Projective Virtual Reality

yardbellΤεχνίτη Νοημοσύνη και Ρομποτική

14 Νοε 2013 (πριν από 3 χρόνια και 9 μήνες)

80 εμφανίσεις

Projective Virtual Reality

Freund and Rossmann

Summarized by Geb Thomas

Objective


Create a scalable approach to graphically
oriented man
-
manchine interfaces


Suited for robotics and automation in a
factory environment and for robot control
over long distances

Projective Virtual Reality


“project” actions from the virtual world into
the real world through robots


VR automatically translates actions in VR
to change the real world


Changes in the real world change things in
VR

Basic of Projective VR


Quality of man machine interface depends
on the user


“The less the user in the virtual world
needs to know about the means of
automation which carry out the task
physically, the better is the design of the
man machine interface.”

A Straightforward Approach


Teleoperation or hand
-
tracking mode


directly track the data
-
glove with the
robot’s TCP


flexible, OK if there is no direct contact


Three modes:


training


generating robot programs


direct on
-
line connection

Disadvantages of Teleoperation


Time delays


precision of graphical model


precise position and orientation information


hand tremble


Online collision avoidance is necessary


Versatile sensor control is necessary to
avoid unwanted tensions

Task Deduction


Inspired by listening to transcripts from
ground control “move sample A into slot B”
or “replace battery in facility C”


Objective: give commands to the robot
system on different levels of abstraction


Subtasks are automatically recognized
through task deduction

Principles of Task Deduction


3 Different classes of action


interactions with the environment


events related to user movements


events related to VR/robot communication


Advantages


User is out of “realtime control loop”


Subtasks are carried out safely with sensor control


Sensors compensate for VR inaccuracies


Accuracy of data
-
glove less important


IRCS planning allows multi
-
robot coordination


Multi users


No longer a need to show robots


Multi automation controlled in a single environment

Limitation


Must know the type of task in advance


Example tasks set by German Space
Agency were 95% successful


Action Planning


Translate high level goals to specific actions


generates an ordered set of elementary
actions


Demands are represented by a constraint net


Selection is performed by task’s subject or
by a selection rule


Images and Metaphors
Simplifying the Work


Locations of physical objects


Visualizing Sensor Information


Placement Aids


Virtual Time


TV
-
view into reality