2
« on: December 28, 2006, 10:33:56 pm »
Hello All,
I am trying to come up with a voice controlled home management system and I want it to have voice recognition like HAL does. Is it possible to run multiple sessions of HAL? I plan to have touchscreens in each room and want each room to be a separate feed to HAL. If I could run HAL multiple times and assign him to multiple sound inputs, each room have it's own dedicated access to HAL. I know the machine could bog down when all stations are speaking to HAL, but it would be fun to find out what the tipping point is.
I imagine it going like this. I rig up HAL to use multiple sound inputs by setting each session of HAL to use one of these sound sources.
All sources use the same global brain. This helps in keeping a global family calendar. But since HAL can switch users, each user would get their own private information store with HAL. The user can speak a voice command such as "Hal, it's George" to swap the user file to their personal one. Terminating the session closes the user file. If no user file is selected it functions solely off the global brain file.
Can HAL do this? If HAL could be invoked multiple times (as many as you have sound devices) you could have HAL even take commands from a VOIP client and have it respond to each user uniquely.
Just wondering what your thoughts are.
Jeff