06-09-2017, 11:57 AM
The method used is simple, which implies a simple count of the zero crossing frequency. Two zero-crossing characteristics are used, namely: maximums and current sums that increase the precision of cognition. The DSP calculates the zero crossings of the spoken words and, consequently, generates different analog signals at its output. These analog signals are further processed to operate the apparatus. The document includes two approaches to implementing speech recognition in DSP, using the Matlab Simulink approach and secondly using Code Composer Studio (CCS). In addition, first-approach performs offline processing and the other performs word processing in real time. The results at the end describe the efficiency of the system.