SPEECH RECOGNITION SYSTEM

摘要

A microcomputer is used for speaker independent speech recognition with a carefully selected vocabulary which may be manufactured at extremely low cost for specialised applications. A circuit (13) generates an AC electrical signal having a frequency determined by the speech input. A detector (33) produces digital signals by comparing the AC electrical signal with a threshold electrical signal level. A counting circuit (31) is connected to the detector for counting the digital signals within time intervals defined by a clock circuit (35) to generate digital count signals. The digital count signals are analysed to produce a speech identifying digital signal, and the speech identifying digital signal is compared with a plurality of speech template digital signals stores in a memory (19) to determine if the speech identifying digital signal corresponds to a matching one of the plurality of speech templates. An output signal generator (23) identifies the matching speech template.