Tallinn buses help detect missing and damaged traffic signs

Tallinn University of Technology researchers have equipped three city buses with sensors and artificial intelligence powered camera systems to identify missing, poorly visible and damaged traffic signs, ERR reported. The project aims to improve oversight of the signs and help strengthen road safety.

Tallinn has about 50,000 traffic signs, making regular inspections a resource intensive task. Signs are currently checked by people, and identifying defects can take weeks.

The sensors installed on the fronts of three buses collect visual information about traffic signs as the vehicles travel through the city. Buses can cover a large share of Tallinn’s streets, but they do not travel everywhere where information may be needed, said Uljana Reinsalu, a senior researcher at Tallinn University of Technology and the project’s lead.

Researchers are therefore considering other types of vehicles that could provide better coverage, including garbage trucks, Reinsalu said.

The data is processed inside the vehicle using artificial intelligence, and only the necessary results are sent to the city. Because the video also captures other road users, the researchers consulted the university’s research ethicists and Estonia’s Data Protection Inspectorate. The current algorithms detect people and vehicle number plates and automatically blur them, Reinsalu said.

Possible future uses

The same platform could eventually be used to detect other problems in the urban environment, such as damaged road surfaces and street lighting failures. Information about air quality is already being collected, Reinsalu said.

The researchers aim to show how mobile sensors and artificial intelligence can make urban management more efficient, data based and sustainable. Urban environment information is currently scattered across different locations, Reinsalu said, adding that the project seeks to organise monitoring so city employees can access information more quickly and use it to find solutions and make decisions.

The researchers will collect data from city streets until the end of November. They will then assess how accurately the system identifies traffic signs and how well the solution fits Tallinn.

The project, TayuTech, is an international initiative led by the FinEst Centre for Smart Cities. In Dublin, Ireland, the project measures air quality, while the Croatian city of Pula is also involved.