NAOMI Plugin Testing Status

These are the “Modules” that I have tested with NAOMI.  They will require some tweaking when I get them all working together on the framework I am using.
These modules, are used in combination with a python based framework, that will give it a more organic feel.  I am re-writing some of the modules to be lighter.  Currently 99.9% of NAOMI can run on a Raspberry PI 2.  However, being that is mostly written in python, it can be installed on almost any

 

Tested:
  • Wifi
    • Connectivity
    • AP Mode
      • Simultaneous AP and Client
        • Bridged
          • to “Borrow” a connection
  • Voice Recognition (STT)
    • Wit.ai
      • dependent on a pseudo-service
      • Owned by an Evil Company (Facebook)
        • Not a good long term solution
    • PocketSphinx [Passive Listener]
      • Tested with stock training
        • passive listener needs to be setup again
          • after more STT training
        • Very poor STT performance out of the box
      • Needs to be setup, and trained to at least catch the activation word (NAOMI)
        • Word may be changed to something easier for STT to catch.
      • Needs considerable audio training
        • this process would have to be automated
  • Text to Speech (TTS)
    • maryTTS
      • Works Well
      • Heavy Memory Footprint
      • Hosted on Naomi
  • Reading In Remote Android Sensors [see NOTE 4 – Sensor Output]:
    • These are just the base sensors w/o any programming on the android side:
      • Mag Field
        • AKA Compass
      • Temp
      • Motion Detection
      • Light
      • Acceleration
      • Gryoscope
      • Barometer
      • Proximity Sensor
      • Step Counter
      • Step Detector
      • LED State
      • Gravity Sensor
      • Game Rotation
      • Screen Rotation
  • Servo Control
    • Home Automation
      • Lighting
        • PWM HAT Required (need more)
        • Currently handles 16 devices
      • Security
        • MAG Locks needed (on order)
          • 1200lb x 3
          • 300lb x 3
        • Relay HAT needed (on order)
    • Kid Carrier Project
  • OpenCV Facial Recognition
    • Used for user profile
  • Notifications – AKA “Communication Methods”
    • NMA – Notify My Android
      • BASH
      • Python
    • LAN BASED
      • bonjour
      • ssh
        • screen popups
        • Physical Actions
    • XMPP
      • Google Talk (AKA Hangouts)
      • HipChat
    • Audio
      • TTS (maryTTS)
        • Resides on NAOMI
      • espeak
        • most debian machines
    • Email
      • python plugin (audio only atm)
        • checks GMAIL
      • fetchmail
    • WOL
      • Only for specific devices
Not-Tested:
  • Facial Recognition:
    • Profile Switching
    • Other methods of Facial Recognition
  • The Machine Learning Layer
    • To be reworked in Python
    • Original Was in Java
      • OpenCyc
        • Too Resource intensive
        • Memory Waste
        • Wont run on RPI
  • The Natural Language Processing Layer
    • To be reworked in Python
      • Good Libraries for this in place
  • GPS Navigation
  • Proximity Sensors
  • Steering By Software
    • “Tank Style”
    • Virtual Joystick
    • Hardware
      • USB Controller
      • PS3/4 Controller
        • Internal Gyro
        • Touch-pad
        • ETC
  • Calendar Based Controls
    • Door Locked During Meeting
      • Overridden Easily (aka Yellow Mode)
    • Keep Light RED during meetings/phone calls/etc
  • Door Controls
    • 3 Different Modes
      • Green
        • Unlocked “Unlocked Mode”
      • Yellow
        • Locked “Privacy Mode”
          • Easily Unlocked
      • Red
        • Locked “Locked Mode”