GopiGo integration with Google Cloud API Demo

Just, a quick demo that show my result of what Gopigo and Google Cloud API can do together:

In this demo, Gopigo is invoking Google Cloud Vision and Google Cloud Text-to-Speech API.

I had on top of my mind to have a helper for managing Alexa :slight_smile:

2 Likes

I love it. It can read itself to sleep, queue up its own entertainment, and it looks sharp too! Great job.

What are the other two sensors on the left side of your bot?

1 Like

One is a temperature and humidity sensor (DHT11), the other is a compass which I still need to figure out how to integrate in my project. My goal is to localize GoPiGo in my house.

1 Like

You might be interested in this PiWars talk about using the magnetometer and encoders with montecarlo simulation to localize in an a priori map, and also the bit about mounting.

There are known I2C issues interfacing to some IMU devices. I abandoned using a 9250 and have been investigating the DI BNO055 because they already figured out the clock stretching interface for it… I run the distance sensor using hardware I2C and access the IMU using software I2C but Ihave not endurance tested the combination. (When I ran the distance sensor on software i2c it would kill i2c after a short while.)

1 Like

Woohoo! You got it to work!

High fives!

1 Like

Whoo-wee!  Hot diggity damn!!  (Carl, are you watching this?)

@cyclicalobsessive’s robot, (Carl), can wander around (relatively) independently, knows how to feed itself, (though potty-training is on the TODO: list :wink:), and sounds like a jolly-old British Squire :face_with_monocle: while doing it all.

Another capable individual, (@brad_nielsen29), is designing/has designed a fire-fighting robot

Others have aspirations of Jetson Nano powered AI BEASTS  that make the Mars Rover look like a cartoon bicycle.

My poor 'bot, (Charlie), still has aspirations of wandering around while tethered to a (remote) joystick without decapitating himself or boxing himself into a corner somewhere.  Independent “thought” and action is just a dream to him.

I’m depressed. . . .

Charlie’s jealous, and I get such a feeling of minuscule inadequacy when compared to Rocket Scientists like you folks.
:sob:

P.S. What’s the huge brass-looking thing on the 'bot’s back?

P.P.S. We’re eagerly awaiting a link to your GitHub repo!

Perhaps " PTron Sonor Portable Wireless Bluetooth Stereo Speaker with Bass Sound for All Smartphones (Gold)"

1 Like

I have called my robot “Phil” because my kids (5 and 7) love the movie “marmaduke” especially when they hear the dog says to his master “I hate you, Phil!”
so Phil is like a familiar name for us! Then, when I have started to work back on GoPiGo last month (had it closed in the box for years after the born of my son)…after some tests and headaches while trying to get Google Cloud API working I have started to say “I hate you Phil!” as well. :slight_smile:
At the end, I have discovered that Google Cloud API didn’t support
my original piboard, so I had to buy a pi3b+.

That is correct!

Absolutely, as soon as I will have chance to put my code in good and clean format.

2 Likes

I apologize if I am looking too closely at the magic. I am curious to a fault sometimes.

Please correct me where I have not understood what I see:

  1. The RPi is commanded to turn an angle

  2. snaps a frame

  3. sends the frame to Google Cloud vision api for client.text_detection()

  4. If text detected, format (first | longest | ??) text.description from response.text_annotations
    for input to google cloud text-to-speech

  5. send request up to client.synthesize_speech

  6. write cloud response to mp3 file

  7. play response mp3 file (via Bluetooth audio device)
    (Beyond all the other magic, if you got a bluetooth audio device to be convenient and reliable, I am in awe of this alone.)

  8. External: Alexa Echo? hears “Alexa, Play Queen’s Music” and starts playing selection

  9. RPi: Rinse and repeat from step 1

Q) What is connected to the Aux input of the Bluetooth speaker?

1 Like

Everything you said is correct except that:

  • google cloud text-to-speech return a mp3 audio
  • the audio is played through Aux Input via a jack 3.5
2 Likes