Engineers develop 'smart' stop sign to warn drivers and self-driving vehicles approaching junctions

'Smart' system uses an infrared sensor to detect the velocity and trajectory of vehicles as they approach the stop sign

Researchers at the University of Texas at San Antonio (UTSA) have developed a low-cost smart stop sign, powered by an integrated solar panel, that they claim can detect vehicles as they approach an intersection.

In the US, about half of all traffic fatalities occur on roads in rural areas, despite those roads carrying less than half of the total traffic in the country. In 2015, a total 35,092 people died on road accidents, with rural areas accounting for 54 per cent of the fatalities. A large number of accidents happen as drivers fail to notice stop signs at road intersections in time.

The technology developed by UTSA researchers is basically a self-powered thermal system, which uses a multi-pixel passive infrared sensor to detect the velocity and trajectory of vehicles as they approach the stop sign.

The sensor captures the thermal signature of the passing vehicles and also determines their type or classification. The system then alerts drivers about approaching vehicle by flashing a red light above the stop sign.

"It distinguishes the vehicle's direction of travel, estimates the velocity of its thermal signature and determines the classification of the vehicle," said Zachary Balcar, a master's student in the UTSA Department of Electrical and Computer Engineering.

According to the research team, the flashing red light gives drivers ample time to safely stop their vehicles.

In the tests carried out so far, the smart system demonstrated an accuracy of 90 per cent in detecting vehicles, which can be considered better than technologies currently in use, including microwave radar, video image processors, and magnetic loop indicators. The system also showed a vehicle classification accuracy of 72 per cent in the tests performed.

The smart system consumes less power to operate, according to researchers, and should be less expensive to produce on a large scale. While existing systems may cost up to $5,000 per unit, the new vehicle detection system would be available for less than $100 per unit.

The researchers also reveal that they are planning to adapt the technology for pedestrian detection and for vehicle-to-infrastructure communication.