Self-Driving Cars Need to Learn More About Humans- Says Nissan
At the 2017 Management Briefing Seminars, a presentation was discussed that talked about the possibility of self-driving cars needing a lot to learn before these can be allowed on the road.
Melissa Cefkin, Principal Researcher, human centered systems at the Nissan’s Research was of the opinion that there is a need for autonomous vehicles to properly understand and learn social behaviors such as the assumption that a pedestrian standing on the car’s left side is about to get in and drive and not necessarily walk into the street. This should be done before manufacturers can deploy them successfully.
Even though most human behaviorism related to driving is inherent, however, some are learned through experience. The challenge according to her lies in programming the norms as well as behaviors in a self-driving car.
Initially, when the autonomous cars start being driven, a lot of people will be experiencing them from outside, either sharing the road as a biker or a pedestrian or as a driver of another car.
Nissan is working on to address the education challenge for the EVs by analyzing the human driving behavior for determining which reactions that are natural and which are learned through experience.
When a differentiation can be made, developers will be able to teach what reactions to expect in the situations in which there is no adherence to the natural assumptions.