2010-06-29 <br>

  • Tilar interface Design
    • Need a good name
    • Main Goal: Make it so non-programmers can program robots
    • p. 91 in notebook for interview with ComDis
    • DATA LOGGING!!!!!!!
      • Show clock on robot's face to sync video with events
      • Save video from robot also, either all of it, or 5 seconds or so surrounding robot events
    • UNDO!!!!
    • Support multiple robot platforms
      • Use Sockets or TCP/IP to communicate btw server(interface) and client(robot)
        • Need detailed spec for network communications interface
        • Actually have separate programs for each robot that can be written in any language
        • Better for vision/CPU intensive
      • Support sending commands to robots and receiving output from robots (sensors)
    • Remote input devices for within session robot control
      • Maybe TCP/IP also
      • Global vs. local remote programming within the interface
      • Initiate action at any time, like sound, facial expression, change motion.
        • Need global remote listener at least
      • Also need to display to these devices (what do we need?)
        • Wiimote: vibration, LEDs, sound
          • Haptic signal for Troy to acknowledge command
        • iPhone: Graphics, etc.
      • Separate input/output channels for smart remotes (PDA, phone, etc.)
        • Inputs: camera (gestures, object locations), speech, remotes, etc.
          • Tellme studio for cellphone input?
            • Can do if we use TCP/IP and have a well-designed module interface
        • Outputs: vibration, sound, LEDs, images
      • Wii remote chording or other way to “add more buttons”
    • Remote input devices for between session robot programming
      • Honda's motion learner
      • Robot puppet programming by demonstration
      • Wii remote motion capture
    • UI Frontend
      • Get rid of tool modes, have everything as direct or proximal interaction.
      • Display animated previews of robot animations
        • With animated previews, fairly easy to do simulated runs without robot
        • Supply server with xml file for robot configuration and then can display any robot server side
      • Keep it simple, have limited number of choices at any time.
        • Hierarchical action selection
      • Draggable action creator
        • Timing will be an issue here too
        • Make more general by xml file to define DH parameters and constraints
        • Need xml file and geometry for each robot piece.
      • Record sounds within the interface
        • Need simple editing tools to crop ends, or maybe just crop silence.
        • Also allow to import sounds, and maybe that's all we need.
    • Robot animation support
      • Support head motion
      • See if Troy can move faster, more enthusiastically
      • Support animations that end in a non-zero position
        • Need to blend to start position of next animation gracefully
      • Pointing arm in particular direction (at person or object)
      • Support pause button
        • Robot should freeze, then resume what it was doing when resuming.
      • Support simultaneous sound and action and facial expression, etc.
        • Specifying when the sound should play
          • Slider on the arrow
          • “Junction” point along arrow that can slide and shows the simultaneous action
        • Marker in animation file for ideal sound trigger time
    • Intelligence support
      • Gaze direction detection
      • Robot needs to sing more songs
      • Joint attention, have robot look at object, maybe same object child is looking at
      • Sounds templates
        • “Hello <name>” becomes “Hello Alan”
        • Troy says “Troy push <object> to <name>” becomes “Troy push truck to Alan”
ar/tilar-v2-brainstorming-notes.txt · Last modified: 2014/08/13 20:46 by tmburdge
Back to top
CC Attribution-Share Alike 4.0 International
chimeric.de = chi`s home Valid CSS Driven by DokuWiki do yourself a favour and use a real browser - get firefox!! Recent changes RSS feed Valid XHTML 1.0