Technical skill sheet

Real-time Perception and Comprehension of Situations by Multisensorial Fusion and Learning Techniques

  • Centre of research : MINES ParisTech, centre de robotique - CAOR
  • Prestation : Characterization, tries, tests, checks, formulation - Advice, assessment, training - Prototyping, Preseries, Industrialisation - Research and Development

Contact Expert


Development of sensor systems designed to increase the perception capacities of robotic modules :
- Increased robot autonomy
- Assistance of man, improved security
- Real-time use of data captured for modelling the environment.
Research themes : On-board perception and comprehension of situations in real time by multisensorial fusion and learning techniques.


Major applications in the road safety sectors :
- Detection of obstacles and other items by the vehicles
- Detection and visual recognition of road signs
- Telecommunications facilities for road safety and traffic flow management
Applications in the fields of videosurveillance : Surveillance for the safety of persons and property.


Methodology of fusion sensor grouping :
- Use of absolute and relative localisation sensors (GPS and SIG, inertial navigation unit, odometry) combined with a frontal camera for accurate localisation of the mobile
- Fusion methods based on Kalman filters (EKF, UKF)
- Combined use of cameras, laser telemetry and radar telemetry