Speech recognition system for an automotive vehicle

• US 4,637,045 A
• Filed: 12/17/1985
• Issued: 01/13/1987
• Est. Priority Date: 10/22/1981
• Status: Expired due to Fees

• Current Assignee: Nissan Motor Co., Ltd.
• Original Assignee: Nissan Motor Co., Ltd.
• Inventors: Futami, Toru, Noso, Kazunori, Kishi, Norimasa
• Primary Examiner(s): KEMENY, EMANUEL
• Application Number:
  US06/809,882
• Time in Patent Office: 392 Days
• Field of Search: 381/41-43, 364/513.5
• US Class Current: 704/275
• CPC Class Codes: B60R 16/0373 Voice control in general G10L

Abstract

A speech recognition system for an automotive vehicle enables desired reference spoken instruction phrases to be easily re-recorded by repeatedly depressing a manual switch arrangement that commands skipping. Each activation of the switch shifts the system from a ready state for recording the current spoken instruction to a ready state for recording a succeeding spoken instruction. By repeatedly activating the skip switch ready state for re-recording any given spoken instruction is reached. In one embodiment, a separate skip switch is provided. In a second embodiment, logic circuitry responds to simultaneous activation of record and recognition switches to command skipping.

Claims

1. A speech recognition system for an automotive vehicle for activating vehicle actuators in response to predetermined spoken instructions supplied to the system via a microphone, which comprises:

(a) a manually controlled record switch for deriving a record signal when activated;

(b) a manually controlled recognition switch for deriving a recognition signal when activated;

(c) a speech recognizer for sequentially recording a plurality of reference spoken instructions whenever one reference spoken instruction is supplied to the system through the microphone while said record switch is activated, a memory having a storage area for each spoken instruction, means for shifting access to each storage area for each spoken instruction has been recorded in the storage area provided therefore, means for activating a plurality of vehicle actuators sequentially whenever one recognition spoken instruction is supplied to the system via the microphone while said recognition switch is activated and when the spoken instruction to be recognized is similar to the reference spoken instruction; and

(d) means for deriving skip instruction signal and for coupling the skip instruction signal to the speech recognizer to shift access from a currently accessed storage area for recording a current reference spoken instruction to a succeeding storage area for recording a succeeding reference spoken instruction even when the current reference spoken instruction is not supplied to the system through the microphone.

3. The automotive speech recognition system for a vehicle as set forth in claim 1, wherein said means for deriving a skip instruction S_c includes a manually activated switch which when activated is connected to energize said speech recognizer to cause access to a currently accessed storage area to be incrementally shifted to the succeeding storage area, even when the current reference spoken instruction is not supplied to the system through the microphone.

4. The speech recognition system for an automotive vehicle as set forth in claim 1 wherein said means for deriving a skip instruction signal S_c comprises:

(a) a latch circuit connected to said record switch for deriving a H-voltage level signal in response to derivation of the record signal S_a ;

(b) a first AND gate having first and second input terminals respectively connected to respond to an output of said latch circuit and to an output of said recognition switch, said first AND gate deriving a H-voltage level skip instruction signal S_c in response to said record switch and said recognition switch being activated simultaneously;

(c) an inverter connected to said latch circuit; and

(d) a second AND gate having first and second input terminals respectively connected to respond to an output of said recognition switch and an output of said inverter for deriving a recognition signal S_b in response to said recognition switch being activated while said record switch is not activated.

6. The speech recognition system for an automotive vehicle as set forth in claim 1 which further comprises a display unit for indicating some instructions thereon appropriate to the case where a plurality of reference spoken instructions are recorded in a predetermined sequence.

2. A speech recognition system for an automotive vehicle for activating actuators in response to predetermined spoken instructions supplied to the system via a microphone, the system comprising:

(a) a manually operated record switch for deriving a record signal S_a when activated;

(b) a manually operated recognition switch for deriving a recognition signal S_b when activated;

(c) a speech anaylsis section connected to said microphone for analyzing the predetermined spoken instructions supplied via the microphone to the system and for deriving analyzed spoken instruction signals;

(d) a memory unit connected to said speech anlysis section for temporarily storing the analyzed spoken instruction signals;

(e) a reference pattern memory unit including plural storage areas M₁ and M_n assigned for plural reference spoken instructions respectively;

(f) a switching section connected to said memory unit and said reference pattern memory unit for connecting said memory unit to a first of the storage areas and for sequentially transferring analyzed reference spoken instruction signals stored in said memory unit to one of the storage areas in said reference pattern memory unit;

(g) a control section connected to said speech analysis section and said switching section for supplying a command signal to said switching section for selecting the first stroage area in response to the record signal S_a being derived and for supplying the same command signal to said switching section for selecting a succeeding one of said storage areas in response to the analyzed reference spoken information signals being stored in a preceding one of said storage areas; and

(h) manually controlled means for selectively deriving a skip instruction signal S_c, said selective deriving means being connected to said control section, said control section supplying the same command signal to said switching section in response to the skip instruction signal S_c for selecting the succeeding storage area even when the current reference spoken instruction is not supplied to the system via microphone.

10. The automotive speech recognition system for a vehicle as set forth in claim 2 wherein said means for deriving a skip instruction S_c includes a manually activated switch which when activated is connected to energize said speech recognizer to cause access to a currently accessed storage area to be incrementally shifted to the succeeding storage area, even when the current reference spoken instruction is not supplied to the system through the microphone.

11. The speech recognition system for an automotive vehicle as set forth in claim 2 wherein said means for deriving a skip instruction signal S_c comprises:

(a) a latch circuit connected to said record switch for deriving a H-voltage level signal in response to derivation of the record signal S_a ;

(b) a first AND gate having first and second input terminals respectively connected to respond to an output of said latch circuit and to an output of said recognition switch, said first AND gate deriving a H-voltage level skip instruction signal S_c in response to said record switch and said recognition switch being activated simultaneously;

(c) an inverter connected to said latch circuit; and

(d) a second AND gate having first and second input terminals respectively connected to respond to an output of said recognition switch and an output of said inverter for deriving a recognition signal S_b in response to said recognition switch being activated while said record switch is not activated.

12. The speech recognition system for an automotive vehicle as set forth in claim 2 which further comprises a display unit for indicating some instructions thereon appropriate to the case where a plurality of reference spoken instructions are recorded in a predetermined sequence.

5. A speech recognition system for an automotive vehicle for activating vehicle actuators in response to predetermined spoken instructions supplied to the system via a microphone, the system comprising:

(a) a record switch for deriving a record signal S_a when activated;

(b) a recognition switch for deriving a recognition signal S_b when activated; and

(c) a microcomputer connected to said record switch and said recognition switch, said microcomputer;

(1) analyzing the predetermined reference spoken instructions supplied to the system through the microphone,

(2) temporarily storing the analyzed reference spoken instruction signals;

(3) shifting access from a currently accessed storage area to a succeeding storage area in response to simultaneous derivation of both of the recording signal S_a and the recognition signal S_b even when the current reference spoken instruction is not supplied to the system via microphone; and

(4) transferring the stored reference spoken instruction signals to a first storage area in response to derivation of the recording signal S_a and to the succeeding stroage area whenever analyzed reference spoken instruction signals have been stored in a preceding storage area.

13. The speech recognition system for an automotive vehicle as set forth in claim 5 which further comprises a display unit for indicating some instructions thereon appropriate to the case where a plurality of reference spoken instructions are recorded in a predetermined sequence.

7. A method of loading signals representing N predetermined uttered phrase to be recognized by a speech recognition system into a memory of the system, the memory initially including N command signals indicative of N uttered phrases to be stored in the memory in response to N phrases to be uttered by an operator, the system, when operating, comparing signals representing phrases uttered by the operator with the signals stored in the memory representing the N phrases to be uttered and deriving one of N signals to indicate which one, if any, of the stored signals representing the N phrases correlates with the phrase uttered by the operator, the loading method comprising:

(a) activating the system to sequentially read out the N command signals to the operator to sequentially provided the operator with N instructions as to the phrases he is to utter, the operator responding to the N instructions by sequentially uttering N phrases, each uttered phrase being associated with one of the commands;

(b) the system sequentially responding to each phrase sequentially uttered by the operator in response to the N instructions by deriving a separate signal for each uttered phrase and sequentially storing each separate signal in a separate memory location,

(c) the operator then supplying an input signal to the system to command skipping of the sequential storing of the separate signals at the separate memory locations without uttering phrases from memory locations 1 . . . (K-1) in the event of the operator wanting to re-record the spoken instructions at locations 1 . . . (K-1), where K is any integer greater than 1 and no greater than N, the system responding to the skipping command input signal to cause the memory to skip over and have no effect on instructions stored at memory locations 1 . . . (K-1) and to advance the memory to memory location K, the system supplying the command signal for memory location K to the operator when the memory is advanced to memory location K, the system responding to a phrase uttered by the operator while the memory is at memory location K by replacing the signal previously stored at memory location K with a signal indicative of the uttered phrase while the memory is accessed to memory location K.

8. The method of claim 7 wherein the operator causes a separate input signal to be supplied to the system to command skipping of memory locations 1 . . . (K-1) without uttering a phrase, the system responding to the separate input signals to sequentially skip memory locations 1 . . . (K-1), the system responding to each input separate signal to provide the operator with a separate instruction corresponding to each of the separate skipped memory locations.

9. A speech recognition system comprising a memory, means for initially loading the memory with N stored signals indicative of N phrases uttered by an operator while the system is in a recording mode, means operative while the system is in a recognition mode for comparing signals representing phrases uttered by the operator with the signals stored in the memory representing the N uttered phrases and for deriving one of N signals to indicate which one, if any, of the stored signal representing the N stored phrases correlates with the phrase uttered by the operator, indicator means for indicating to the operator that one of the N memory locations is ready to be loaded while the system is in the record mode, the means for loading including:

(a) means for sequentially reading out the N command signals to the operator to sequentially provided the operator with N instructions as to the phrases he is to utter, the operator responding to the N instructions by sequentially uttering N phrases, each uttered phrase being associated with one of the commands;

(b) means sequentially responsive to each phrase sequentially uttered by the operator in response to the N instructions for deriving a separate signal for each uttered phrase and for sequentially storing each separate signal in a separate memory location,

means responsive to an operator operation indicating that the operator wants to re-record a spoken instruction at memory location K without re-recording the spoken instructions at locations 1 . . . (K-1), where K is any integer greater than 1 and no greater than N, skipping by the sequential storing of the separate signals at the separate memory locations 1 . . . (K-1) operator does not phrases for memory locations 1 . . . (K-1) in the event of and have no effect on instructions stored at memory locations 1 . . . (K-1) and to advance the memory to memory location K, means for supplying the indicator means with a command signal for memory location K in response to the memory being advanced to memory location K after having skipped memory locations 1 . . . (K-1), means responsive to a phrase uttered by the operator while the memory is at memory location K for replacing the signal. previously stored at memory location K with a signal indicative of the uttered phrase while the memory is accessed to memory location K.