RAMON MORROS. Researcher at the Image and Video Processing Group (GPI), Universitat Politècnica de Catalunya · BarcelonaTech
Sustainable mobility in cities is a matter of vital importance. Public transport is clearly a key factor in the implementation of a more reasonable model than the current one, which is strongly focused on private combustion vehicles. However, public transport by land has some major problems that make it unattractive. Perhaps the main one is the duration of journeys, which can be much longer than in private vehicles. A relevant factor in the reduction of journey time is better interaction with private vehicles, which can enter lanes reserved for public transport at times of transport saturation. Other crucial aspects are the number of stops and the frequency of service. More stops enable an increase in granularity and proximity to the user, but also entail longer journey times. To correctly plan public transport networks it is key to have information to optimise all of these parameters.
The research group GPI@IDEAI of the UPC is working on computer vision-based solutions and artificial intelligence to obtain real-time information on aspects such as number of people at bus stops, average waiting time, and number of people who get on and off the bus. This information enables more informed decisions to be made in the design of networks. Data collection is always carried out with respect for privacy. Information such as people’s faces and vehicle registration plates is deleted. Our technology can also be used to determine the occupancy of reserved lanes by unauthorised vehicles. These tools can be of great help for local government to plan mobility properly in cities.
In addition to public transport, in recent years the use of micromobility, that is, bicycles, electric scooters and other similar vehicles, has increased considerably. These vehicles can complement public transport without polluting and increase the flexibility and versatility of journeys. Cities’ regulations are gradually being adapted to these modes, but many users are still not aware of them. This means that conflicts can arise between pedestrians and drivers and the number of accidents can increase.
UPC, through GPI@IDEAI and RSLAB, is a member of the consortium that is developing RideSafeUM, a project funded by the EIT Urban Mobility whose aim is to increase the safety of micromobility vehicles. Using a camera and geolocation systems, it can be determined whether a vehicle is circulating on the pavement, the road or the bike lane, and the user can be notified in real time if the regulations are being violated on the specific path on which they are travelling. For example, the system will detect that a vehicle is travelling on the pavement and will inform the user that this is prohibited. In addition, potential accidents can be detected through the use of an accelerometer, and emergency services alerted if necessary.
Finally, anonymised information on journeys made, potential infractions and accidents is sent to a centralised server where government bodies can gather information on the use of microvehicles that can be used to improve the urban mobility networks. Notably, the system is never used to fine users, as the data are anonymised before they are sent. Currently, pilot tests on the system are being carried out in three European cities: Barcelona, Rome and Thessalonica. The RideSafeUM solution should be available on the market from 2023.
In addition to these mobility projects, GPI@IDEAI researchers are investigating other related aspects using computer vision and artificial intelligence such as medical images, satellite image analysis, precision agriculture and analysis of sports.