The trolley problem: the answer is a smarter autonomous car
A popular scenario is the “trolley” dilemma, conceived in the seventies and used extensively in discussions about future autonomous cars. In the original version, we are asked to imagine someone driving a trolley headed for five people standing on tracks who will be killed if the trolley proceeds on its current course. We cannot stop the trolley by any means, but it can be turned onto an alternate track where it will kill one person instead of five.
While quite a lot of research considers trolley dilemma to be a model case to map cognitive responses, does it really apply to the incoming reality of self-driving cars?
One of Volvo’s leading experts, Jonas Nilsson, thinks that the technology should be even smarter. “These cars will avoid getting into such extreme situations in the first place. Unlike humans, who can get tired and distracted, vehicles driving in autonomous mode will be designed to plan the route in such a way that the remaining risk of unsafe surprising situations is acceptably low”.
Handling surprising situations
Cars with autonomous capabilities must be able to consider their own operational limitations to handle surprising situations – and adjust their tactical behaviour accordingly. “For instance, when driving in the fog, the autopilot must adjust its speed to be able to stop within its visibility distance”, says Nilsson.
“With safety as a core value of Volvo Cars for nearly 90 years - and a vision that no one should be killed or seriously injured in a new Volvo car by 2020 - we argue that vehicles driving in autonomous mode shall have the responsibility not to get surprised in a way that any reasonably foreseeable accident, including the trolley problem, can even show up”, he explains.