Tonight I experimented with different types of motors while thinking about the conversations Yiyang and I have been having about our final project. After doing the two H-Bridge labs, I made this with a servo motor:
I think this type of servo-motion will be perfect for the Surveillance Bird. We just need to get motors with enough torque to support the weight of the bird + camera + platform. The weight is still a variable, but we have a much better idea of what we’re working towards with the final project.
We started off with too many things we want to accomplish. At first we had multiple bird-shaped cameras scattered about a room. From there we wanted the birds to make sounds / talk / record video.
We’ll start off focusing on getting one bird with a camera in its head to rotate so that faces are the cent
er of its video. This is the first systems diagram I’ve included.
We got a lot of feedback that if we used video, we should complete the feedback loop because people want to see the video. I would prefer to work with sound and use sound to allow people to have a conversation with the Surveillance Bird. The bird would either repeat things it heard when volume was above a certain threshold (like a parrot) or it would say things. We talked about allowing one of us to control the bird but I’m more interested in letting the program dictate the bird’s interactions. If we incorporate sound, this would require installing a speaker and microphone into the bird. It would also be nice to have a third motor tilt the bird’s head while it’s “listening.”
If we can get sound working, it would be fun to complete the concept of Surveillance Bird by sending audio snippets to a Speech to Text API, and tweeting the results from @birdveillance. We could then display the text on a screen (perhaps the screen of the computer that we’re using to process the video) or have the bird read the text and add “follow me @birdveillance on twitter.”
INITIAL BILL OF MATERIALS
- Bird Doll
- Small USB Camera (from ER)
- Wood for base
- Two (or three?) servo motors (one of X turn, one for Y tilt, and possibly one more to tilt the head to the side like an inquisitive dog when “listening”) strong enough to rotate camera/bird/base.
- Maybe foam or something to keep the motors and camera in place?
- Arduino
- Computer with Processing OpenCV face detection library
- Microphone (part of the camera?)
- Speaker
- Cables to connect everything
TIMELINE
- 11/19 – functional prototype that can detect faces and rotate at least on the X axis.
- 11/26 – tweak the bird’s looks and movement. Create bass and figure out how to hold everything together.
- 12/2 – add sound elements to complete the surveillance feedback loop.
- after PComp final, we might continue to tweak and try to install this in a bird-like location during the winter show.