AI technologies applied to image analysis address the field of passenger and goods transportation, both for the autonomous vehicle and for assistance and optimization of multimodal transportation.

AI for autonomous vehicles

CEA LIST has been developing technologies dedicated to autonomous vehicle and its ecosystem.

Perception of the environment for autonomous vehicles and ADAS

  • Deep MANTA

Deep learning algorithm for detection of road users in 2D and 3D, traffic lights, traffic signs and ground markings, free space semantic segmentation and depth map prediction

YouTube video
  • V-SLAM

Indoor geolocalization by vision: demonstrator of level 4 autonomous bus for intelligent parking in RATP underground bus central station

YouTube video

Perception of the road environment with cameras embedded on the infrastructure

YouTube video

Driver and passenger monitoring in autonomous vehicle and shuttle

  • Concentration level and emotion recognition

  • Position, activity of passengers, interactions between passengers

  • On-board events

User experience and innovative services

  • Customized information in natural language

  • Additional information displayed in augmented reality

YouTube video

AI for sustainable mobility

Pedestrian and cyclist counting with wide angle stereoscopic camera

CEA LIST develops for Eco-Compteur computeur vision technologies for real-time users counting at the service of active transportation.

YouTube video

Smart management of parking lots

Carpooling detection

AI for multimodal mobility : MaaS

MaaS (Mobility-as-a-Service) is a service platform to plan and book travel and to pay for multimodal journey, with the aim to reduce personal vehicle use and solo driving.

CEA LIST technologies apply to MaaS area and to new service development to ease multimodal transport and connections of vulnerable travellers in underground train station and in airports areas.

  • Guidance and vision-based localization

YouTube video

Natural Language Processing for quality and system engineering management process

LIMA is CEA LIST natural language processing tool managing 60 languages and dialects. It enables automatic analysis of large documentary base and automatic generation of reports in natural language.

Regulatory watch for new regulations and developments: automatic information retrieval and structuration

Opinion analysis on social media and customer satisfaction survey

  • Information retrieval

  • Automatic report filling

  • Criticisms and potential root causes analysis

Automatic vocal report

  • Automatic creation of quality survey analysis