Pedestrians pose problems for self-driving cars in NYC

General Motors plans to become the first company to test self-driving cars in New York City.

Bart Selman, an expert on artificial intelligence technology and professor of computer science at Cornell University, says the GM announcement illustrates the fast pace of artificial intelligence development – and will serve to accelerate the introduction of more self-driving technologies.

Bio: http://www.cs.cornell.edu/selman/

Selman says:

“This really shows how fast the self-driving technology is moving. GM was not much of a player until recently but car companies are realizing that self-driving cars will mean a significant change for their business models. Traditional car companies need to adapt and compete in terms this technology or risk becoming obsolete.

“These developments are driven in part by continuing rapid improvements in computer vision and general artificial intelligence technologies for self-driving cars. Many aspects of the sensor and vision components are reaching superhuman levels. The cars are designed at this point to be overly cautious. In cities, speeds are quite limited and simply having good obstacle detection of human and non-human objects with fast-braking capabilities can make the cars quite a bit safer than any human-driven car.

“One issue that will need careful attention in New York City is how the cars deal with pedestrians simply stepping in front of the car. Cars right now hit the brakes when an object appears in front of them but this could become a problem with pedestrians ‘bullying’ self-driving cars. Pedestrians don’t step in front of human-operated cars because they are not sure the driver will stop – but self-driving cars will not have a choice but to stop.

“Assuming the experiment goes well, this will be the start of the phasing in of self-driving cars into New York City, which will impact a large number of taxi, Uber and Lyft drivers. Interestingly, having self-driving cars in full operation for commuting may actually boost the overall employment opportunities in a city. People will be able to work during the ride and commuting itself may become more efficient.”

————————————————-

Art Wheaton is an automotive expert with Cornell University’s School of Industrial and Labor Relations. Wheaton says autonomous vehicle testing success is critical for General Motors, as many auto industry players are anxious about the future of transit in New York City.

Bio: https://www.ilr.cornell.edu/people/arthur-wheaton

“General Motors is moving ahead with plans for autonomous vehicles. Choosing New York City is appropriate considering the density of population and likelihood of fleet sales. The important safety feature mentioned in the article is limiting the range or target zone of testing to a 5-mile radius. The complexity of mapping out large areas of the country and factoring all of the construction and other foreseeable objects are huge barriers.

“Autonomous cars need to have an incredible amount of data and information. Limiting the area to a heavily congested part of New York City is also problematic. Pedestrian, bicycle, individual cars, and mass transit busses make this an incredibly difficult scenario. GM must have a high degree of confidence in their Bolt fleet of test cars to attempt this test.

“If they are successful it improves their reputation and puts them near the top as a high-tech firm. Failure on the other hand would be a significant setback. There are many players involved with competing agendas such as Uber, Taxis, auto competitors, and limousines that are anxious about the future of urban transit solutions.”

 

For interviews contact:
Daryl Lovell
office: 607-254-4799
cell: 607-592-3925
dal296@cornell.edu

 

Cornell University has television, ISDN and dedicated Skype/Google+ Hangout studios available for media interviews.

 

– 30 –