CN105068782A

CN105068782A - Recognition system for voice instruction of robot

Info

Publication number: CN105068782A
Application number: CN201510487475.2A
Authority: CN
Inventors: 胡达广
Original assignee: Individual
Current assignee: Individual
Priority date: 2015-08-11
Filing date: 2015-08-11
Publication date: 2015-11-18

Abstract

A recognition system for a voice instruction of a robot relates to the technical field of a recognition system for a voice instruction of a computer. In order to enable the computer to only recognize the voice instruction from a set region or a set direction, the recognition system adopts the following technical measures: a plurality of sound pick-ups are arranged at proper positions in a set region range or outside the set region range; the sound pick-ups are connected with analog-digital conversion circuits; the analog-digital conversion circuits are connected with the computer; electronic data of voice prints or voice print features of each executable voice instruction is stored in the computer; the advantages of the technical measures are as follows: the recognition system for the voice instruction enables the computer to only carry out recognition and execution on the voice instruction in the appointed region or the appointed direction, and can very effectively exclude interference outside the region or the set direction, so that the anti-interference capability of the recognition system for the voice instruction of the computer is greatly enhanced.

Description

Robotics Sound's instruction recognition system

Technical field

The present invention is a technical field of computing machine sound instruction recognition system.

Background technology

So-called robot is exactly the combination of computing machine and mechanical hook-up in fact, and robot employs computing machine, and the use of computing machine be unable to do without the input of instruction.People, in the process using robot, use more and more and send the mode of instruction to computer input instruction with sound.The instinct of reason to be sounding the be mankind be also people send instruction, seek help, exchange the most directly, the soonest, the most effective mode, the benefit that this mode is maximum is simultaneously that it meets the habits and customs of people completely.Therefore, sound as carrier to computer input instruction be the most convenient only.The vocal print of alternative sounds or vocal print feature associate with computer instruction by people, and these sound and vocal print just can represent the instruction of input computing machine.Computing machine can perform specific instruction according to specific vocal print.But because the propagation of sound is than air almost without any condition, so sound can be propagated to any direction, the sound in any direction also can be come by air borne.When computer user uses transfer voice computer instruction, be just easy to the interference by periphery sound source, the judgement making computing machine do making mistake.Resist external interference, good way to allow the sound of computing machine only in " listening " a certain scope or in a direction, the sound transmitted other positions or direction " is turned a deaf ear to " and just can be improved greatly the antijamming capability of Robotics Sound's instruction input system like this, reduces the probability judged by accident.

Summary of the invention

Be only limited to from the sound instruction in setting regions or direction initialization to allow computer recognizing, this invention takes following technical measures: within the scope of setting regions or the extraneous appropriate location of setting regions multiple acoustic pickup is set, acoustic pickup connects analog to digital conversion circuit, analog to digital conversion circuit connects computing machine, and calculator memory contains the vocal print of executable each sound instruction or the electronic data of vocal print feature.

The benefit done like this is: in setting regions or the extraneous appropriate location of setting regions multiple acoustic pickup is set, acoustic pickup connect analog to digital conversion circuit, analog to digital conversion circuit connect computing machine.Make like this acoustic pickup can pick up in setting district or in direction initialization sound, mistiming when simultaneous computer can arrive each acoustic pickup according to the locus of the concrete installation of each acoustic pickup and sound, calculate accurate location and the direction of sound source.When computing machine receives the vocal print signal of the sound instruction that people sends, and confirm its sound source from setting region in or setting direction time, the vocal print electronic data in it and storer is just carried out identification comparison by computing machine; The vocal print signal of the sound instruction that the people received when computing machine sends, and confirm its sound source from setting regions or direction initialization outer time, refusal is compared to this vocal print signal by computing machine.Computing machine is so just made to be only limited to the sound instruction identified in setting district or on direction initialization; When being currently received from the vocal print in setting district or on direction initialization and the vocal print in storer or vocal print aspect ratio to time successful, then computing machine sends the corresponding machine instruction of the sound instruction performed representated by this vocal print.Generally speaking, the present invention is exactly the sound instruction recognition system that one can make computing machine only identify the sound instruction in appointed area (setting district) or assigned direction and perform, it can interference effectively outside the outer or direction initialization of exclusionary zone, and the antijamming capability of the sound instruction recognition system of computing machine is strengthened greatly.

Embodiment

Setting regions scope described in the application or the direction of setting, refer to that the user of computing machine or robot is when operating computing machine or robot, reasonably or the activity space scope needed or command source direction, also can be the direction of being originated by the scope of manually specifying arbitrarily or sound instruction simultaneously; Described in the application within the scope of setting regions or the extraneous suitable place of setting regions multiple acoustic pickup (microphone) is set, refer at the sound that can clearly pick up in setting range or on direction initialization with by calculating function and to judge in setting range Sounnd source direction and position or directly can judge that multiple acoustic pickup is installed in the place of Sounnd source direction; Sound instruction described in the application refers to that the executable instruction of a computing machine and a certain language or sound are associated together, and makes this language or sound become the input instruction of computing machine.

The direction of determination sound source of the present invention and the principle of position are that the sound sent based on space distribution position and the sound source of multiple acoustic pickup arrives the result because of Number synthesis computing such as mistiming of each acoustic pickup.Therefore, combination installation (quantity the is different) form of multiple acoustic pickup and the array configuration of installation site have varied, and these forms can meet requirement of the present invention; When have in setting range people sound instruction time, when computing machine confirms that sound instruction is the direction from setting range or setting, the vocal print stored in the vocal print electronic data that acoustic pickup just transmits by computing machine and storer or the electronic data of vocal print feature are analysed and compared.

Principle of work of the present invention: be only limited to from the sound instruction in setting regions or on direction initialization to allow computer recognizing, this invention takes following technical measures, suitable place in setting range or outside setting range arranges multiple acoustic pickup, acoustic pickup connects analog to digital conversion circuit, analog to digital conversion circuit connects computing machine, and calculator memory contains the vocal print of executable various sound instruction or the electronic data of vocal print feature.Suitable place in setting range or outside setting range arranges multiple acoustic pickup, and acoustic pickup connects analog to digital conversion circuit, and analog to digital conversion circuit connects computing machine.Make acoustic pickup can pick up sound in setting district like this, mistiming when simultaneous computer can arrive each acoustic pickup according to the locus of the concrete installation of each acoustic pickup and sound, calculate accurate location and/or the direction of sound source.When computing machine receives the vocal print signal of the sound instruction that people sends, and confirm its sound source from setting regions or setting direction on time, the vocal print electronic data in it and storer is just carried out identification comparison by computing machine.After comparison success, computing machine enters the executive routine of this instruction; When computing machine receives the vocal print signal of the sound instruction that people sends, and when confirming its sound source from outside setting district or the direction of non-setting, computing machine will the comparison of refusal to this vocal print signal.Computing machine so just can be made to be only limited to the sound instruction identified in setting district or direction initialization, the antijamming capability of the sound instruction input system of computing machine and robot is strengthened greatly.

Claims

1. Robotics Sound's instruction recognition system, mainly comprise acoustic pickup, computing machine, it is characterized in that: within the scope of setting regions or the extraneous appropriate location of setting regions multiple acoustic pickup is set, acoustic pickup connects analog to digital conversion circuit, analog to digital conversion circuit connects computing machine, and calculator memory contains the vocal print of executable each sound instruction or the electronic data of vocal print feature.