and decision making.
And Google’s view on the automated vehicle is vastly opposed to the traditional auto industry’s approach on autonomous technology that calls to be introduced in so-called “baby steps” in vehicles drivers will buy in two to three years. They push down the path proposed by Rajkumar. Autonomous driving will come in stages, with automakers from GM and Mercedes to Tesla testing systems that in certain conditions can handle driving on their own – on the highway or in parking lots.
The automated parking feature is already here and numerous safety systems include automated brakes – especially in the city and at lower speeds – while others correct the steering to keep the specified lane or a preset distance from the vehicle in front.