We're back together, and we're shouting at our robot again. Only this time it's way better:
We've got a new batch of overhauled voice code, thanks to James The Hatchett Matchett. It uses Azure for voice recognition instead of Google Cloudspeech, and seems to be working way better.
We've got a system that queues the recognisable commands (so if you say "forward forward left right forward" it can store and release them
We've got better response time between listening and sending the commands
We've added the Gyro sensor to keep our robot from drifting off track.
And, just to make sure we're not up to anything else...
We DO NOT HAVE a line sensor or colour sensor
We DO NOT HAVE a Pi camera
We DO NOT HAVE a bluetooth PS4 controller hooked up to this challenge
WE ARE ALL ABOUT THE VOICE
Our Voice Commands have been modified for even better accuracy:
"Forward": Go forward 30cm
"Fettuccine": Go forward 15cm
"Left": Turn left 20 degrees
"Lasagne": Turn left 10 degrees
"Right": Turn right 20 degrees
"Ratatouille": Turn right 10 degrees
And, for the big turns, we encoded the names of local politicians:
"Michelle O'Neill": leans to the left
"Arlene Foster": leans to the right
Instead of the terrible time of 4:46 we got last time, we got our time down to 2:00, then 1:45, then 1:30, and our best time of the day was 1:21.
We may submit this, but may try again if we meet up one more time.
And here's a photo of me playing Animal Restaurant.
Comments