JPS60501086A - Simultaneous transmission of voice and data over analog channels - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed description of the invention] Simultaneous Transmission of Voice and Data on Analog Channels techniques for restoring analog voice and modulated data transmitted simultaneously through related. In this technology, the receiver separates the two signals received at the same time. phase jitter and frequency offset in the recovered data signal. and the ability to substantially improve the removal of demi-waves from audio signals by compensating for is being given.

Current analog transmission equipment can simultaneously transmit voice and data over the same channel. It would be even more efficient if it could be sent. Preferably such a suggestion would be a restored sound It must not reduce the quality of voice and data, and it must not reduce the quality of voice and data, and It doesn't have to be anything. At the same time, we aim to provide a simple and inexpensive system. is desirable.

Data and voice signals in telephone systems where communication is carried out via wireless coupling. One way to send this is by L. E. Schnall (L.) on July 21, 1981. No. 4,280,020 issued to E., 5chnurr). ing. Data and voice signals are separated in the frequency domain and each has its own separate size. transmitted on a low-band channel.

The data sideband channel converts an otherwise continuous wave signal into a time-marked signal. Contains sidebands generated by encoding.

For multiplexing or demultiplexing audio and non-audio signals The distributed spectrum instrument of N.F. U.S. Patent No. 4.3 to N.F. Maxemchuk]31 No. 97. Here, at the transmitter, the audio signal of the block is The time domain is transformed into the frequency domain using the Rie transform. That's the Fourier component. It can be selected almost at random from a set of eel ingredients. responsive to its selected ingredients and the predicted value of that component replaces it. The predicted value is then modulated. example For example, to reflect a multiplex of non-speech signals whose amplitudes are logical 1 or 0. , increased or decreased. The modulated predicted value is sent to the receiver for transmission to the receiver. It is transformed back into the time domain. At the receiver, the multiplexed signal On the other hand, parallel demultiplexing is generated to extract audio and non-audio signals. Jiru.

Recently, we have been using the characteristics of the short-time fast Fourier transform (CFFT) and the statistical characteristics of speech. Several systems have been proposed for transmitting voice and data simultaneously. example Based on the “Chord Playing Technique” by Baar, Steel (R, 5teele) et al. Simultaneous transmission of voice and data used ("Sim-ultaneOus Tran" mission of 5peach and Data usingCod e-Breaking Techniques”), BSTJ Vol. 60 No. 9, November 1981 issue, p. 2081-2105:; There are three systems used as carriers. Proposed. More specifically, the audio is sampled at 8 kHz and The correlation coefficient and root mean square value of the sample are divided into blocks of samples. When the stem threshold value is exceeded, the data enters the image sending state. If the data is logical 0 The sample is sent without modulation, but the presence of a logic 1 causes the frequency inversion shift of the sample. Clamp linking (scrambling) occurs. Receiver for voice and data Perform the inversion process to restore both. These techniques are very complex and requires careful timing and non-distributed channels.

The remaining problem is how to handle the various effects produced by the analog channel at the receiver. A technique that simultaneously transmits voice and data over an analog channel while compensating The goal is to provide a technology that is simple, inexpensive, and does not require increased bandwidth. Is Rukoto.

The aforementioned problems have been solved by the present invention. The invention provides It relates to technology for simultaneously transmitting analog voice and modulated data. This technique The technique is the ability to separate the two signals received at the same time on the receiver side, to compensate for phase jitter and frequency offset in the recovered data signal. It has the ability to substantially improve the removal of data signals from voice signals.

Simultaneously received analog signals from an analog transmit channel containing the default channel bandwidth. Log audio signal and modulated data signal It is one feature of the present invention to provide a receiver capable of restoring each.

where the analog audio signal has a predetermined power over the bandwidth of the analog transmission channel. The data signal has the highest power density characteristic of this analog audio signal. It is received in the part of the analog transmission channel frequency band where the level is low. to the receiver At the time, the data is detected and remodulated and then passed through an adaptive filter to and subtracted from the received signal to produce recovered audio. In adaptive filter The weights used in are adjusted by the device implementing the least mean squares algorithm. to maximize removal of the data signal from the received synthesized speech and data signal. It works just like that.

During the data detection process, information related to phase jitter and frequency offset is To substantially improve the removal of data signals from formed and recovered audio signals It is used when remodulating data.

Other features of the invention will be apparent from the following description and the attached drawings. It will be.

Referring to the drawings, like parts are designated by like numbers in the several figures. Figure 1 shows the audio signal and multi-level phase shift key modulation. l Phase 5hift’Keyed; MPSK) same as the data signal. A block diagram of a preferred transceiver arrangement for transmitting data; FIG. Typical voices generated by male and female speakers and default port Plot the power density (db) versus frequency averaged for the speed data signal. Figure shown; Figure 3 shows the BPSK data carrier frequency range from 500 to 2500Hz (D range) and Regarding loud Gaussian noise, 5o.

data-to-voice power ratio (data-t) for a data bit rate of bits per second Bit error for o-speech power ratio: DSPR) - A diagram showing a typical curve of speed (Bit Error Rate: BER); Figure 4 shows 250 and 10 when the BPSK data carrier frequency is 2500Hz. Produce a typical ER vs. DSPR curve for bit rates between 0.00 bit/s. This is a diagram that was cut.

Preferred systems for simultaneously transmitting analog voice and data signals according to the invention A blower diagram of a preferred embodiment is shown in FIG. This system includes a transmitter 10 receiving as input audio and data signals from a signal source; There is. The audio signal can be bandpass filtered in an optional filter 2, if desired. For example, a typical frequency band of 200 Hy to 3200 Hz can be used. Ru. The resulting audio signal 5(t) is then scale-linked by a factor α in a multiplier 14. It is sent to adder 16.

The input data signal is at a predetermined carrier frequency f at modulator 18. Modulated by

This carrier can be used for multiple levels within the analog signal frequency band of, for example, 2500H2. Multilevel Phase 5hift Keyed: MPSK) typically in the form of a carrier. This modulation is an amplified cosine pulse performed to generate an MPSK modulated data signal D(t) that may include a signal waveform. be exposed. The resulting typical MPSK modulated data signal is applied to adder 16. transmission on analog transmission channel 20. A signal X(t) is generated. The transmitted signal is expressed as x(t)-D(t)+α5(t) It will be done.

In this system, the transmitted signal X(t) travels on an analog transmission channel. first As an approximation to +n　)　. That is, x(t) - (D(t) ten α5(t))"Hch(t) - (D(t)*Hch( t))+(α5(t)*Hch(t)) (1) The receiver 30 Rear reconstruction device 32 and o: converts the received signal X in a conventional manner using an MPSK demodulator 33. Restore the data part of (t). The demodulator 33 includes an encoding and determining section, which is (a) the received data signal for transmission to both the receiver first output and the remodulator 34; (b) Phase jitter and frequency in the data signal of the received composite signal X(t) It has the ability to generate phase error information related to wavenumber offsets. phase error The information signal may be e.g. Contains raw information related to the source frequency and certain harmonics of this. United States 7 For example, these frequencies are 60, 1'20, and 180 Hz. Yo For example, in a Dew. This raw phase error signal is processed in a phase error tracking circuit 38 to find an appropriate position. Generates a mismatch signal. Phase error tracking circuit 38 may include any suitable method known to those skilled in the art. It is to be understood that circuitry may be included. For example, that of the frequency of interest Separate bandpass filters for each: e.g. a low pass filter that passes 500H21 Or, for example, on March 16, 1982, R. D. Gitlin (R. .

Gitlin, D., U.S. Pat. No. 4,320,526. This is an adaptive phase jitter tracker. The efficiency of the data signal restoration section of the receiver 3o is determined by the system greatly depends on the parameter α. From equation (1), data signal D(t) is an audio signal It is clear that 5(t) must be detected where it is present. Sith system parameter α is the amount of audio power σ2 required for reliable data restoration. It is adjusted to be as small as possible.

The audio signal is converted from the properly synchronized composite signal X(t) to the data △ It is restored by subtracting the signal D(t) component. This first This is done by generating a data signal D(t) in modulator 34. 34 is sent It corresponds in function to the MPSK modulator 18 of the device 1o. MPSK remodulator 34 The second timing is obtained from the carrier recovery circuit 32. In addition, the phase error tracking circuit The phase error information from path 38 is channeled into the regenerated data signal D(t). child This is the received signal when the regenerated data signal is subtracted from the received composite signal X(t). substantially alters the removal of the data signal within the resulting recovered audio signal at the second output of the device. If you do good, it will be done in the υ direction. The remodulator may include any suitable circuitry such as: I can do it. For example, if the data signal is modified by the recovered phase error information, a first phase encoder including a phase modulator for converting the signal into a phase difference encoded signal; , a second part modulating the resultant signal from the first part into a regenerated data signal D(t). This is the modulation section. to data signal D(t) is the audio signal 5(t) until the influence of channel 20 becomes apparent. is not directly subtracted from the received composite signal '52'(t) to recover . To do this, we must estimate the channel response Hch(t). do not have. Thereafter, the audio signal (5ct) is restored as follows.

↑(t)-[(D(t)*ach(t))+(α5(t)%rch[t))]- i(Tsuji1ch(t))(2) Data signal D(1) is known and changes arbitrarily Estimate the channel response Heh(t) when the audio signal 5(t) is unknown. This problem is solved by using an adaptive filter 35 according to the present invention. Currently heavy The detection is performed using the least mean square (Least Mean 5 square) through the device 36. Adaptive finite impulse response (F inite ImpulseRespOnse: F I R) If the filter is It is used in the response filter 35. Typical devices are Bee Widrow (B, W Ad apt 1veNO4se Cancelling: Pr1nciples and Applications-Proceedings PrOceedings of the IEEE, Volume 63 No. 12, December 1975, page 1709, Figure 29.

The efficiency of the MPSK receiver including the carrier recovery circuit 32 and the MPSK demodulator 33 is This can be easily understood along with the interference caused by interference. However, when the interference is audio, the receiver Efficiency requires special attention. Is the frequency distribution of white Gaussian noise uniform? Then, the data bit error rate (Bit-Error-Rate: BER) In this case, the MPSK carrier frequency is not important. Audio power density varies with frequency It is not uniform, but rather as the frequency increases, as shown by curve 40 in Figure 2. It decreases rapidly over time. In this case, the MPSK carrier frequency is seems to play an important role. Because it's damaging data. This is because only the interference portion is within the same band as the signal. Binary phase shift key (B inary Phase 5hiftKeyed: BPSK) Career lap A typical data signal with a wave number of 2500 Hz and a paw rate of, for example, 250 is shown in Figure 2. is shown as a curve 40 superimposed on an audio signal curve 40 in FIG.

For a given data-to-8peec POwe r RatiO: DSPR) using a matched FIJ letter receiver. , better BER when selecting a higher carrier frequency as shown in Figure 3. It has been found that efficiency can be achieved. Figure 4 shows when different data rates are used. , shows the BER efficiency for different DSPHs. In Figure 4, the BP used The SK carrier frequency is typically 2.5 kHz and the given Higher data rates require higher DSPR for lower BER. . As mentioned above, the parameter α determines the voice power for reliable data recovery. is adjusted to be small enough to The value of α can be easily determined from the DSPR. I can't stand it. That is, The audio signal is recovered from the received data and audio composite signal at the receiver 30. The degree to which It is limited by the ability of the party to estimate the cost. The adaptive FIR filter 35 However, the regenerated data from the remodulator 34 If the data signal D(t) is convolved with the arbitrarily changing impulse response H(t), It has been found to be very effective in solving such problems. the The resulting signal is then subtracted into D(t) by any suitable means in subtractor 3t. Synchronized composition/＼ Subtracted from signal X(t) to leave restored audio S(t). appropriate means is, for example, the delay at the ten input terminal of subtractor 37 in FIG. restored sound To improve the voice estimate, we use the least mean square (Least Mean 5 square : LMS) A11 After that, H(t) converges to Hch(t) while the response H(t) changes arbitrarily. So The restored voice at the output of the subtractor 37 has a No or almost no noise is included. The phase error information is re-modulated into the data signal D ( t), the adaptive filter 35, which does not have the ability to compensate for phase errors, is Can assist.

The parameter μ controls how quickly filter 35 converges. The older you get, the faster you adapt. However, if the value of μ is too large, it becomes unstable. In addition, if μ is small, the final’ The error between 9(t) and Hch(t) becomes much smaller. The theory of adaptive filters is The aforementioned article by Widrow et al. (December 1975 issue of Proceedings Procedures of the I EEE)).

As a typical example, a length 64 FIR filter and a 1o-90μ are used to Data removal of around 3db was achieved.

The application of the adaptive filter 35 described herein is based on the bandwidth of the input data signal D(t). is a special case when the data does not occupy the entire analog transmission channel bandwidth. in this case There exists a group of solutions for adaptive f12 78 Table Sho GO-501086 (6). circuit 3 After the LMS algorithm from 6 converges, H(t) is will continue to change until one of those solutions is reached, which produces an arithmetic error when this problem A simple solution for The solution is to remove the filter. The resulting signal D(t) should cover a wide band. cormorant. There is only one solution for the adaptive filter, and the RC filter response and the convolved chip The channel response will consist of He h (t).

The recovered voice is calculated by removing channel response, additive channel noise, and imperfect power of the data signal. Please understand that it deteriorates due to aging. Audio signal to measure the quality of restored audio The noise ratio (Signal-t, :+-Noi+8e Ratio: SNR) used.

SNR is expressed as follows.

Nch is the power of the additive channel noise, and ND is the power of the removed data signal D(t). The power of the noise generated by σ8 is the power of the audio signal. Previously, smaller It was mentioned that a better BER can be obtained by increasing the value of α. However, equation (5) SN of the restored speech decreases with Riiα, and furthermore, if α is a very small value It can be said that the quality of the restored audio deteriorates if is a critical system parameter in determining the best compromise between recovered data and voice efficiency. This is one of the parameters.

3 It is understood that the foregoing examples are merely illustrative of the invention. stomach. Various other modifications and variations embodying the principles of the invention and falling within the spirit and scope of the invention. Variations will occur to those skilled in the art. The analog transmission channel 20 can take various forms. , e.g. a general signal operating in the range 0-4000Hz with unknown amplitude and frequency distortion. It should be understood that the telephone channels may include telephone channels.

Brief description of the drawing F/ni, / bit error rate bit error rate

Claims (1)

[Claims] 1. Predetermined power density characteristics over a predetermined bandwidth There is an analog audio signal having a of the bandwidth of a received analog audio signal such that the power density characteristic of the signal has a low value. An input terminal that has the ability to simultaneously receive a modulated data signal received by a part and; first and second output terminals; Demodulates the data signal from a composite signal of the received analog voice and modulated data. a phase error detected in the received data signal; means (33) capable of generating a signal; at the output of said demodulating and restoring means; The phase jitter and frequency offset while remodulating the recovered data signal An information signal is introduced to add substance to the data signal received at the input terminal of the receiver. means (34) capable of generating an output signal corresponding to the phase error signal and including the phase error signal; ; A number representing the estimated impulse response of the channel connected to the input terminal of the receiver. 1 signal, the first signal being the remodulated data output signal from the remodulating means. Adaptive filter means 19 (3) capable of convolving to generate a resulting output signal. 5) and; The resulting output signal generated by the adaptive filter means (35) is transmitted to the receiver. subtracted from the composite signal of analog audio and modulated data received at the input terminal. removes the data signal forming part of the combined received signal; Ability to generate a resultant output signal comprising the recovered analog audio signal at an output terminal. means (37) for having; A receiver comprising: 2. In the receiver according to claim 1, the phase error multiple signal from the demodulating means in response to the composite signal traveling through the channel connected to the input terminal of the receiver. one of the selected frequencies or selected frequencies associated with the interfering signal resulting in a phase error; phase error tracker with the ability to track the phase error in either of the frequency bands A receiver characterized by trace means (38). 3. In the receiver according to claim 2, the remodulation means (34) comprises: converting the demodulated data signal at the output of the demodulating means (33) into a phase difference encoded signal; and a phase encoding means capable of reintroducing the phase error signal into the encoded signal. , of the remodulation means in response to the phase differentially encoded signal at the output of the phase encoding means. modulation means for generating the output signal. 4. In the receiver according to claim 1, the adaptive filter means (35) comprises: l6 generating the first signal;1. The first signal is converted into the remodulation data from the remodulation means (34). means having the ability to convolve with the data output signal; the resulting output signal from said subtraction means (3γ) and said remodulation data from said remodulation means; the first signal generated by the generation/convolution means in response to the data output signal; the second output terminal of the receiver to maximize the rejection of the data signal; and means for forming a resultant output data signal of the corresponding filter means. receiver. 5. In the receiver according to claim 4, the adaptive filter means (35) The modifying means comprises successive synchronous samples of the output signals of both the remodulation means and the subtraction means. Apply the least mean square algorithm to the pull to control the generation/convolution method of the impulse response of the channel that generates the signal and is connected to the input terminal of the receiver. including a device (36) for converging the estimated value to the actual channel impulse response; A receiver characterized by. 6. In the receiver according to claim 1, the adaptive filter means (35) to generate a first signal that is an estimate of the impulse response of the analog transmit channel. A receiver characterized in that: 7. In the receiver according to claim 1, the reception modulation at the input terminal A receiver characterized in that the data signal is a multi-level phase shift key data signal.