- Date of entry : 07/10/2026
- Associated laboratory(ies) : Centre INRIA d’Université Côte d’Azur et LISPEN (laboratoire d'Ingénierie des Systèmes Physiques et Numériques) - Arts et Métiers Sciences et Technologies de Chalon-sur-Saône et Lille.
HawQ-ai builds a simulation platform to design and validate camera and AI perception systems before they are
deployed. Today, a video surveillance system is sized on a floor plan, with no evidence that the cameras, and the
analytics running on them, will actually see what they are meant to see. HawQ-ai turns the floor plan or a 3D
capture of the site into a digital twin, simulates the physical behaviour of the sensors in it, and measures the
performance to expect under real operating conditions.
The first application addresses the security of sensitive sites (airports, critical infrastructure, industrial facilities):
optimising camera placement, verifying the installation, auditing existing systems. The second addresses AI
security: testing how robust detection models are against attacks designed to fool them.
Born from PhD research on digital twins and matured within Inria Startup Studio, the project brings together
expertise in computer vision, 3D simulation and cybersecurity.
