Need to Stop a Driverless Car? Try Waving Your Hand

Home > News > Content

【Summary】This technology contains a lot of underlying issues that must be fixed before it can be applied to self-driving cars.

Original Michael Cheng    Feb 18, 2017 12:00 PM PT
Need to Stop a Driverless Car? Try Waving Your Hand

Self-driving cars are being designed to view the world like humans. To achieve this level of awareness, the modern vehicles must be able to detect body language and subtle changes to their environments, such as a biker leaning to one side before making a turn or eye contact and a slight nod when acknowledging and allowing a vehicle to pass through from a crossing lane.

To humans, there's no need to "learn" about these everyday cues. Since birth, people have become naturally acclimated to processing hints about their surroundings. Researchers at the Royal College of Art and Imperial College London are working on applying such complex perceptions to driverless platforms. Called Blink, this technology is powered by machine learning, as it picks up common gestures from pedestrians. In application, an individual crossing the street could stop a self-driving car just by raising his or her hand.

Machine Learning and Sensing

There are two primary actions associated with this technology. The first part involves detecting and processing natural gestures, which could be foot tapping, hand raising, eye scratching, or arm swaying. Through machine learning, Blink learns the meaning of these actions and their applications to interacting with cars. We know that machine learning can be used to classify content quickly and efficiently, using monumental data sets. The level of accuracy is almost surreal, which allows technical, algorithmic learning to go as deep as detecting differences in culture-based and gender-based gestures.

Next, a sensing feature powered by LED components is used to notify pedestrians about the car's reaction to the gesture. Mounted on the windscreen and rear window of the vehicle, the system leverages robust light signals and auditory notifications to acknowledge the human-initiated signal. In application, a raised hand would cause the driverless car to flash a green light, suggesting that it is safe to cross the street. While a side-waving motion could cause the vehicle to flash a red light, indicating that it is not safe to cross the street and the individual should wait for the car to pass through.

Humans are Complicated

This technology contains a lot of underlying issues that must be fixed before it can be applied to self-driving cars. Giving pedestrians the power to stop vehicles is very controversial, as the action can easily be exploited for crimes. Furthermore, cars should not stop for every individual on the street waving his or her hands in the air – this is very unproductive and may cause passengers to become frustratingly late for their personal commitments (imagine how confusing this would be during peak hours, with swarms of people gesturing vehicles to stop). So far, the company has not tested their gesture-based device on an actual self-driving car, which speaks volumes about how difficult it would be to actually use this type of sensing component in real-world environments.

"Aside from the potential of autonomous vehicles to cause far fewer accidents and fatalities in the urban environment, their arrival also provides an opportunity to rebalance the power dynamics and give pedestrians an equal weighting in the conversation between pedestrians and vehicles on the road," said Raunaq Bose, one of a group of four researchers at Imperial College London and Royal College of Art who developed the design.

Prev                  Next
Writer's other posts
    Related Content