Skip to main content

Audio, vision, and touch

Neural Cube perceives the world through a small set of sensors and turns what they pick up into something it can reason about. Audio and vision are the two main channels, and touch fills in a third direct way to interact with the robot without needing to speak or use another device.

All of this happens on the robot itself. Audio, video, and touch input are not streamed to any outside service.

THE ROOM AS IT ACTUALLY ISwho is speakingis this meant for mewho just tapped itspeech, speakers, tonedoorbell, footsteps, musicchatter it filters outHEARINGscenery, faces, emotionobjects and places it knowsa closer look on requestSEEINGtap frequencytouch durationrate of tappingTOUCHWHAT ISHAPPENINGNOW
hearingseeingtouchsensor into its channelone channel checking anotherinto the picture of now
Each sensor is one point of contact with the room. Audio, vision, and touch combine to inform what the robot says, does, and remembers.