Three Successful Pilots to Improve Public Transport User Experience
- As part of the Antitrash project, three pilots took place during 2021 in the cities of Hamburg, Zalaegerszeg, and Barcelona.
- An Artificial Intelligence (AI) computer vision algorithm was designed to process the images collected by cameras and identify trash or potential damage inside vehicles.
The goal of the project is to revolutionise user experience and maintenance processes of shared and public transport vehicles. With this in mind, the Antitrash project put to test a system developed to identify trash or potential damage inside a vehicle as well as unpleasant odours. In order to assess the designed algorithm, pilots were conducted in the cities of Hamburg, Zalaegerszeg, and Barcelona, which served as proof of concept of the solution in different environments and means of transport (metro, tram, and bus).
The consortium created a novel system to automatically assess the cleanliness level inside a vehicle. This system relies on a computer-based solution and an air quality control module to detect litter, broken items, or displeasing smells. CARNET played a key role, together with Aalto University, in developing the AI computer vision algorithm that processes the images collected by cameras. In addition to trash identification, the algorithm also had the capacity to detect valuable objects that could be left behind by users (e.g. smartphones or laptops).
Regarding the Barcelona pilot, CARNET ran the demonstrations together with its industrial partner TUGSAL. Trials took place from October to December 2021 and tested the computer vision feature on a pilot bus on line B23 (linking the cities of Barcelona and Badalona). The pilot bus was fitted with four cameras and an industrial computer to capture and pre-process images. Pictures were regularly taken and sent to the Antitrash server at Aalto University, where AI would assess the presence of items of interest (trash or valuable objects). In addition, image lighting conditions were also evaluated to avoid misdetections. Furthermore, compliance with GDPR was guaranteed by analysing images with high-accuracy human-detection systems that disregarded any pictures where people were present.
A remarkable aspect of the Barcelona demonstration consisted of the use of standard surveillance cameras. While other pilots relied on specific camera prototypes, the Barcelona case proved the feasibility of using the Antitrash system with commercial cameras, which are already installed in many public transport vehicles. This possibility reduces deployment and hardware maintenance costs associated with the system, making it more attractive for transport operators and other companies.
Following the gathering of valuable insights resulting from the pilots’ stage, the project’s conclusion was marked with a meeting organised by SEAT S.A., which took place on 14 December 2021. The aim for the future is to create a spin-off company to bring the Antitrash solution to the market.
This project has been funded by EIT Urban Mobility.