Automated speech recognition filter

Abstract

An automated speech recognition filter is disclosed. The automated speech recognition filter device provides a speech signal to an automated speech platform that approximates an original speech signal as spoken into a transceiver by a user. In providing the speech signal, the automated speech recognition filter determines various models representative of a cumulative signal degradation of the original speech signal from various devices along a transmission signal path and a reception signal path between the transceiver and a device housing the filter. The automated speech platform can thereby provide an audio signal corresponding to a context of the original speech signal.

Description

BACKGROUND OF THE INVENTION

[0001] 1. Field of the Invention

[0002] The present invention generally relates to automated speech recognition (“ASR”) methods and systems. The present invention particularly relates to ASR methods and devices for providing an input speech signal to an ASR platform that approximates an original speech signal.

[0003] 2. Description of the Related Art

[0004] An ASR platform as known in the art is designed to respond to a reception of a transmitted speech signal (e.g., voice commands) from a transceiver (e.g., mobile phones, embedded car phones, and phone enabled personal data assistants) with an audio signal that corresponds to the context of the transmitted speech signal. However, the performance of prior art ASR platform is adversely affected by any signal degradation of the transmitted speech signal (e.g., acoustical coupling and signal distortion) along a transmission signal path from a user of the transceiver to the ASR platform. Prior art attempts to minimize, if not eliminate, any signal degradation of the transmitted speech signal have generally proven to be unacceptable. The present invention addresses this problem.

SUMMARY OF THE INVENTION

[0005] The present invention relates to an automated speech recognition filter that overcomes the aforementioned disadvantages of the prior art. Various aspects of the invention are novel, non-obvious, and provide various advantages. While the actual nature of the present invention covered herein can only be determined with reference to the claims appended hereto, certain features, which are characteristic of the embodiments disclosed herein, are described briefly as follows.

[0006] One form of the present invention is an automated speech recognition filter including means for determining one or more models representative of a signal degradation of a first speech signal transmitted from a transceiver to the automated speech recognition filter, and means for providing a second speech signal as a function of the model(s) whereby the second speech signal is an approximation of the first speech signal.

[0007] A second form of the invention is an automated speech recognition filtering device comprising means for determining one or more models representative of a signal degradation of a first speech signal transmitted from a transceiver to said automated speech recognition filtering device, and means for providing a second speech signal as a function of the model(s) whereby the second speech signal is an approximation of the first speech signal. The automated speech recognition filtering device further comprises a database operable to store a user profile corresponding to the first set of one or more models.

[0008] A third form of the invention is an automated speech recognition system comprising means for determining one or more models representative of a signal degradation of a first speech signal transmitted from a transceiver to the automated speech recognition system, and means for providing a second speech signal as a function of the model(s) whereby the second speech signal is an approximation of the first speech signal. The automated speech recognition system further comprises an automated speech recognition platform operable to provide an audio signal in response to a reception of the second speech signal with the audio signal corresponding to a context of the first speech signal.

[0009] The foregoing forms, and other forms, features and advantages of the invention will become further apparent from the following detailed description of the presently preferred embodiments, read in conjunction with the accompanying drawings. The detailed description and drawings are merely illustrative of the invention rather than limiting, the scope of the invention being defined by the appended claims and equivalents thereof.

BRIEF DESCRIPTION OF THE DRAWINGS

[0010] FIG. 1 is an illustration of a hands-free, in-vehicle environment in accordance with the present invention;

[0011] FIG. 2 is an illustration of a transmission signal path and a reception signal path between a vehicle and a wireline network as known in the art;

[0012] FIG. 3 is a block diagram of a first embodiment of an automated speech recognition filtering device in accordance with the present invention; and

[0013] FIG. 4 is a block diagram of a second embodiment of an automated speech recognition filtering device in accordance with the present invention.

DETAILED DESCRIPTION OF THE PRESENTLY PREFERRED EMBODIMENTS

[0014] FIG. 1 illustrates a hands-free, in-vehicle environment containing a conventional vehicle 10, a conventional wireless network 30, a conventional wireline network 40, an ASR filtering device 50 of the present invention, and a conventional ASR platform 60. A user 11 of a transceiver in the form of a mobile phone 20 is seated within vehicle 10. In other embodiments of the present invention, the transceiver can be in the form of an embedded car phone, a phone enabled personal data assistant, and any other transceiver for transmitting and receiving a phone call.

[0015] A more detailed explanation of the invention will now be provided herein. Those having ordinary skill in the art will appreciate that the various described signals are based upon a discrete time instant k and the various described filters and models are based upon a discrete time, frequency domain operator z. Specifically, the operator z is used to represent the frequency response characteristics of the filters and the models described herein. Additionally, those having ordinary skill in the art will appreciate that an audio signal of the present invention can include one or more forms of sound.

[0016] Mobile phone 20 conventionally transmits a transmission signal T1[k] to wireless network 30 in response to user 11 articulating a speech signal U1[k] in a direction of a microphone (not shown) of mobile phone 20. Speech signal U1[k] is a main component of transmission signal T1[k]. A noise signal N1[k] consisting of noise emanating from various sources of vehicle 10 (e.g., an engine, a heater/air conditioner, a radio, and a pair of wiper blades) are also components of transmission signal T1[k]. In addition, an audio signal R5[k] (FIG. 2) being an acoustically coupled form of an audio signal R4[k] is a component of transmission signal T1[k]. Transmission signal T1[k] therefore ranges from a slightly distorted version of speech signal U1[k] to a significantly distorted version of speech signal U1[k] as a function of an intensity of the vehicle noise signal N1[k] and an intensity of audio signal R5[k] generated by mobile phone 20, wireless network 30, and wireline network 40.

[0017] Wireless network 30 (e.g., an advanced mobile phone service, a time division multiple access network, a code division multiple access network, and a global system for mobile communications) conventionally transmits a transmission signal T2[k] to wireline network 40 in response to a reception of transmission signal T1[k] by wireless network 30. The conventional transmission of transmission signal T2[k] involves a degree of signal distortion and a degree of signal attenuation of transmission signal T1[k] by wireless network 30. Transmission signal T2[k] therefore ranges from a slightly distorted version of transmission signal T1[k] to a significantly distorted version of transmission signal T1[k] as a function of an intensity of the signal distortion and an intensity of the signal attenuation by wireless network 30 upon transmission signal T1[k].

[0018] Wireline network 40 (e.g., a Public Switched Telephone Network, and VOIP network) conventionally transmits a transmission signal T3[k] to ASR filtering device 50 in response to a reception of transmission signal T2[k] by wireline network 40. The conventional transmission of transmission signal T3[k] involves a degree of signal distortion and a degree of signal attenuation of transmission signal T2[k] by wireline network 40. Transmission signal T3[k] therefore ranges from a slightly distorted version of transmission signal T2[k] to a significantly distorted version of transmission signal T2[k] as a function of an intensity of the signal distortion and an intensity of the signal attenuation by wireline network 40 upon transmission signal T2[k].

[0019] ASR filtering device 50 uniquely transmits a speech signal U2[k] to ASR platform 60 (e.g., a computer platform employing commercially available speech recognition software from Nuance of Menlo Park, Calif. or SpeechWorks of Boston, Mass.) in response to a reception of transmission signal T3[k] by ASR filtering device 50. The unique transmission of speech signal U2[k] by ASR filtering device 50 involves a utilization of profile based characteristics of vehicle 10, mobile phone 20, wireless network 30, and wireline network 40 as well as a utilization of real-time signal characteristics of transmission signal T3[k], audio signal R1[k], and an estimate of vehicle noise signal N[k]. The result, as will be further described herein in connection with FIGS. 3 and 4, ASR filtering device 50 transmits speech signal U2[k] to ASR platform 60 as an approximation of speech signal U1[k]. An improved performance of ASR platform 60 is therefore facilitated by a reception of speech signal U2[k] by ASR platform 60.

[0020] FIG. 2 illustrates a transmission signal path from vehicle 10 to wireline network 40 and a reception signal path from wireline network 40 to vehicle 10. In vehicle 10, a speech signal U3[k], a vehicle noise signal N2[k], and audio signal R5[k] serve as inputs of a summation process for generating a speech signal U4[k] that is a degraded version of speech signal U1[k]. A processing function in the form of an acoustic model 12 represents an acoustical effect on speech signal U1[k] that produces speech signal U3[k] as a result of a distance between user 11 (FIG. 1) and a microphone (not shown) of mobile phone 20. A processing function in the form of an acoustic model 13 represents an acoustical effect on vehicle noise signal N1[k] by vehicle 10 that produces vehicle noise signal N2[k]. A processing function in the form of an acoustic model 14 represents an acoustical effect on audio signal R4[k] that produces audio signal R5[k] resulting from an acoustical coupling between the microphone and a speaker (not shown) of mobile phone 20. Variable n within acoustic models 12-14 corresponds to a particular type of vehicle 10 (FIG. 1).

[0021] The transmission signal path includes processing functions in the form of a transmission model 21 of mobile phone 20, a transmission model 31 of wireless network 30, and a transmission model 41 of wireline network 40. The reception signal path includes processing functions in the form of a reception model 22 of mobile phone 20, a reception model 32 of wireless network 30, and a reception model 42 of wireline network 40.

[0022] Transmission model 21 represents a signal degradation of speech signal U1[k] by mobile phone 20 in a generation of transmission signal T1[k] by mobile phone 20 in response to a reception of speech signal U4[k] during a transmission mode of operation of mobile phone 20. Reception model 22 represents a signal degradation of audio signal R1[k] by mobile phone 20 in a generation audio signal R4[k] by mobile phone 20 in response to a reception of audio signal R3[k] during a reception mode of operation of mobile phone 20. Variable m within transmission model 21 and reception model 22 corresponds to a particular type of mobile phone 20.

[0023] Transmission model 31 represents a signal degradation of speech signal U1[k] by wireless network 30 in a generation of transmission signal T2[k] by wireless network 30 in response to a reception of transmission signal T1[k] during a transmission mode of operation of wireless network 30. Reception model 32 represents a signal degradation of audio signal R1[k] by wireless network 30 in a generation of audio signal R3[k] by wireless network 30 in response to a reception of audio signal R2[k] during a reception mode of operation of wireless network 30. Variable j within transmission model 31 and reception model 32 corresponds to a particular type of wireless network 30.

[0024] Transmission model 41 represents a signal degradation of speech signal U1[k] by wireline network 40 in a generation of transmission signal T3[k] by wireline network 40 in response to a reception of transmission signal T2[k] during a transmission mode of operation of wireline network 40. Reception model 42 represents a signal degradation of audio signal R1[k] by wireline network 40 in a generation of audio signal R2[k] by wireline network 40 in response to a reception of audio signal R1[k] during a reception mode of operation of wireline network 40. Variable within transmission model 41 and reception model 42 corresponds to a particular type of wireline network 40.

[0025] FIG. 3 illustrates ASR filtering device 150 as one embodiment of ASR filtering device 50 (FIG. 1). ASR filtering device 150 includes a noise suppression module 151 and an ASR filter 152 which both may be comprised of digital circuitry, analog circuitry, or both (e.g., an application specific integration circuit). Also, noise suppression module 151 and ASR filter 152 may be programmable, dedicated state machines, or a hybrid combination of programmable and dedicated hardware. To implement the principals of the present invention, ASR filtering device 150 can further include any control clocks, interfaces, signal conditioners, filters, Analog-to-Digital (A/D) converters, Digital-to-Analog (D/A) converters, communication ports, or other types of operators as would occur to those having ordinary skill in the art.

[0026] In one embodiment, ASR filtering device 150 includes one or more commercially available microprocessors (not shown) operatively coupled to one or more commercially available solid-state memory devices (not shown) containing programming corresponding to noise discrimination module 151 and ASR filter 152 as software modules. In the illustrated embodiment, ASR filtering device 150 further includes a database 153 containing user profiles including known and/or unknown values of variables n, m, j and i illustrated in FIG. 2. The following TABLE 1 is an exemplary listing of user profiles within database 153:

TABLE I
USER	VARIALBE	VARIABLE	VARIABLE	VARIABLE
PROFILE	n	m	j	i
1	1 (Car X)	1 (Phone X)	1 (AMPS)	1 (PSTN)
2	1 (Car X)	2 (PDA X)	2 (TDMA)	1 (PSTN)
3	2 (Car Y)	3 (ECP X)	3 (CDMA)	2 (VolP)
4	2 (Car Y)	1 (Phone X)	Unknown	Unknown
5	3 (Car Z)	2 (PDA X)	Unknown	Unknown
6	3 (Car Z)	3 (ECP X)	Unknown	Unknown

[0027] In TABLE 1, PDA is an abbreviation for phone enabled personal data assistant and ECP is an abbreviation for an embedded car phone.

[0028] The operation of ASR filter device 150 will now be described herein. First, noise suppression module 151 conventionally provides a transmission signal T4[k] to ASR filter 152 in response to a reception of transmission signal T3[k] from wireline network 40 (FIG. 1) by ASR filtering device 150. In one embodiment, noise suppression module 151 conventionally attenuates vehicle noise N2[k] from transmission signal T3[k] in generating transmission signal T4[k].

[0029] Second, ASR filter 152 receives transmission signal T4[k] and audio signal R1[k] to provide a speech signal U2[k] as an approximation of speech signal U1[k]. In one embodiment, ASR filter 152 identifies a user profile in response to an identification signal (not shown) to ASR platform 60 via ASR filtering device 150 prior to the reception of transmission signal T3[k] by ASR filtering device 150, and thereafter reads the values of variables n, m, j and i from the identified user profile as stored in database 153. The identification signal can include a data component or a characteristic indicating a frequency response characteristic of mobile phone 20 or wireless network 30 as would occur to those having ordinary skill in the art.

[0030] In response to variables n, m, j and i, ASR filter 152 establishes a filter Jt1(z,ij,m,n)[k] in accordance with the following equations [1]-[3]: 1$\begin{array}{cc}{P}_{\mathrm{t1}}=C_t^m*W_t^j*F_t^i& \left[1\right]\\ P_{\mathrm{r1}}=A_r^n*C_r^m*W_r^j*F_r^i& \left[2\right]\\ J_{\mathrm{t1}}\left(z,i,j,m,n\right)\left[k\right]=\frac{T_4\left[k\right]-\left(P_{\mathrm{t1}}*P_{\mathrm{r1}}*R_1\left[k\right]\right)}{P_{\mathrm{t1}}*A_u^n*T_4\left[k\right]}& \left[3\right]\end{array}$

[0031] ASR filter 152 filters transmission signal T4[k] and audio signal R1[k] through filter Jt1(z,ij,m,n)[k] to provide speech signal U2[k] as an approximation of speech signal U1[k].

[0032] In embodiments of the present invention wherein wireless network 30 is omitted, ASR filter 152 establishes a filter Jt2(z,i,m,n)[k] in accordance with the following equations [4]-[6]: 2$\begin{array}{cc}{P}_{\mathrm{t2}}=C_t^m*F_t^i& \left[4\right]\\ P_{\mathrm{r2}}=A_r^n*C_r^m*F_r^i& \left[5\right]\\ J_{\mathrm{t2}}\left(z,i,m,n\right)\left[k\right]=\frac{T_4\left[k\right]-\left(P_{\mathrm{t2}}*P_{\mathrm{r2}}*R_1\left[k\right]\right)}{P_{\mathrm{t2}}*A_u^n*T_4\left[k\right]}& \left[6\right]\end{array}$

[0033] whereby ASR filter 152 filters transmission signal T4[k] and audio signal R1[k] through filter Jt2(z,i,m,n)[k] to provide speech signal U2[k] as an approximation of speech signal U1[k].

[0034] In embodiments of the present invention wherein wireline network 40 is omitted, ASR filter 152 establishes a filter Jt3(zj,m,n)[k] in accordance with the following equations [7]-[9]: 3$\begin{array}{cc}{P}_{\mathrm{t3}}=C_t^m*W_t^j& \left[7\right]\\ P_{\mathrm{r3}}=A_r^n*C_r^m*W_r^j& \left[8\right]\\ J_{\mathrm{t3}}\left(z,j,m,n\right)\left[k\right]=\frac{T_4\left[k\right]-\left(P_{\mathrm{t3}}*P_{\mathrm{r3}}*R_1\left[k\right]\right)}{P_{\mathrm{t3}}*A_u^n*T_4\left[k\right]}& \left[9\right]\end{array}$

[0035] whereby ASR filter 152 filters transmission signal T4[k] and audio signal R1[k] through filter Jt3(zj,m,n)[k] to provide speech signal U2[k] as an approximation of speech signal U1[k].

[0036] In embodiments of the present invention wherein wireless network 30 and wireline network 40 is omitted, ASR filter 152 establishes a filter Jt4(z,m,n)[k] in accordance with the following equations [10]-[12]: 4$\begin{array}{cc}{P}_{\mathrm{t4}}=C_t^m& \left[10\right]\\ P_{\mathrm{r4}}=A_r^n*C_r^m& \left[11\right]\\ J_{\mathrm{t4}}\left(z,m,n\right)\left[k\right]=\frac{T_4\left[k\right]-\left(P_{\mathrm{t4}}*P_{\mathrm{r4}}*R_1\left[k\right]\right)}{P_{\mathrm{t4}}*A_u^n*T_4\left[k\right]}& \left[12\right]\end{array}$

[0037] whereby ASR filter 152 filters transmission signal T4[k] and audio signal R1[k] through filter Jt4(z,m,n)[k] to provide speech signal U2[k] as an approximation of speech signal U1[k].

[0038] FIG. 4 illustrates ASR filtering device 250 as one embodiment of ASR filtering device 50 (FIG. 1). ASR filtering device 250 includes a noise discrimination module 251 and an ASR filter 252 which both may be comprised of digital circuitry, analog circuitry, or both (e.g., an application specific integration circuit). Also, noise discrimination module 251 and ASR filter 252 may be programmable, dedicated state machines, or a hybrid combination of programmable and dedicated hardware. To implement the principals of the present invention, ASR filtering device 250 can further include any control clocks, interfaces, signal conditioners, filters, Analog-to-Digital (A/D) converters, Digital-to-Analog (D/A) converters, communication ports, or other types of operators as would occur to those having ordinary skill in the art.

[0039] In one embodiment, ASR filtering device 250 includes one or more commercially available microprocessors (not shown) operatively coupled to one or more commercially available solid-state memory devices (not shown) containing programming corresponding to noise discrimination module 251 and ASR filter 252. In the illustrated embodiment, ASR filtering device 250 further includes a database 253 containing user profiles including known and/or unknown values of variables n, m, j and i similarly to the description of database 153 (FIG. 3).

[0040] The operation of ASR filter device 250 will now be described herein. First, noise discrimination module 251 provides a noise discrimination signal y[k] to ASR filter 252 in response to a reception of transmission signal T3[k] from wireline network 40 (FIG. 1) by ASR filtering device 250. In one embodiment, noise discrimination module 251 inputs transmission signal T3[k] to compute noise discrimination signal y[k] in accordance with the following equations [1] and [13]: 5$\begin{array}{cc}{P}_{\mathrm{t1}}=C_t^m*W_t^j*F_t^i& \left[1\right]\\ y\left[k\right]=P_{\mathrm{t1}}*A_o^n*N_1\left[k_n\right]& \left[13\right]\end{array}$

[0041] where k, is a time sample during which speech signal U1[k] (FIG. 2) is not present within transmission signal T3[k] and vehicle noise signal N1[k] (FIG. 2) is within transmission signal T3[k]. Noise discrimination module 251 determines an absence of speech signal U1[k] and a presence of vehicle noise signal N1[k] within transmission signal T3[k] by distinguishing a significant difference in different energy level ranges and spectral characteristics of transmission signal T3[k]. Noise discrimination module 251 assumes a steady state noise characteristics of vehicle noise signal N1[k] to thereby generate noise discrimination signal y[k] in accordance with the following equation [14]: 6$\begin{array}{cc}y\left[k\right]\cong P_{\mathrm{t1}}*A_o^n*N_1\left[k\right]& \left[14\right]\end{array}$

[0042] Second, ASR filter 252 receives transmission signal T3[k], noise discrimination signal y[k], and audio signal R1[k] to provide speech signal U2[k] as an approximation of speech signal U1[k]. In one embodiment, ASR filter 252 identifies a user profile in response to an identification signal (not shown) transmitted to ASR platform 60 via ASR filtering device 250 prior to the reception of transmission signal T3[k] by ASR filtering device 250, and thereafter reads the values of variables n, m, j and i from the identified user profile as stored in database 253. The identification signal can include a data component or a characteristic indicating a frequency response characteristic of mobile phone 20 or wireless network 30 as would occur to those having ordinary skill in the art.

[0043] In response to variables n, m, j and i, ASR filter 252 establishes a filter Jt5(z,i,j,m,n)[k] in accordance with the following equations [1], [2], [14] and [15]: 7$\begin{array}{cc}{P}_{\mathrm{t1}}=C_t^m*W_t^j*F_t^i& \left[1\right]\\ P_{\mathrm{r1}}=A_r^n*C_r^m*W_r^j*F_r^i& \left[2\right]\\ y\left[k\right]\cong P_{\mathrm{t1}}*A_o^n*N_1\left[k\right]& \left[14\right]\\ J_{\mathrm{t5}}\left(z,i,j,m,n\right)\left[k\right]=\frac{T_3\left[k\right]-y\left[k\right]-\left(P_{\mathrm{t1}}*P_{\mathrm{r1}}*R_1\left[k\right]\right)}{P_{\mathrm{t1}}*A_u^n*\left(T_3\left[k\right]-y\left[k\right]\right)}& \left[15\right]\end{array}$

[0044] ASR filter 252 filters transmission signal T3[k], noise discrimination signal y[k], and audio signal R1[k] through filter Jt5(z,ij,m,n)[k] to provide speech signal U2[k] as an approximation of speech signal U1[k].

[0045] In embodiments of the present invention wherein wireless network 30 is omitted, ASR filter 252 establishes a filter Jt6(z,i,m,n)[k] in accordance with the following equations [4], [5], [16] and [17]: 8$\begin{array}{cc}{P}_{\mathrm{t2}}=C_t^m*F_t^i& \left[4\right]\\ P_{\mathrm{r2}}=A_r^n*C_r^m*F_r^i& \left[5\right]\\ y\left[k\right]\cong P_{\mathrm{t2}}*A_o^n*N_1\left[k\right]& \left[16\right]\\ J_{\mathrm{t6}}\left(z,i,m,n\right)\left[k\right]=\frac{T_3\left[k\right]-y\left[k\right]-\left(P_{\mathrm{t2}}*P_{\mathrm{r2}}*R_1\left[k\right]\right)}{P_{\mathrm{t2}}*A_u^n*\left(T_3\left[k\right]-y\left[k\right]\right)}& \left[17\right]\end{array}$

[0046] whereby ASR filter 252 filters transmission signal T3[k], noise discrimination signal y[k], and audio signal R1[k] through filter Jt6(z,i,m,n)[k] to provide speech signal U2[k] as an approximation of speech signal U1[k].

[0047] In embodiments of the present invention wherein wireline network 40 is omitted, ASR filter 252 establishes a filter Jt7(zj,m,n)[k] in accordance with the following equations [7], [8], [18] and [19]: 9$\begin{array}{cc}{P}_{\mathrm{t3}}=C_t^m*W_t^j& \left[7\right]\\ P_{\mathrm{r3}}=A_r^n*C_r^m*W_r^j& \left[8\right]\\ y\left[k\right]\cong P_{\mathrm{t3}}*A_o^n*N_1\left[k\right]& \left[18\right]\\ J_{\mathrm{t7}}\left(z,j,m,n\right)\left[k\right]=\frac{T_3\left[k\right]-y\left[k\right]-\left(P_{\mathrm{t3}}*P_{\mathrm{r3}}*R_1\left[k\right]\right)}{P_{\mathrm{t3}}*A_u^n*\left(T_3\left[k\right]-y\left[k\right]\right)}& \left[19\right]\end{array}$

[0048] whereby ASR filter 252 filters transmission signal T3[k], noise discrimination signal y[k],and audio signal R1[k] through filter Jt7(zj,m,n)[k] to provide speech signal U2[k] as an approximation of speech signal U1[k].

[0049] In embodiments of the present invention wherein wireless network 30 and wireline network 40 is omitted, ASR filter 252 establishes a filter Jt8(z,m,n)[k] in accordance with the following equations [10], [11], [20] and [21]: 10$\begin{array}{cc}{P}_{\mathrm{t4}}=C_t^m& \left[10\right]\\ P_{\mathrm{r4}}=A_r^n*C_r^m& \left[11\right]\\ y\left[k\right]\cong P_{\mathrm{t4}}*A_o^n*N_1\left[k\right]& \left[20\right]\\ J_{\mathrm{t8}}\left(z,m,n\right)\left[k\right]=\frac{T_3\left[k\right]-y\left[k\right]-\left(P_{\mathrm{t4}}*P_{\mathrm{r4}}*R_1\left[k\right]\right)}{P_{\mathrm{t4}}*A_u^n*\left(T_3\left[k\right]-y\left[k\right]\right)}& \left[21\right]\end{array}$

[0050] whereby ASR filter 252 filters transmission signal T3[k], noise discrimination signal y[k], and audio signal R1[k] through filter Jt8(z,m,n)[k] to provide speech signal U2[k] as an approximation of speech signal U1[k].

[0051] ASR filtering device 150 (FIG. 3) and ASR filtering device 250 (FIG. 4) have been described herein as pre-ASR filtering devices in electrical communication with ASR platform 60 (FIG. 1). In alternative embodiments of the present invention, ASR filtering device 150 and ASR filtering device 250 may be incorporated into ASR platform 60.

[0052] ASR filtering device 150 (FIG. 3) and ASR filtering device 250 (FIG. 4) have also been described herein in the context of an employment within a telecommunication system having a transceiver situated within a vehicle. In alternative embodiments of the present invention, ASR filtering device 150 and ASR filtering device 250 may employed within various other systems used for audio communication purposes such as, for example, a video conferencing system, and the transceivers of such systems can be situated within the system as would occur to those having ordinary skill in the art.

[0053] While the embodiments of the present invention disclosed herein are presently considered to be preferred, various changes and modifications can be made without departing from the spirit and scope of the invention. The scope of the invention is indicated in the appended claims, and all changes that come within the meaning and range of equivalents are intended to be embraced therein.

Claims

1. An automated speech recognition filter, comprising: means for determining one or more models representative of a signal degradation of a first speech signal transmitted from a transceiver to said automated speech recognition filter; and means for providing a second speech signal as a function of the one or more models, the second speech signal being an approximation of the first speech signal.

2. An automated speech filtering device, comprising: means for determining one or more models representative of a signal degradation of a first speech signal transmitted from a transceiver to said automated speech recognition filtering device; and means for providing a second speech signal as a function of one or more models, the second speech signal being an approximation of the first speech signal; and a database operable to store a user profile corresponding to the first set of one or more models.

3. An automated speech recognition system, comprising: means for determining one or more models representative of a signal degradation of a first speech signal transmitted from a transceiver to said automated speech recognition system, and means for providing a second speech signal as a function of the one or more models, the second speech signal being an approximation of the first speech signal; and an automated speech recognition platform operable to provide an audio signal in response to a reception of the second speech signal, the audio signal corresponding to a context of the first speech signal.

4. An automated speech recognition filtering method, comprising: determining one or more models representative of a signal degradation of a first speech signal transmitted from a transceiver; and providing a second speech signal as a function of the one or more models, the second speech signal being an approximation of the first speech signal.

5. An automated speech recognition filtering device, comprising: a database operable to store a user profile corresponding to a transceiver, the user profile including a first variable indicative of an identification of the transceiver; and an automated speech recognition filter operable to determine a transceiver transmission model in response to a reception of the first variable, the transceiver transmission model being representative of a first signal degradation on a first speech signal by the transceiver.

6. The automated speech recognition filtering device of claim 5, further comprising: a noise discrimination module operable to provide a noise discrimination signal in response to a reception of the first speech signal by said automated speech recognition filtering device, wherein said automated speech recognition filter is further operable to provide a second speech signal as a function of the transceiver transmission model and the noise discrimination signal.

7. The automated speech recognition filtering device of claim 5, further comprising: a noise suppression module operable to provide a second speech signal in response to a reception of the first speech signal by said automated speech recognition filtering device, the second speech signal being a product of an attenuation of a noise component from the first speech signal, wherein said automated speech recognition filter is further operable to provide a third speech signal as a function of the transceiver transmission model and the second speech signal.

8. The automated speech recognition filtering device of claim 5, wherein said automated speech recognition filter is further operable in response to a reception of the first variable to determine a transceiver reception model representative of a second signal degradation of the first speech signal by the transceiver.

9. The automated speech recognition filtering device of claim 8, further comprising: a noise discrimination module operable to provide a noise discrimination signal in response to a reception of the first speech signal by said automated speech recognition filtering device, wherein said automated speech recognition filter is further operable to provide a second speech signal as a function of the transceiver transmission model, the transceiver reception model, and the noise discrimination signal.

10. The automated speech recognition filtering device of claim 8, further comprising: a noise suppression module operable to provide a second speech signal in response to a reception of the first speech signal by said automated speech recognition filtering device, the second speech signal being a product of an attenuation of a noise component from the first speech signal, wherein said automated speech recognition filter is operable to provide a third speech signal as a function of the transceiver transmission model, the transceiver reception model, and the second speech signal.

11. The automated speech recognition filtering device of claim 5, wherein: the user profile further includes a second variable indicative of an identification of wireless network within a transmission signal path between the transceiver and said automated speech recognition filtering device; and said automated speech recognition filter is further operable to determine a wireless transmission model in response to a reception of the second variable, the wireless transmission model being representative of a second signal degradation of the first speech signal along the transmission signal path by the wireless network.

12. The automated speech recognition filtering device of claim 11, further comprising: a noise discrimination module operable to provide a noise discrimination signal in response to a reception of the first speech signal by the automated speech recognition filtering device, wherein said automated speech recognition filter is further operable to provide a second speech signal as a function of the transceiver transmission model, the wireless transmission model, and the noise discrimination signal.

13. The automated speech recognition filtering device of claim 11, further comprising: a noise suppression module operable to provide a second speech signal in response to a reception of the first speech signal by said automated speech recognition filtering device, the second speech signal being a product of an attenuation of a noise component from the first speech signal, wherein said automated speech recognition filter is further operable to provide a third speech signal as a function of the transceiver transmission model, the wireless transmission model, and the second speech signal.

14. The automated speech recognition filtering device of claim 5, wherein: the user profile further includes a second variable indicative of an identification of wireless network within a reception signal path between the transceiver and said automated speech recognition filtering device; said automated speech recognition filter is further operable to determine a wireless reception model in response to a reception of the second variable, the wireless reception model representative of a second signal degradation of the first speech signal along the reception signal path by the wireless network.

15. The automated speech recognition filtering device of claim 14, further comprising: a noise discrimination module operable to provide a noise discrimination signal in response to a reception of the first speech signal by said automated speech recognition filtering device, wherein said automated speech recognition filter is further operable to provide a second speech signal as a function of the transceiver transmission model, the wireless reception model, and the noise discrimination signal.

16. The automated speech recognition filtering device of claim 14, further comprising: a noise suppression module operable to provide a second speech signal in response to a reception of the first speech signal by said automated speech recognition filtering device, the second speech signal being a product of an attenuation of a noise component from the first speech signal, wherein said automated speech recognition filter is further operable to provide a third speech signal as a function of the transceiver transmission model, the wireless reception model, and the second speech signal.

17. The automated speech recognition filtering device of claim 5, wherein: the user profile further includes a second variable indicative of an identification of wireline network within a transmission signal path between the transceiver and said automated speech recognition filtering device; said automated speech recognition filter is further operable to determine a wireline transmission model in response to a reception of the second variable, the wireline transmission model being representative of a second signal degradation of the first speech signal along the transmission signal path by the wireline network.

18. The automated speech recognition filtering device of claim 17, further comprising: a noise discrimination module operable to provide a noise discrimination signal in response to a reception of the first speech signal by said automated speech recognition filtering device, wherein said automated speech recognition filter is further operable to provide a second speech signal as a function of the transceiver transmission model, the wireline transmission model, and the noise discrimination signal.

19. The automated speech recognition filtering device of claim 17, further comprising: a noise suppression module operable to provide a second speech signal in response to a reception of the first speech signal by said automated speech recognition filtering device, the second speech signal being a product of an attenuation of a noise component from the first speech signal, wherein said automated speech recognition filter is further operable to provide a third speech signal as a function of the transceiver transmission model, the wireline transmission model, and the second speech signal.

20. The automated speech recognition filtering device of claim 5, wherein: the user profile further includes a second variable indicative of an identification of wireline network within a reception signal path between the transceiver and said automated speech recognition filtering device; said automated speech recognition filter is further operable to determine a wireline reception model in response to a reception of the second variable, the wireline reception model being representative of a second signal degradation of the first speech signal along the reception signal path by the wireline network.

21. The automated speech recognition filtering device of claim 20, further comprising: a noise discrimination module operable to provide a noise discrimination signal in response to a reception of the first speech signal by said automated speech recognition filtering device, wherein said automated speech recognition filter is further operable to provide a second speech signal as a function of the transceiver transmission model, the wireline reception model, and the noise discrimination signal.

22. The automated speech recognition filtering device of claim 20, further comprising: a noise suppression module operable to provide a second speech signal in response to a reception of the first speech signal by said automated speech recognition filtering device, the second speech signal being a product of an attenuation of a noise component from the first speech signal, wherein said automated speech recognition filter is further operable to provide a third speech signal as a function of the transceiver transmission model, the wireline reception model, and the second speech signal.

23. The automated speech recognition filtering device of claim 5, wherein: the user profile further includes a second variable indicative of an identification of a vehicle having said transceiver therein; said automated speech recognition filter is further operable to determine one or more acoustical models in response to a reception of the second variable, the one or more acoustical models being representative of a second signal degradation of the first speech signal by the vehicle.

24. The automated speech recognition filtering device of claim 23, further comprising: a noise discrimination module operable to provide a noise discrimination signal in response to a reception of the first speech signal by said automated speech recognition filtering device, wherein said automated speech recognition filter is further operable to provide a second speech signal as a function of the one or more acoustical models, the transceiver transmission model, and the noise discrimination signal.

25. The automated speech recognition filtering device of claim 23, further comprising: a noise suppression module operable to provide a second speech signal in response to a reception of the first speech signal by said automated speech recognition filtering device, the second speech signal being a product of an attenuation of a noise component from the first speech signal, wherein said automated speech recognition filter is further operable to provide a third speech signal as a function of the one or more acoustical models, the transceiver transmission model, and the second speech signal.

26. The automated speech recognition filtering device of claim 5, further comprising: a noise discrimination module operable to provide a noise discrimination signal in response to a reception of the first speech signal by said automated speech recognition filtering device, wherein said automated speech recognition filter is further operable to receive an audio signal from an automated speech recognition platform, and wherein said automated speech recognition filter is further operable to provide a second speech signal as a function of the transceiver transmission model, the noise discrimination signal, and the audio signal.

27. The automated speech recognition filtering device of claim 5, further comprising: a noise suppression module operable to provide a second speech signal in response to a reception of the first speech signal by said automated speech recognition filtering device, the second speech signal being a product of an attenuation of a noise component from the first speech signal, wherein said automated speech recognition filter is further operable to receive an audio signal from an automated speech recognition platform, and wherein said automated speech recognition filter is operable to provide a third speech signal as a function of the transceiver transmission model, the second speech signal, and the audio signal.

28. A method of processing voice commands spoken into a first transceiver and transmitted over a first wireless network, said method comprising: storing a plurality of processing functions, wherein each processing function optimizes voice recognition characteristics for a discrete set selected from at least one of a plurality of transceivers, a plurality of wireless networks and a plurality of wireless network components, wherein the first transceiver is one of the plurality of transceivers and wherein the first wireless network is one of the plurality of wireless networks; receiving a signal having an identifying data component, wherein the identifying data component indicates at least one of (a) first frequency response characteristics of the first transceiver, and (b) second frequency response characteristics of the wireless network; and responsive to the identifying component, selecting from the plurality of processing functions a selected processing function, wherein the selected processing function optimizes processing of the spoken voice commands for at least one of the first transceiver and the first wireless network.

29. A method of processing voice commands spoken into a first transceiver and transmitted over a first wireless network, said method comprising: storing a plurality of processing functions, wherein each processing function optimizes voice recognition characteristics for a discrete set selected from at least one of a plurality of transceivers, a plurality of wireless networks and a plurality of wireless network components, wherein the first transceiver is one of the plurality of transceivers and wherein the first wireless network is one of the plurality of wireless networks; receiving a signal having an identifying characteristic, wherein the identifying characteristic indicates at least one of (a) first frequency response characteristics of the first transceiver, and (b) second frequency response characteristics of the first wireless network or components thereof; and responsive to the identifying characteristic, selecting from the plurality of processing functions a selected processing function, wherein the selected processing function optimizes processing of the spoken voice commands for at least one of the first transceiver and the first wireless network.