| The COVID-19 pandemic has created significant restrictions to passenger mobility through public transportation. Several proximity rules have been applied to ensure sufficient distance between passengers and mitigate contamination. In conventional transportation, abiding by the rules can be ensured by the driver of the vehicle. However, this is not obvious in Autonomous Vehicles (AVs) public transportation systems, since there is no driver to monitor these special circumstances. Since, AVs constitute an emerging mobility infrastructure, it is obvious that creating a system that can provide a sense of safety to the passenger, when the driver is absent, is a challenging task. Several studies employ computer vision and deep learning techniques to increase safety in unsupervised environments. In this work, an image-based approach, supported by novel AI algorithms, is proposed as a service to increase the COVID-19 safety rules adherence of the passengers inside an autonomous shuttle. The proposed real-time service, can detect deviations from proximity rules and notify the authorized personnel, while it is possible to be further extended in other application domains, where automated proximity assessment is critical. |
*** Title, author list and abstract as seen in the Camera-Ready version of the paper that was provided to Conference Committee. Small changes that may have occurred during processing by Springer may not appear in this window.