COACHES ART-ADN Le numérique au service du handicap Travaux du laboratoire GREYC - Organisé par le CCAS - la Ville de Caen - le Dôme
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Le numérique au service du handicap Travaux du laboratoire GREYC COACHES ART-ADN Organisé par le CCAS – la Ville de Caen – le Dôme
CHIST- ERA/2014 COACHES COoperative Autonomous Robots in Complex and Humans EnvironmentsS Coordinateur : Abdel-Illah Mouaddib Partenaires :Unicaen (F), U.Sapienza (I), U. Vreije (B), U. Sebanci (T)
Motivation Deploying robots in public area to assist, guide, and escort visitors and surveillance Visual applications are born: Guidance and Personal assistance in public area Shop malls, touristic sites, airport, train stations, hospitals Surveillance and patrolling in cooperative way with intervention units (firefighters, police, medical units, and tourist office staff, …) Our target use cases: Caen Rive de l’orne MALL in cooperation with Caen city
COACHES OBJECTIVES OF THE PROJECT Public spaces in large cities are increasingly Assis ng Surveilling Informing Services Manager or becoming complex and unwelcoming Guiding People Monitoring the environment Shopkeepers environments. COACHES project addresses Sensor! Seman8c! Knowledge! Ac8ons! Sensor! Seman8c! Knowledge! Ac8ons! Processing! Interpreta8on! Update! Processing! Interpreta8on! Update! ! ! Sensor! Seman8c! Knowledge! Ac8ons! Software network! network! KB! ! KB! ! ! ! Processing! Interpreta8on! Update! ! ! ! ! ! ! ! Scene!Analysis!and!understanding!(WP2)! ! Knowledge>based!environment!! network! ! ! Scene!Analysis!and!understanding!(WP2)! ! Knowledge>based!environment!! ! KB! ! ! ! ! modelling!!and!reasoning!(WP1)! Mul= >robot! ! ! ! ! modelling!!and!reasoning!(WP1)! Mul= >robot! T21! coopera= ! ve! ! T21! coopera= ! ve! ! T22! ! planning!under! ! ! Scene!Analysis!and!understanding!(WP2)! ! Knowledge>based!environment!! ! T22! ! planning!under! RealB8me!image! RealB8me!image! S! and!video! Situa8on! ! awareness ! ! uncertainty!(WP4)! ! ! ! modelling!!and!reasoning!(WP1)! ! Mul= >robot! S! and!video! Situa8on! ! awareness ! ! uncertainty!(WP4)! ! processing! T21! coopera= ve! processing! E!! ! ! ! T42! ! T22! ! ! planning!under! E!! ! ! ! T42! N! T11! T12! ! Mul8Brobot! A! S! RealB8me!image! Situa8on! uncertainty!(WP4)! ! N! T11! T12! ! Mul8Brobot! A! ! ! coopera8ve! and!video! ! awareness ! ! ! ! coopera8ve! S! Mul= >modal!short>term!human>robot! KnowledgeB Spa8al,!common! ! planning! C!! E!! processing! ! ! T42! S! Mul= >modal!short>term!human>robot! KnowledgeB Spa8al,!common! ! planning! C!! ! ! sense!and!!non! ! ! ! sense!and!!non! fundamental issues related to the design of a based! based! O! interac= on(WP3)! environment! monotonic! T! N! T11! T12! ! Mul8Brobot! A! O! interac= on(WP3)! environment! monotonic! T! ! ! modelling!and! reasoning! ! ! ! coopera8ve! ! ! modelling!and! reasoning! ! R! Seman8c!maps! U! S! Mul= >modal!short>term!human>robot! KnowledgeB Spa8al,!common! ! planning! C!! R! Seman8 c!maps! U! ! ! ! T41! interac= on(WP3)! ! ! based! sense!and!!non! ! ! ! T41! S! T31! T32! A! O! environment! monotonic! ! T! S! T31! T32! A! ! ! ! Robust! ! ! modelling!and! reasoning! ! ! ! Robust! Mul8Bmodal! Human!needs! naviga8on T! R! Seman8c!maps! ! U! Mul8Bmodal! Human!needs! naviga8 on T! shortBterm!HRI es8ma8on ! ! ! ! T41! shortBterm!HRI es8ma8on ! ! O! S! T31! T32! ! A! O! ! ! Mul8Bmodal! ! Human!needs! ! Robust! ! ! R! shortBterm!HRI es8ma8 on naviga8on T! R! ! Shopkeepers! ! ! ! Shopkeepers! S! O! S! Visitors! Managers! ! ! Visitors! Managers! R! ! Shopkeepers! Managers! S! Visitors! robust system of self-directed autonomous robots with high-level skills of environment modeling and scene understanding, distributed Robots autonomous decision-making, short-term interacting with humans and robust and safe navigation in overcrowding spaces. Sensor Seman c Knowledge Ac ons Processing Interpreta on Update network KB Scene Analysis and understanding (WP2) Knowledge-based environment modelling and reasoning (WP1) Mul -robot T21 T22 coopera ve ROBOT-1 Goals Gi* accomplished Goal G* of the robot planning under Real- me sensor, WP2: Percep on by the other robots sent to the other robots S image and video Situa on awareness uncertainty (WP4) processing GOALS π E T42 KB Message of the A DEC EXEC N T11 T12 Mul -robot Extracted features end of the goal G* coopera ve S Mul -modal short-term human-robot Knowledge- Spa al, common sense and non planning C interac on(WP3) based O monotonic T PERRCEPTION WP2: WP2 KB environment reasoning External Cameras WP1 modelling and ROBOT-2 Goal G* selected R Seman c maps U NETWORK KB reasoning PRUs-task Library by the robot T31 T32 T41 A GOALS S Robust Informa on π List of goals Mul -modal Human needs naviga on T KB DEC EXEC short-term HRI es ma on Generated O WP4 at me t: Gt R MDP-based planning DECISION Shopkeepers List of goals Selected PRU Managers S Visitors ROBOT-3 Computed policy π π(G*) policy GOALS π KB DEC EXEC Petri-Net execu on plan EXEC Naviga on Assistant !
Assistance des personnes avec des capacités limitées Défintion de nouvelles modalités d’interaction avec des limitations de la vue de l’ouie de la parole
Nouvelles modalités d’interaction Vision robotique robuste Reconnaissance de gestes et langage des signes Reconnaissane de l’émotion Capteurs de forces Une lesse et capteurs de contact Synthèse de la parole
Fabrice Maurel – Université de Caen Normandie – Laboratoire GREYC équipe HUman Language TECHnology (HULTECH) IHM : TAL : Interactions Traitement Homme Automatique Machine des Langues
Accès non visuel aux pages Web
Accès non visuel aux interfaces tactiles (VoiceOver – Talkback…)
Stratégies de lecture non visuelle de haut niveau Skimming non visuel : Scanning non visuel : On crée des paysages ! On crée des chemins ! Approches orientées « texte » (Oral/Braille) : On traduit ! Approches orientées « image » (Tactile) : On transpose !
Deux projets ZONE 1 ZONE 2 ZONE 3 ZONE 4 en cours ZONE 6 ZONE 7 ZONE 8 ZONE 5 ZONE 9 Menu haut 1 Vidéo CAC 40 Menu haut 2 Les plus partagés Couac trains Nous suivre régionaux Éditions abonnés
TACTINET : premiers résultats Etudes perceptives et de reproduction de formes
Métaphore de l’effet « cocktail party » TAGTHUNDER : première approche
https://tagthunder.greyc.fr/demo/ TAGTHUNDER : première réalisation
TAGTHUNDER : premiers résultats • Utilisateurs voyants • Expérience ressentie comme difficile • Très peu d‘erreurs
Merci de votre attention
You can also read