Real-world event recognition using multiple instant learning sensors

This paper discusses meta-event recognition methods using multiple smart sensors. We previously proposed a smart sensor, which is implemented on small and low-cost microcontroller device, and can instantly learn and detect sound events. We consider an aggregation method for multiple event information to improve recognition rate and estimate position of sound source, reasoning relations among detected events. Using collaborated smart sensors, developer will easily build smart objects and smart spaces utilizing real world sounds without signal processing programming.