DC11. AI for proactive safety detection using conflict techniques

Hosts: UH & CEGEKA

Objectives:

• To propose improved AI techniques for the analysis of traffic conflict video data in terms of correct road user type classification vs. privacy perspectives
• To combine, using AI techniques, traffic conflict indicators for a holistic safety assessment
• To translate the results from a conflict observation analysis into suggested safety improvements, via AI-based decision mechanisms To forecast the achievable reduction in conflicts for different potential infrastructure changes

Expected results:

• AI software with privacy-by-design for improved analysis of conflict video data will be developed and validated
• A holistic safety assessment methodology will be developed based on the fusion of multiple surrogate safety indicators
• A translation of conflict indicator results into actionable safety improvements on location will be proposed

Planned secondment(s):

Citymesh NV –BE, Purpose: to be involved in customer projects, to experience built end-to-end AI solutions. The office has agreed to host the researcher; formal arrangements to be made during the project preparation.