News

Deep learning-sensor fusion architecture for pedestrian and object detection in scattering media

Last April 18th Dr. Alfredo Chávez finished his two-year stay at CD6, completing successfully the TECNIOSPRING PLUS project he had been awarded on "Deep learning-sensor fusion architecture for pedestrian and object detection in scattering media".

Camera and LIDAR sensor have significant hindrances when working in scattering media, such as fog and smoke. Further, imaging radar sensors are available at reasonable prices and are well-known in automotive environments. Along his stay, Dr. Chavez implemented the development of imaging radar sensors in the data fusion framework already existing at CD6, fusing radar data with LIDAR point clouds and 2D imaging cameras in RGB and thermal, completing the sensor fusion suite which an autonomous vehicle is usually carrying out. His landmark work (presented in one article https://doi.org/10.3390/s23084167 and one conference communication) serves as setpoint for CD6 to further pursue research in multimodal sensing for pedestrian perception and support the development of advanced ADAS algorithms. His results will be directly applied to project USEFUL (TED2021-132338B-I00, "Percepción y sensado multimodal para vehículos autónomos), where a sensorized car is being developed at CD6 for advanced dataset generation. Research activities were transferred to a company (Beamagine) via a 3 month secondment.

Thanks, Alfredo, for your contribution and best of lucks!

This project has received funding from the European Union's Horizon 2020 research and innovation program under the Marie Sklodowska-Curie grant agreement No. 712949 (TECNIOspring PLUS), as well as from the Agency for Business Competitiveness of the Generalitat de Catalunya.