model of the human name stan emotion happy command watch me face location (x,y,z) = (122, 34,...

8
Model of the Human Name Stan Emotion Happy Command Watch me Face Location (x,y,z) = (122, 34, 205) Hand Locations (x,y,z) = (85, -10, 175) (x,y,z) = (175, 56, 186)

Upload: aubrey-cox

Post on 04-Jan-2016

212 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Model of the Human  Name Stan  Emotion Happy  Command Watch me  Face Location (x,y,z) = (122, 34, 205)  Hand Locations (x,y,z) = (85, -10, 175) (x,y,z)

Model of the Human

NameStan

EmotionHappy

CommandWatch me

Face Location(x,y,z) = (122, 34, 205)

Hand Locations(x,y,z) = (85, -10, 175)(x,y,z) = (175, 56, 186)

Page 2: Model of the Human  Name Stan  Emotion Happy  Command Watch me  Face Location (x,y,z) = (122, 34, 205)  Hand Locations (x,y,z) = (85, -10, 175) (x,y,z)

Model of the Human

NameStan

EmotionSad

CommandWatch me

Face Location(x,y,z) = (122, 34, 205)

Hand Locations(x,y,z) = (85, -10, 175)(x,y,z) = (175, 56, 186)

(x, y,z)(x, y,z)

(x, y,z)

Stan

Page 3: Model of the Human  Name Stan  Emotion Happy  Command Watch me  Face Location (x,y,z) = (122, 34, 205)  Hand Locations (x,y,z) = (85, -10, 175) (x,y,z)

Model of the current human: description of the current human

Human activity: description of what the user is doing

User’s request: the nature of the interaction, the task the user request of the robot

Human Agent Internal Model

Page 4: Model of the Human  Name Stan  Emotion Happy  Command Watch me  Face Location (x,y,z) = (122, 34, 205)  Hand Locations (x,y,z) = (85, -10, 175) (x,y,z)

Motivated by desire for natural human-robot interaction

Encapsulates what the robot knows about the human

Identity Location Intentions

Human Agent

Page 5: Model of the Human  Name Stan  Emotion Happy  Command Watch me  Face Location (x,y,z) = (122, 34, 205)  Hand Locations (x,y,z) = (85, -10, 175) (x,y,z)

Detection module

Monitoring module

Identification module

Human Agent Modules

Page 6: Model of the Human  Name Stan  Emotion Happy  Command Watch me  Face Location (x,y,z) = (122, 34, 205)  Hand Locations (x,y,z) = (85, -10, 175) (x,y,z)

Detection Module

Allows the robot to detect human presence

Uses multiple sensor modalities

IR motion sensor array Speech recognition Skin-color segmentation Face detection

Page 7: Model of the Human  Name Stan  Emotion Happy  Command Watch me  Face Location (x,y,z) = (122, 34, 205)  Hand Locations (x,y,z) = (85, -10, 175) (x,y,z)

Monitoring Module

Keeps track of the detected human

Localization and tracking algorithms Face tracking Finger pointing gesture Basic speech interface

Page 8: Model of the Human  Name Stan  Emotion Happy  Command Watch me  Face Location (x,y,z) = (122, 34, 205)  Hand Locations (x,y,z) = (85, -10, 175) (x,y,z)

Identification ModuleUnder development

Attempts to identify detected human based on stored model and current model

Voice pattern comparison Name Height Clothing color

Detects changes in dynamic model Clothing color Height