Development of a real-time monitoring tool for an autonomous vehicle fleet within a research project.
This European research project aimed to validate the feasibility of setting up an autonomous vehicle fleet in a real situation:
- Bring users from a station to a shopping center,
- Move users around the shopping center.
Many different actors have been involved (industrialists, academics, private digital companies) to propose a realistic and coherent solution to these constraints. Each actor brought his own solution to arrive at the final project.
Two types of vehicles have been integrated, a passenger car (Renault Zoé) for trips from the station and a tram-type shuttle for trips in the shopping center. This project was carried out in collaboration with Exoskills who outsourced the technical architecture, development and management of the team to us.
We were mainly involved in the Backend part, the objective of which was to provide an API that would allow a management interface to work properly. Because of the real-time aspect of the handled data (the sensors integrated into the vehicles transmit information very regularly) the usual solutions did not meet expectations (database exposed through a standard HTTP API). We have chosen to use WebSockets to communicate vehicle information (position, indications) to the interface in real time. The Backend therefore exposed two different APIs: one for “static” data and one for real-time data.
The city of Rambouillet, sponsor of the project, made it possible to carry out tests in real conditions (on a closed road). We were thus able to validate the operation of our solution. The project will ultimately not result in the installation of vehicles because of the human risk. However, there will remain a great challenge both on the collaboration with all these actors and on the atypical side of the solution to be proposed.