Voiced instruction identification system

Claim 2

A voiced instruction identification system comprising a means for detecting sound which is generated in a form divided into moras, through a microphone with provisions for mounting on a human body; a means for sampling a fundamental wave from the detected sound; a means for determining a value of at least one feature of the fundamental wave for each mora; a means for forming a pattern in respect to relative changes of the value of said feature between contiguous moras; and a means for identifying an instruction from the formed pattern.

Claim 3

A voiced instruction identification system as defined by claim 1, wherein said value determining means includes a means for determining a representative value of frequencies from the fundamental wave of each mora.

Claim 4

A voiced instruction identification system as defined by claim 1, wherein said value determining means includes a means for determining a representative value of frequencies from the fundamental wave of each mora and a means for determining a representative value in respect to at least one of the amplitudes of the fundamental wave of each mora and the duration of each mora.

Claim 5

A voiced instruction identification system as defined by claim 2, wherein the fundamental wave for each mora has several periods, said frequency representative value determining means includes means for eliminating an initial few of said periods of the fundamental wave for each mora and determining a representative value from the succeeding few periods of the fundamental wave for such mora.

Claim 6

A voiced instruction identification system as defined by claim 3, wherein the fundamental wave for each mora has several periods, said frequency representative value determining means including means eliminating the initial few of said periods of the fundamental wave for each mora and determining a representative value from the succeeding few periods of the fundamental wave for such mora.

Claim 7

A voiced instruction identification system as defined by claim 2, wherein said frequency representative value determining means includes a means for determining a representative value for each mora from frequencies of a few periods of the fundamental wave when said frequencies are within a certain range.

Claim 8

A voiced instruction identification system as defined by claim 3, said frequency representative value determining means includes a means for determining a representative value for each mora from frequencies of a few periods of the fundamental wave when said frequencies are within a certain range.

Claim 9

A voiced instruction identification system, particularly for controlling a powered device by voiced instructions of a source person, comprising: means including a microphone and band pass filter for passing a limited frequency band of electrical signals corresponding to spoken moras; Schmidt trigger means for converting said limited frequency band of electrical signals from sine wave form to square wave pulses with several Schmidt pulses per mora; means for eliminating the initial unstable portion of the Schmidt pulses for each mora; a clock pulse source of frequency greater than frequencies in said band, and counter means responsive in each mora to Schmidt pulses occurring subsequent to said initial unstable portion for counting the number of clock pulses in each of a selected number of said subsequent Schmidt pulses so as to produce a selected number of clock pulse totals for each mora; symbolic value sampling means for selecting a representative one of said clock pulse totals and producing an output corresponding quantatively to said representative total, for each mora; tonal pattern change detecting means for detecting a change in said symbolic value sampling means output between consecutively occurring ones of said moras, such that the detected changes provide a tonal change pattern usable to identify the spoken command comprising said moras.

Claim 10

A voiced instruction identification system, particularly for controlling a powered device by voiced instructions of a source person, comprising: means including a microphone and band pass filter for passing a limited frequency band of electrical signals corresponding to spoken moras; Schmidt trigger means for converting said limited frequency band of electrical signals from sine wave form to square wave pulses with several Schmidt pulses per mora; means for eliminating the initial unstable portion of the Schmidt pulses for each mora; a clock pulse source of frequency greater than frequencies in said band, and counter means responsive in each mora to Schmidt pulses occurring subsequent to said initial unstable portion for counting the number of clock pulses in each of a selected number of said subsequent Schmidt pulses so as to produce a selected number of clock pulse totals for each mora; symbolic value sampling means for selecting a representative one of said clock pulse totals and producing an output corresponding quantatively to said representative total, for each mora; tonal pattern change detecting means for detecting a change in said symbolic value sampling means output between consecutively occurring ones of said moras, such that the detected changes provide a tonal change pattern usable to identify the spoken command comprising said moras.

Claim 11

The apparatus of claim 8, including means responsive to completion of said selected number of Schmidt pulses in a given mora and to a subsequent mora pause exceeding a reference time for resetting said eliminating means and counter means upon restarting of Schmidt pulses for the next mora.

Claim 12

The apparatus of claim 9, including means responsive to a mora pause for occurrence of a preselected number of moras corresponding to the number of moras in the spoken command to be identified, for causing said tonal pattern change detecting means to detect said changes.

Claim 13

The apparatus of claim 8, in which said tonal pattern change detecting means includes a pitch data stack for sequentially storing said symbolic value sampling means outputs each quantatively corresponding to a respective clock pulse total, said apparatus further including a rhythm counter and means supplying same with further clock pulses for counting the time interval during which the sound for each mora is present, and a rhythm data stack for receiving count data from said rhythm counter for each of the desired number of moras.

Claim 14

The apparatus of claim 11, further including means responsive to the output of said microphone for detecting the peak amplitude value of each sound mora and providing a digital representation thereof, and a stress data stack for storing said digitized amplitude values for the desired number of moras, whereby changes, from mora to mora, in the data in the pitch, rhythm and stress data stacks may be compared to corresponding changes in a reference instruction word to identify the spoken word applied to the microphone.