Researcher develop AI warns pedestrians wearing headphones on the street

Researcher develop AI warns pedestrians wearing headphones on the street


Today, many people wear headphones in public spaces. To limit accidents as much as possible, a researcher based in the United States has developed a prototype audio alert system boosted with artificial intelligence (AI) intended for this type of user.

AI in the helmet
Xiaofan Jiang is an assistant professor of electrical engineering and a member of the Data Science Institute at Columbia University in the United States. In a report published in the IEEE Internet of Things Journal (vol 6) in October 2019, the applicant presented a prototype of an innovative audio alert system. This is to warn pedestrians wearing headphones on the street. Remember that wearing a headset with music does not guarantee safety in public space. The risks of collision with vehicles are indeed much greater due to the decrease in vigilance.

According to the researcher, there is talk of a system with AI directly integrated into the pedestrian helmet and connected to a dedicated smartphone application. The algorithm is capable of interpreting the surrounding sounds and alerting the user if necessary. It turns out that the system is able to detect a vehicle at a maximum distance of 60 meters! This logically gives the person time to free himself and thus become safe.

A system for the moment limited
You should know that this system works with the sensors fitted to a good part of the most recent vehicles. These therefore make it possible to warn the pedestrian. On the other hand, the system is not able (yet?) To predict the trajectory of the vehicle potentially representing a danger.

There is also talk of an inexpensive solution. Indeed, a series of microphones fitted to the prototype would not increase the price of the helmet! More specifically, we are talking about four sensors placed at different locations on the helmet in order to optimize detection. Note also the presence of a cheap circuit board. The latter is dedicated to extracting the most relevant sounds.

Finally, the project leaders are aware that such a system could greatly annoy users. Indeed, if the music turns off every time a vehicle approaches, the utility is no longer really there. However, Xiaofan Jiang said that the alert would only be given in the event of a real risk of collision.