Time diversity for data communication over a voice channel

TECHNICAL FIELD [0001] The present invention relates to wireless telecommunications, and in particular to combining time diversity of digital data transmitted over a voice channel of a wireless telecommunications network.

BACKGROUND OF THE INVENTION [0002] A lot of telephony used, transmit communication.

Telecommunications terrestrial networks is up to voice effective channels (vocoder) components in cellular designed voice signals

For example, a digital voice coder uses linear prediction coding techniques to represent voice signals. Such linear predictive coders filter out the noise (non-voice signals), while compressing and estimating the frequency components of the voice signals before transmitting them on the voice channel.

[0003] It is sometimes desirable to transmit both audio signals and digital data in a wireless telecommunications network. For example, the cellular makes a combination of getting help in an emergency state, the user may wish to send digital location data to a call center on the same channel that is used to verbally describe the conditions of the operator's condition. However, it may be difficult to have digital data signals on the channel when the telephone user with the number "911" in order to threaten the person transmitting the voice network

As such, these signals are subject to several types of distortion.

[0004] For example, a digital data signal propagated on the wireless network voice channel may be distorted by the vocoder effects caused by the voice compression algorithm. In addition, digital data signals may be distorted by network effects caused by poor RF conditions and / or high traffic in the network. These distortions introduce bit errors that can be overcome with the use of techniques such as, for example, error correction (FEC) and repeated bit sequence transmissions.

[0005] Due to the fact that there are many types of vocoders (e.g. EVRC, AMR and others) and many possible network conditions, it is difficult to anticipate the quality of the voice channel and the associated bit error rate in advance. In addition, the quality of the voice channel can change rapidly over time. Therefore, it is difficult to design an efficient FEC scheme that minimizes the number of bits required for error correction while providing acceptable transmission performance in a low quality channel environment. For example, a FEC scheme with a very small number of error correction overhead bits can produce acceptable performance on a high quality channel with a small number of errors requiring correction, but if the quality of the channel deteriorates,

SUMMARY OF THE INVENTION [0006] The above-mentioned existing systems are successfully provided without solved solved occurring in the examples

To make the present application understandable by reading and studying the following description.

[0007] The embodiments of the invention are set out in the appended claims.

[0008] The foregoing and other features and advantages of the invention will be better understood from the following detailed description of preferred embodiments of the invention that will be shown with reference to the accompanying drawings.

BRIEF DESCRIPTION OF THE DRAWINGS [0009] The same reference numbers and markings in the various drawings indicate the same elements.

[0010] FIG. 1 is a diagram showing a wireless communication network that realizes in-band signaling (IBS).

[0011] Figure 2 is a schematic diagram of digital tone output from an IBS modem.

[0012] FIG. 3 shows a process for transmitting digital data in a wireless communication network.

[0013] FIG. 4A is a diagram of a conventional receiver for receiving digital data transmitted over a wireless communication network.

[0014] Figure 4B shows a schema of a receiver with a component combining time diversity.

[0015] Figure 5 is a flowchart showing the operation of the receiver shown in Figure 4B.

DETAILED DESCRIPTION OF THE INVENTION [0016] In the following detailed description, reference is made to the accompanying drawings, which form a part thereof and which are illustrated by way of illustration

Certain embodiments in which the invention may be embodied. These embodiments are described in sufficient detail to enable those skilled in the art to practice the present invention and it should be understood that other embodiments may be employed and various changes may be made without departing from the scope of the present invention as defined by the appended claims. The following detailed description should not therefore be interpreted as a restriction.

[0017] Referring to Figure 1, the wireless communication network 12 includes a mobile phone 14 that receives voice signals 22 from the user 23. The voice encoder (vocoder) 18 present in the mobile phone 14 encodes the voice signals 22 to form the encoded digital voice signals 31, which are then transmitted on a wireless digital voice channel 34 (cellular call). The cellular telephone 14 transmits the encoded voice signals 31 to the cellular communication side (cell side) 36, which forwards the cellular call to the cellular telecommunications switching system (CTSS) 38.

[0018] The CTSS system 38 either combines a cellular call with another cellular telephone or wireless cellular network 12, with a fixed line telephone in the PTSN 42 network as a peripherally switched connection, or routes a cellular call on a packet switched network according to the Internet Protocol (IP) protocol as Voice Over IP (VOIP) connection. The cellular call may also be routed from the PSTN 42 network back to the cellular network 12 or from the PSTN 42 network to the IP 46 network or vice versa. The cellular call finally arrives at telephone 44, which corresponds to the target telephone number originally entered in the cell phone 14.

[0019] The in-band signaling (IBS) 28 enables the mobile phone 14 to transmit digital data 29 from the data source 30 on the digital voice channel 34 of the cellular network 12. The IBS 28 modem modulates digital data 29 to synthesized digital data tone 26. The term "digital data tones" as used herein refers to sound tones that are modulated to encode bits of digital data. Digital data tones 26 prevent the coding components in the cellular network 12, such as vocoder 18, from making excessive digital data distortions. The coding and modulation scheme used in the IBS 28 modem allows transmission of digital data 29 through the same vocoder 18 used in the cellphone 14 to encode the voice signals 22. The IBS 28 modem allows transmission of voice signals 22 and digital data 29 on the same voice channel using the same electronic cell phone circuits. This prevents the user from transmitting digital data using a separate wireless modem and allows the user of the mobile phone to talk and send data during the same digital wireless connection. The digital data 29 is modulated to form a sound signal in the voice band. This prevents the cellphone 18 from filtering or over-distorting the binary values ​​associated with the digital data 29. The same transceiver and coding circuit of the cellular phone is used to transmit and receive both voice and digital data. This allows the IBS 28 modem to be much smaller, less complex and more energy efficient than a free-standing wireless modem. In some embodiments, the IBS 28 modem is implemented entirely on the z software platform that the IBS 28 modem could be much smaller, less complicated and more energy efficient than a free-standing wireless modem. In some embodiments, the IBS 28 modem is implemented entirely on the z software platform that the IBS 28 modem could be much smaller, less complicated and more energy efficient than a free-standing wireless modem. In some embodiments, the IBS 28 modem is implemented entirely on the z software platform

By using only existing hardware components in a mobile phone 14.

[0020] One or more servers 40 are located at different locations in the wireless network 12, PSTN 42 network or IP 46 network. Each server 40 includes one or more IBS 28 modems that encode, detect and decode digital data 29 transmitted and received on the digital voice channel 34. The decoded digital audio tones 26 are either processed in the server 40 or routed to another computer, such as a computer 50.

[0021] Figure 2 shows one embodiment of synthesized digital data tone 26 that is transmitted and received by the IBS modem 28. In the illustrated embodiment, the IBS 28 modem uses a frequency shift binary keying (FSK) modulation scheme in which each data bit digital 29 is converted into one of two different tones. In other embodiments, other various suitable modulation schemes may be used. For example, the IBS 28 modem can use a 4-ton FSK scheme in which a different sinusoid frequency is assigned to each of the four possible quadruple values ​​(represented by two-bit sequences: "00", "01", "10" and "11"). Alternatively, a binary phase shift keying (PSK) modulation scheme may be used,

[0022] Referring again to the example with the binary FSK shown in figure 2, the first tone is generated at frequency fx and represents the binary value "1" and the second tone is generated at frequency fi represents the binary value "0". For each bit in the transmission sequence,

The transmitter sends a sine wave with the frequency fi (for the binary value "1") or the frequency f (for the binary value "0") within the duration of the interval One In some embodiments, the frequencies fi and fo are in the range of about 200 Hertz ( Hz) to about 3,500 Hertz, which, as noted, is an effective frequency range for generating tone 26 data that represents binary bit values. For example, in one embodiment, the frequency fi is about 500 Hertz, while the frequency f0 is about 900 Hertz. In another embodiment, the frequency fi is about 2100 Hertz, and the frequency f is about 2500 Hertz. The IBS modem 28 Sines and Kosinuses, which are generating digital values ​​that represent different amplitude and phase values ​​for the frequencies fi and f0.

[0023] In some embodiments, the digital data is outputted on the digital voice channel 34 at a transmission rate in the range of about 100 bits per second to about 500 bits per second, which, as noted, is an effective bandwidth range to prevent degradation of digital audio data by various voice coders of mobile phones. For example, in one embodiment, digital data is outputted on a digital voice channel 34 at a bit rate of 400 bits per second. In this embodiment, it contains tables used for a sinusoid for each tone f of the zero point amplitude of a millisecond. At speed and f0, they start and last through sampling 8,000 and end with about 2.5 samples per second,

[0024] Figure 3 shows the process of transmitting a digital data packet 70 on a digital voice channel 34 of the wireless communication network 12 that implements

Combining time diversity according to embodiments of the present invention. In the illustrated embodiment, the digital data packet 70 comprises a sequence of K bits that may represent a single packet or a frame of a longer message payload. The message payload can be divided into packages of various sizes and formats using a variety of appropriate techniques, such as those described in U.S. Patent No. 6,690,681 entitled "InBand Signaling For Data Communications Over Digital Wireless Telecommunications Network" published February 10 2004. In some embodiments, each data packet 70 includes about 100 data bits (i.e., K & 100), which may include a plurality of header bits, synchronization pattern bits, bits of a checksum,

[0025] Block 200 adds error detection bits, such as a cyclic redundancy check (CRC) code to digital data packets 70 to be transmitted. This creates a data sequence containing M bits, where (M - K) represents the number of error detection bits. In some embodiments, block 200 adds about 16 error detection bits (i.e. M - K & 16). Block 205 adds bits of the error detection overhead to the M-bit data sequence, such as the Bose-Chaudhuri-Hocquenghem code (BCH), the Reed-Solomon code, or the convolutional error correction code. This creates a complete transmission sequence containing N bits, where (N - M) represents the number of error correction overhead bits. In some embodiments, the complete transmission sequence comprises a total of about 186 bits (i.e., N & num number)

[0026] Bl ok. 210 modulates the N-bit data sequence to form synthesized digital data tones 26 including a voice band audio signal suitable for

The transmission of data on the digital voice channel 34 of the wireless telecommunications network 12 as described above. After the transmission, the block 215 demodulates the tones 26 of the digital data and generates a feature vector that is used to create an estimate of the transmitted data sequence. As described in more detail below, if the feature vector contains errors, the block 220 may perform combining the temporal diversity of the feature vector with the previous feature vectors 225 of the same N-bit data sequence (if any) that were previously transmitted.

[0027] Bl ok 230 performs error correction of the demodulated N-bit data sequence, and block 235 performs error detection of the resulting M-bit estimated data sequence. Error correction and error detection of the demodulated data signal can be made of appropriate techniques, which using many different ones are well known to the average software used by specialists. If no errors are detected, then the K-bit sequence of digital data is fed to its target recipient. Otherwise, the digital data packet 70 is retransmitted on the digital voice channel 34 of the wireless telecommunications network 12.

[0028] Figure 4A shows a traditional receiver 300 comprising a feature extraction module 310 and a bit sequence estimation module 320. The term "module" as used herein may refer to software for performing specific functions. It is recognized that the functions performed by the modules described herein may be implemented within a larger or smaller number of modules than is described in the attached text. For example, a single function can be performed as a result of the operation of multiple modules or more functions can be performed by the same module. In addition, the modules described may fit in a single one for any combination of hardware or equipment of a particular function or

Locations or at different locations connected via a wired or wireless telecommunications network.

[0029] As illustrated in Figure 4A, at time t1, the first demodulated signal is received by the feature extraction module 310. This demodulated signal includes a sinusoidal waveform divided into a series of sequential bit intervals. The feature extraction module 310 processes the waveform in each bit interval independently and in sequence to generate the first feature vector X. In general, the feature vector comprises a set of measurements taken on the demodulated signal to estimate the transmitted bit sequence.

[0030] For example, in embodiments implementing the binary FSK modulation scheme, as shown in figure 2, the feature vector comprises a Fourier size sequence. For each bit interval (e.g., 2.5 milliseconds at a bit rate of 400 bits per second), the feature extraction module 310 calculates the Fourier magnitude for each frequency f0 f0. These two quantities are marked as S (fx) and S (f0) respectively. The size S (fx) - S (f0) is then recorded as X ;, the "soft value" for the i-th bit interval. The sequence N of soft values ​​(Xi, X2, X3, ..., Xn) represents the vector of the characteristics of the N-bit data sequence. [0031] In other embodiments, the feature vectors may contain various other appropriate measurements. For example, in embodiments implementing a binary PSK modulation scheme, the feature vector contains a sequence of cross-correlation values. For each bit interval, the feature extraction module 3i0 calculates the cross correlation value between the received waveform and each of the two sinusoids: one with a phase of 0 degrees and one with a phase of 90 degrees. These two correlation values ​​are designated as Sq and Si respectively. Size Sq

53 / 51P28826PL00

- Si is then recorded as Xi, the "soft value" for the bit interval.

[0032] The feature vector X is sent to the bit sequence estimation module 320, which analyzes the feature vector X and generates the corresponding estimated bit sequence. In the above-described example with the binary FSK modulation scheme, the size S (%) for each bit interval is proportional to the probability that the corresponding bit is a binary value of "1" and the magnitude of S (f0) is proportional to the probability that the corresponding bit is the binary value of "0". Therefore, if Xi or S (f1 S (fo) has a positive value then the bit sequence estimation module 320 means i-th bit of the estimated bit sequence as binary value "1". Otherwise, i-th bit is marked as binary value "0".

[0033] The purpose of the bit sequence estimation module 320 is to estimate the most likely sequence of bits represented by the feature vector X. In some embodiments, the bit sequence estimation module 320 applies the decision rule of one bit at a time, as described above. In other embodiments, other bit sequence estimation techniques may be used. when decoding the code, the bit decision is subject to neighboring bit periods.

[0034] In some cases, the estimated bit sequence generated by the bit sequence estimation module 320 contains errors and the sequence does not pass the next error detection test. In such cases, the traditional receiver 300, shown in Figure 4A, discards the first feature vector X and waits for retransmission of the N-bit data sequence.

[0035] At time t2, the feature extraction module 310 receives a signal,

For example, an individual convolutional, each influenced by the second demodulated observation which contains a repeated transmission of the original N-bit data sequence. Module 310

Since the traditional receiver 300 has rejected the first feature vector X, the bit sequence estimation module 320 analyzes the second feature vector Y independently of the first feature vector X. If the second estimated bit sequence also contains errors, then the process will be repeated until the error-free copy of the Nbit sequence or the transmission is overdue until the wireless network 12 can receive the error.

[0036]

Figure 4B shows a receiver including a feature extraction module 340 and a bit sequence estimation module including a component

330

350

360 combining time diversity according to embodiments of the present invention. Similar to the traditional receiver 300, at time t1, the first demodulated signal is received by the feature extraction module 340, which generates the first feature vector X representing the demodulated signal. The feature vector X is then fed to the bit sequence estimation module 350, which analyzes the feature vector X and generates the estimated bit sequence as described above. In the example shown in Figure 4B, the estimated bit sequence contains errors. However, the receiver 330 does not reject the feature vector X. Instead, the receiver 330 saves the first feature vector X in memory for later use by the component 360 combining time diversity.

[0037] At time t2, the feature extraction module 310 receives a second demodulated signal that includes repeated transmission of the original N-bit data sequence. The feature extraction module 310 then generates a second feature vector Y. If the estimated bit sequence based on the second feature vector Y contains errors, then the time combining component 360 may preferably analyze the second feature vector Y in combination with the first feature vector X that is stored in memory. Therefore, the estimation module 350

The bit sequence may generate an additional estimated bit sequence based on the combination of feature vectors X and

Y.

[0038] In certain embodiments, the feature vectors X and

Y are summed or averaged together to generate an additional estimated bit sequence. For example, in the binary FSK example described above, if the feature vectors X and Y contain errors, then the component 360 combining temporal diversity can generate a third feature vector V by adding the corresponding soft values ​​to each other as follows:

Vi = Χι + Υι, V2 = Χ2 + Υ2,, V_N - X_N+ Y_N [0039] The module 350 then creates the use of this bit sequence can the bit sequence by estimating the estimated same bit sequence estimation described above: if Vi is positive, then the ity bit is marked as binary value "1", otherwise i-th bit is marked as a binary value of "0".

[0040] For any bit rate error rate, the associated V vector will have a higher probability of producing a correct sequence estimate than the X or Y vector itself. The combined V vector is typically more accurate than the X vector itself or Y vector itself, since when a low quality channel is transmitted two separate copies of a given signal, they will probably be subject to distorting effects in different ways. Therefore, when they are taken together, their related feature vectors X and Y usually produce a better estimate of the transmitted sequence than any of them could alone.

[0041] If the combined V vector does not produce the correct estimated bit sequence, then the first and second feature vectors X and Y remain stored in memory. If

The next received feature vector Z (not shown) also fails, then the time combining component 360 may construct another new connected vector W by summing soft values ​​from all three received vectors as follows:

Wi = Xi + Yi + Zi, W2⁼X2 + Y2 + Z2, ..., W_N= X_N+ Y_N+ Z_N.

[0042] Alternatively, because there are an odd number of feature vectors, the value of the i-th bit can be assigned by the majority selection from Xi, Yi and Zi. The combined vector W is even more likely to produce an effective estimate than the vector V. Accordingly, by using the time diversity of the repeated transmissions, the receiver 330 can perform better and better estimates of the transmitted data sequence along with each repeated transmission.

[0043] Figure 5 shows the operation of a receiver 330 having a time diversity combining function. Block 400 represents the start of the process when the data signal is received and demodulated by the IBS 28 modem. At this point in the process, the IBS 28 modem detected a data signal and performed synchronization and other steps necessary to demodulate the signal. As described above, the data signal may indeed contain any desired sequence of bits, such as, for example, a data packet representing a portion of a payload of the message data.

[0044] Bl about 405 extracts the feature vector from the demodulated data signal. Block 410 generates an estimated bit sequence of some component based on the extracted feature vector. In embodiments, the estimation requires error correction, such as BCH code, ReedSolomon code, or convolutional error correction code. Decision block

415 determines if the estimated bit sequence contains

53 errors. In some embodiments, the determination includes a CRC error checking algorithm.

If no errors are detected then block 420 removes similar feature vectors (if any) stored in the memory of the receiver 330. Generally two or more feature vectors are considered "similar" if they correspond to a single data sequence. In some embodiments, feature vectors corresponding to different data sequences may be stored in the memory of the receiver 330 at the same time. In these embodiments, similar feature vectors can be identified by determining whether a certain percentage of common bits between two data vectors exceeds a selected threshold value, such as 80%. After identifying and deleting similar feature vectors, block 425 sends a confirmation signal (ACK) to the transmitter and the process ends in block 430.

[0046] If errors are detected in decision block 415, block 435 stores the current feature vector in the memory of the receiver 330. The optional decision block 440 determines whether any similar feature vectors are stored in the memory. As described above, this determination can be made by verifying that any other feature vectors stored in memory have a desirable threshold for the percentage of bits in common with the feature vector under consideration. If not, the digital data signal is retransmitted and the control returns back to block 405, in which the feature vector is extracted from the retransmitted data signal.

[0047] In the illustrated embodiment, if decision block 440 determines that similar feature vectors are stored in memory, block 450 generates one or more additional bit sequence estimates based on a combination of similar feature vectors, as described above . In other embodiments, the block 450 generates

Additional estimated bit sequences by combining all feature vectors stored in the memory, regardless of whether they are similar. Decision block 455 determines whether an additional estimated bit sequence (additional estimated bit sequences) contains any errors. If not, then block 420 deletes similar feature vectors stored in memory and the method continues as described above.

[0048] If errors are detected in decision block 455, then the digital data signal is retransmitted and the control returns back to block 405 as described above. In some cases, the transmitter may terminate the process before receiving a non-error copy of the digital data signal. For example, the transmitter may stop retransmission of the digital data signal after a predetermined number of repeated retransmissions or a selected period of time that elapses from the first failed transmission.

[0049] The transmission duplicating systems and methods described above have a number of distinct advantages over traditional methods. For example, in memory, the time diversity receiver 330 can extract more information from failed transmissions than the traditional receiver 300. Therefore, the receiver 330 can often produce an error-free information sequence using fewer repeats than the traditional receiver 300. Alternatively, more can be used. an efficient FEC scheme with fewer bits of overhead for error correction than is required by traditional systems.

[0050] Although the present invention has been described with reference to certain preferred embodiments, other exemplary embodiments are also within the scope of the present invention, which are obvious to those of ordinary skill in the art, memorizing feature vectors with the 360 ​​combining component

Including embodiments which do not include all the features and advantages set forth herein. Accordingly, the scope of the present invention is defined only by reference to the appended claims.