CONTROL OF HOME APPLIANCES THROUGH VOICE

CONTROL OF HOME APPLIANCES THROUGH VOICE

ABSTRACT

In daily life, there will be many situations where it happens to return all the way from office to home to switch on or off the device, which is very difficult. That is the situation where we imagine a system, which enable us operating devices from remote locations. When not only far away from home but also when busy at home, we imagine of operating our appliances through voice either it be AC, heater, coffee pot, microwave, lamp or any other appliance. This is no more imagination now. The digital processing capabilities of microcontrollers (MCUs) allow penetration of voice control to embedded systems. Practically any new device containing a microcontroller Unit (MCU) could be controlled by voice.  This paper emphasizes on controlling of lights and heat Using DSP56F805 with Integrated Simple Voice Control. Two basic algorithms that can be used for speech recognition are DTW(Dynamic Time Warping) and HMM(Hybrid Markov Models).This system uses DTW algorithm(speaker dependent) which has fewer requirements on hardware and less  speech data in the "training" phase.
Current system for lights and heating control can be generalized for controlling any device with a limited command set. Speaker dependence is the main restriction but “new” algorithms based on hidden Markov models (HMM) adapted to the MCU environment are going to change this situation soon as DSP56F805 has 32.5K Flash program memory so the recognition algorithm can be easily upgraded.