Close
0%
0%

The Barkenator 9000

Detecting and classifying dog barks and activities with AI precision.

BDM
Similar projects worth following
I have a special dog that loves to bark however each bark or whine has a different meaning.
People who come over may not understand these meanings so I decided to make a device work on my dog that can communicate his intentions, well my thoughts about his intentions via light, sound, and through an application on a smart device.

Here's a few examples:
Total Spazoid - wee-o-wee-o barks coupled with digging
Lurker at the door - Angry viscous bark, very deep and loud
Morning bathroom - Small grunts (think purring sounds or nasally sound)
Needy (food, outside, ball) - small barks, higher pitched than a viscous bark
Hurt - The common yelp

Since we have a 9DOF sensor on board we should detect:
Playing - Small growls coupled with breathing and tons of movement. Don't go down that path.
Eating - Loud crunches with distinct
Drinking - Lapping sound
Sleeping
Running

Edison Details - http://www.intel.com/content/dam/support/us/en/documents/edison/sb/edison-module_HG_331189.pdf

Edison Pinout - https://cdn.sparkfun.com/assets/learn_tutorials/3/2/2/edison-pinout.pdf

License:

 [AGPLv3 Logo]

  • 1 × Intel Edison
  • 1 × 9DOF module for Edison From sparkfun
  • 1 × Battery module for Edison From Sparkfun
  • 1 × USB Mic USB Microphone that plugs directly into USB slot
  • 1 × Micro A USB to USB female Used to connect the microphone to the Edison
  • 1 × NeoPixel Color to associate with a behavior

  • NSA Signals detected

    BDM04/21/2017 at 03:21 0 comments

    I took the rough drawing I made in fusion and glued a few pieces of balsa wood together and tossed it into the mill to make something to hold it all to my dogs collar. Little Velcro and we were good to go and what I have here is a small glimpse at one bark aka the NSA (Not So Anthropomorphic) Signal. I have the threshold to record too low right now and it records every sound however with this visualization I can see the amplitude of a bark as well as some panting. I figure these are the high and lows and everything else should be rejected based on volume.

    My next step will be to discriminate between a bark, pant, grunt, etc... and the normal background noise (people talking, doors, banging, walking, etc....) . With thresholds adjusted I will strap this back on his collar and get some sounds as UPS man arrives, wanting to go outside, and a resource being "stolen" by another dog.

    Can you spot the bark? It's the large green beacon with a red base. The rest is echo's since I have a high ceiling and tile floor. You can also see on the timeline more barking however it is mixed with my other dogs but for this dog he has a distinct bark signature, more to come as I show you comparisons in another log.

  • Idea for housing

    BDM04/16/2017 at 21:57 0 comments

    ere is the rough idea for the housing that I will place the device in. It needs to be cleaned up a ton with inserts, strap inserts, screw holes, fillets, and of course material type/color changes. The inside will be a true pocket when complete.

  • Software installed

    BDM04/06/2017 at 20:10 0 comments

    Finally got around to install all the software that will be used in the application for the deep learning part.

    USB mic was also configured and is working so time to strap this to my dogs neck to get some sample data.

    High Level installed items: (will put in a document later)

    wget install miniconda2

    conda install numpy

    conda install scipy

    pip install librosa

    conda install theano

    conda install h5py

    figure out how to install tensorflow for timing comparison to theano

    pip install keras

    -->update ~/.keras/keras.json (backed=theano)

  • Still updating documentation

    BDM04/03/2017 at 16:25 0 comments

    ploade a new Draw.io file to github for the high level application flow. Still needs ton of work however the high level flow is nearly complete, just need to convert from my notebook to draw.io.

    I have started the Fusion 360 part however that will take time as I am new to it. Screws or snaplid, that is the question I need to solve. Most likely screw lid since I will need in it, battery replacement ease in the future, and of course you can't bite it open. Also would ensure any o-ring seal is solid.

  • Rough circuit concept

    BDM03/29/2017 at 02:50 0 comments

    This is a work in progress however it gets the general idea across. Most likely I will remove the USB charging and make it a standard plug since this is going on a dogs neck and small debris could get lodged in it and prevent charging.


  • Core modules diagram

    BDM03/25/2017 at 00:20 0 comments

View all 6 project logs

Enjoy this project?

Share

Discussions

David H Haffner Sr wrote 2 days ago point

This is the coolist thing !!

  Are you sure? yes | no

Similar Projects

Does this project spark your interest?

Become a member to follow this project and never miss any updates