New bus system to improve pedestrian safety

A new system can improve bus safety by issuing a warning when a potential risk to pedestrians is detected so the driver can react accordingly, or bringing the bus to a halt if they don't.
The system developed by researchers at the Polytechnic University of Valencia (UPV) in collaboration with Spanish company Cognitive Robots, combines video image capturing technology and software that analyses the images and issues warnings to the driver.
The safety system was invented to help avoid accidents in area of the bus stop when the bus restarts its route.
The system incorporates different cameras placed at strategic points of the bus that allow the driver to see where the rear-view mirrors can't; a screen placed next to the vehicle wheel; advanced software for detection of people and different warning devices synchronised with the software.

According to the researchers, this system is especially interesting for the issue of child safety.
In this case, the traditional security systems present some deficiencies, since when the children are near the bus, they are in the blind spot of the rear-view mirrors, which implies a risk.

"Our unit automatically processes the images taken by the cameras and warns the driver even if he is not looking at the screen," said Leopoldo Armesto, researcher at the Institute of Design and Manufacturing of the UPV.
Besides the software, the researchers have developed a mechanism in the accelerator, which, if risk arises, can lock it; and a system that prevents the driver from spinning the wheel towards the area around the pedestrian.

"The system, at first, does not disable driving but when it detects a risk, warns the driver, slightly affecting the accelerator or the wheel. If the driver persists, it can generate an emergency lock," said Armesto.
In order to test the unit features, the researchers did different tests in a controlled environment, causing the bus to brake when it detected the proximity of a person while the vehicle was running.
In order to validate the software processing ability, the researchers created a big bank of images taken from different places in the city of Valencia where there is a significant pedestrian traffic.

"We used a total of 12,000 samples of people seen from a vertical position, the same view than the bus cameras have. This big image bank is what allows the system to identify and detect the presence of a person in front of the bus, warn the driver and, if necessary, take direct action on the vehicle," said Armesto.