TRAM Barcelona Improves Services with People Counting Solution

Barcelona is one of the most visited cities in Europe with people coming from all around the world for tourism or to experience events like football matches or the Smart City Expo World Congress. One thing is clear with so many people arriving in a city every year sometimes services can suffer the consequences. Tram Barcelona knows this as they've experienced days in which their stations are packed. meaning stops take longer and safety is compromised. In terms of safety, sometimes crowd in our platforms can be a concern because, as you can see, our platforms, they are a bit narrow, let's say like that. And if there's a lot of people, can become like a safety issue. Also, we want to improve our user experience. So how can we help them have a better, use of the tram and if it's crowded let's we want to find solutions after identifying the issue tram set up a search for a connected smart solution and was able to move quickly tram is a private operator made up of companies like movendia or alstom that works on concessions so unlike a public company it doesn't need a tender process they found crowd scan analyzed their solution and were able to deploy it very quickly for proof of concept or POC in December. So the end of 2022, we went to an exhibition in Malaga and Tram Barcelona, the innovation manager, Enzo came to us. He was looking for a new technology to measure the amount of people waiting on the platforms and was also looking for something that worked real time and was privacy friendly. So he didn't have to install cameras, et cetera, on the platforms. So we met and then already a month later in January. we were up and running in Barcelona. We wanted to give it a try, we wanted to do a proof of concept and see if the measures were correct, if the system worked and if we could do something out of it. The companies installed the solution on two stations, one with side platforms and one with a central platform. They installed some 10 sensors per platform which were positioned both at hip level and on the roofs. They also do a lot of work the gateway per station. For real-time data the sensors measure the amount of people on the platform every 15 milliseconds and the gateway sends the data every 10 seconds to a cloud-based platform. The way it works we have hardware and these are sensors which we place around an area that you want to monitor and these sensors they will emit an electromagnetic signal and what we do is we measure the signal strength in between all sensors when an environment is empty people move in this area because of their body mass they will weaken the signals and there is a correlation between the number of people and these weakenings it's a mesh network that is set up and we use that seven so it's a bit like lora one or sick fox to steer the network while enzo and i were at the station i realized the platforms already had cameras so knowing that with the crowd scan solution there was a need for 10 sensors per platform for people counting there was an obvious question that arose why not go for computer vision which would have used already existing infrastructure in each stations we have cameras located because of course our control center has to have like a vision of what's happening in our network but the thing is that cameras sometimes are located in certain area in certain places that we cannot see the platform when the tram it's in the stop so this Our solution has this issue, the issue that sometimes we cannot have the full vision of the platform and also that we wanted a solution that was privacy friendly for our users. As they've ended the POC stage, the companies are now looking to deploy more sensors in around 15 other stations. With the learning process complete, Tram has two outcomes in mind for the data collected, which at the moment they visualize on a Crowdscan cloud-based platform. but which in the future could be integrated with other data from other solutions they have deployed, like the overhead cameras they have on the tram doors to monitor the number of riders. We have live data, so in terms of live data, what can we do? The control center can send personnel, inspectors, operators to come to the station and help dispatch, help distribute the people along the platform. We can send automatic alerts through the public address system and on the other hand we have the historical data where we can recap all data and maybe we can anticipate. We can know that for example in a certain platform we can send in that certain time of the day a double configuration tram so there is more place for the people to you enter the tram so the tram is not that crowded for the full implementation the companies are looking to tackle some issues like how to avoid vandalism as devices have been stolen and damaged by passerbys they're considering a change of color and a more fixed installation like screwing it in tram is also looking to make it easier for batteries to be changed as depending on the model they last between one to three years these changes will mean that the poc has led to a more adapted solution we are seeing more and more of these projects come to life gathering mobility data to improve services as there is great potential in this type of data something you'll be able to learn next week as we'll cover the topic at the smart city expo world congress in barcelona