Asked
— Edited
Hi I'm new here. I have been working with Arduino's for a while now. I have a full scale replica R2D2 I built. Its mainly controlled via RC but I have been adding autonomous functions. It appears that EZ-robot is the answer or best way to go a bought this. But I have a few quick questions:
1). Can 2 EZ-B4 boards talk to each other via WIFI 2). Can 2 EZ-B4 boards talk to each other via Serial 3). What is the expansion port for?
If you'd like to see a short video of my project so far:
I'm new here but. You might try one mic, per room and use a motion sensor to activate each room.
The issues were that the room mics would pick up too much noise. Even the best microphone will pick up rubbish if it's not on the user. Motion detectors didn't help much either. I guess it depends on how quiet the house is though.
I have tried a whole range of microphones for Jarvis from the Kinect mic to high quality directional mics with noise reduction. None work very well. I will admit that I do have very high standards though.
I finally found a wireless mic on the person/people speaking worked a lot better. The accuracy was greatly increased and I now have no triggering from false commands. Accuracy of understanding the spoken words is also dramatically increased, previously I would need a confidence level of 75% for Jarvis to trigger events, any higher and it wouldn't accept them. I now have this set to 97% and rarely do I need to repeat commands and never does noise trigger commands.
Another note though, I use dictation with Jarvis for quite a lot of commands rather than have a set phrase list. For a small phrase list I'm sure the setup could be of a much lower quality but do you really want that?
I think what d.cochran ,is trying to do, is make it so that anyone who enters the room can talk with or to the robot. Same as I'm thinking of doing with a C3po in my computer shop. In both cases we would have to issue a mic. to each person that entered the room. In this instance a centralized mic. would be better, because you do not have one central operator.
Rich is correct that a whole room mic gives a lot of false triggers. Also it's hard to get the software to respond to other people that hasn't trained the software.
That said I don't want to ware a mic when talking to my B9 and also want others to try to speak to him. I'm OK with some false triggers as it makes him look more autonomous. The coolest thing that happened recently was a friend of mines young son was sitting on a bar stool in front of B9 watching a game of pinball on one of my games. B9 suddenly bent over him and scared the crap out of him. Poor kid but we all laughed till we had tears.
Anyway, I've had real good success with a stand up mic called The Blue Snowball. It has several settings and even looks cool.