Search
Close this search box.

Luna Systems

Computer vision enabled practical interventions for the reduction of errant rider behaviour

Project summary

Errant shared e-scooter behaviour, including disorderly and illegal parking which creates trip hazards/obstacles, has become an issue in Istanbul and many cities across Europe. This project aimed to prevent disorderly and illegal parking of shared e-scooters, so that cities can confidently introduce and scale public shared micromobility schemes.

Luna Systems’ computer vision technology provides greater fleet control by determining where users are riding and employs incentivisation/gamification to reduce disorderly parking. Luna Systems leverages an artificial intelligence solution that automates the ‘end of ride’ photo verification of correct parking, and ensures riders cannot end their ride until they have parked correctly.

Luna Systems worked with Turkish shared e-scooter operator ‘hop’ to develop an API approach to further enhance accuracy of parking detection. The solution algorithm is tasked with providing accuracy alongside the following validation points: Is there a scooter in the image? Is it a ‘hop’ scooter? Is there more than one ‘hop’ scooter? Is it parked in a bay? Is the kickstand down? Is it locked?
With the Luna Systems solution, it is possible to reduce incidents of disorderly parking, which in turn reduces the amount of time operators need to allocate resources to recuperate the wrongly parked vehicles. Additionally, the solution makes more e-scooters available for end-users.

With scaling, the solution will be able to provide shared micromobility operators with data insights into patterns of behaviour that can further support decision-making.

Project start:

1 June 2023

Budget:

€59,500

Context

The project aimed to solve errant e-scooter rider behaviour, so that cities can confidently introduce and scale public shared micromobility schemes.

Challenge

The project aimed to address disorderly and illegal parking of shared e-scooters creating trip hazards and obstacles.

Expected outcome

The project resulted in the use of computer vision technology for greater fleet control by determining where users are riding and use of incentivisation/gamification to reduce disorderly parking.

Project partners

Silver
France

Ynstant

Silver
Spain

RMIT Europe

Silver
Spain

Parkunload

Project Lead

Maria Diviney

maria.diviney@luna.systems