HK1174160A - A system for communicating digital data over a voice channel of a digital telecommunications network - Google Patents

Description

System for transmitting digital data over voice channel of digital communication network

The present application is a divisional application of the invention patent application having an application number of 018132073, application date of 2001, 6/22, entitled "enhanced in-band signaling for data communication in a digital wireless communication network".

Technical Field

The present invention relates to wireless communications, and more particularly to systems for communicating digital data over voice channels "in-band" in digital wireless networks.

Background

Cellular telephones allow a user to talk to another user without being tied to a "land line". The cellular telephone includes circuitry for extracting speech signal samples from the user's speech. These speech signals are converted to digital form using an a-D converter. The digitized voice signal is encoded by a voice encoder and then modulated onto a carrier frequency that transmits the voice signal over a cellular network. The voice signal is transmitted through the wireless cellular network to another telephone in the wireless cellular network or to another telephone in the landline telephone network.

Different coders/decoders (codecs), modulators, vocoders, Automatic Gain Controllers (AGC), analog-to-digital converters (a/D), noise reduction circuits, and digital-to-analog converters (D/a) are used in cellular and landline telephone networks. These telephony components may implement different encoding schemes for encoding and decoding speech signals.

The communication components are designed to efficiently communicate voice signals over wireless and landline voice communication channels. For example, digital speech coders use predictive coding techniques to represent speech signals. These predictive coders filter out noise (non-speech signals) while compressing and estimating the frequency components of the speech signal before transmitting the speech signal over the speech channel.

Problems occur when voice communication devices, such as voice encoders, are used to communicate digital data. A speech encoder may interpret signals representing digital data as non-speech signals. The speech encoder may completely filter out or corrupt these digital data signals. Thus, digital data cannot be reliably transmitted over the same digital voice channel used to transmit voice signals.

Sometimes a user must simultaneously transmit both voice signals and digital data to another location. For example, when a cell phone user calls "911" for emergency assistance, the user may need to send digital location data to a call center while verbally explaining the emergency to the operator. The digital data is preferably transmitted through the cellular telephone without the use of a separate analog radio modem.

A need therefore exists for communicating digital data over a voice channel of a digital wireless communication network.

Disclosure of Invention

An inband signaling modem communicates digital data over a voice channel in a digital wireless communications network. The input receives digital data. The encoder converts the digital data into audio tones that synthesize frequency characteristics of human speech. The digital data is also encoded to prevent speech encoding circuitry in the communication network from corrupting the synthesized audio tones representing the digital data. The output then outputs the synthesized audio tones to a voice channel of a digital wireless communications network.

According to the present invention there is provided a system for communicating digital data over a voice channel of a digital communications network, comprising:

an input to receive digital data;

a processor for converting said digital data into audio tones; and

outputting the audio tones to an output of a digital transmission circuit, wherein the digital transmission circuit encodes the audio tones in the same manner as the voice signals are encoded and transmits the encoded audio tones over the same voice channel in the digital communication network as is used to transmit the voice signals.

Drawings

The foregoing and other features and advantages of the invention will become further apparent from the following detailed description of the preferred embodiments of the invention, read in conjunction with the accompanying drawings.

Fig. 1 illustrates a wireless communication network providing in-band signaling (IBS) according to the present invention.

Fig. 2 illustrates in detail a cellular phone coupled to an IBS modem according to an embodiment of the present invention.

Fig. 3 is another embodiment of an IBS modem according to the present invention.

Fig. 4 illustrates an IBS modem encoder in detail.

Fig. 5 is a schematic diagram of IBS packets.

Fig. 6 is a schematic diagram of digital data tones output from an IBS modulator.

Fig. 7 shows how the automatic gain controller corrupts the digital data.

Fig. 8 illustrates how a digital wireless network can filter out digital data tones.

Fig. 9 shows in detail a receiving circuit coupled to an IBS modem decoder.

Fig. 10 is a state diagram of the IBS decoder shown in fig. 9.

Fig. 11 is a block diagram illustrating a search state in an IBS decoder.

Fig. 12 is a block diagram illustrating an active state in an IBS decoder.

Fig. 13 is a block diagram illustrating a clock recovery state in an IBS decoder.

Fig. 14 is a schematic diagram of a cellular telephone when the IBS modem is in a detachable battery pack.

Fig. 15 is a schematic diagram showing different data sources coupled to a cell phone via an IBS modem.

Fig. 16 is a schematic diagram showing an implementation of an IBS modem using a sound card.

Fig. 17 and 18 are block diagrams showing how the sound card in dispute 16 works.

Fig. 19 is a block diagram of a synchronization circuit of an IBS modem.

Fig. 20 is a detailed diagram of the synchronization circuit in fig. 19.

Fig. 21 is a timing chart showing how the synchronization circuit in fig. 19 operates.

Fig. 22 shows how the synchronization circuit determines the optimum synchronization start time.

FIG. 23 is an alternative implementation of a synchronization circuit.

Fig. 24 is a diagram of an encoder for a multi-channel IBS modem.

Fig. 25 is a decoder diagram of a multi-channel IBS modem.

Fig. 26 and 27 show different channel configurations for the multi-channel IBS modem shown in fig. 24 and 25.

Fig. 28 is an encoder diagram for a multicarrier IBS modem.

Fig. 29 is a decoder diagram of a multicarrier IBS modem.

Detailed Description

Referring to fig. 1, a wireless communication network 12 includes a cellular telephone 14 that receives a voice signal 22 from a user 23. The speech encoder 18 in the cellular telephone 14 encodes the speech signal 22 into an encoded digital speech signal 31 (cell call) that is transmitted over a wireless digital speech channel 34. Cellular telephone 14 transmits encoded voice signals 31 to a cellular communication site (cell site) 36, which relays the cell call to a cellular communication switching system (CTSS) 38.

CTSS 38 either connects the cell call to another cellular telephone in wireless cellular network 12, or as a circuit-switched call to a landline telephone over PSTN network 42, or sends the cell call as a voice over IP (VoIP) call over packet-switched Internet Protocol (IP) network 46. The cell call may also be sent from PSTN network 42 back to cellular network 12 or from PSTN network 42 to IP network 46, or vice versa. The cell call eventually arrives at telephone 44 corresponding to the destination telephone number originally entered on cellular telephone 14.

Additional data may be inserted at any point in the cellular network 12, such as in the PSTN network 42 and the IP network 46, and the signal remodulated for transmission over a wireline or cellular network. Such data may be system-related data such as routing information, long distance or tariff information, etc.

An in-band signaling (IBS) modem 28 enables the cell phone 14 to send digital data 29 from a data source 30 over a radio channel 34 of the cellular network 12. The IBS modem 28 modulates the digital data 29 into synthesized digital data tones 26. The digital data tones 26 prevent the encoded components in the cellular network 12 and the landline network 42, such as the voice encoder 18, from corrupting the digital data. The coding and modulation scheme used in the IBS modem 28 allows the digital data 29 to be transmitted through the same speech encoder 18 used in the cell phone 14 that encodes the speech signals 22. With this technique, appliances such as vending machines can be enhanced.

A synthetic tone is defined as a signal representing digital data that also has signaling properties that enable the signal to be encoded and decoded by a speech encoder without losing the digital data information in the signal. In one example, a Frequency Shift Keyed (FSK) signal is used to produce synthesized tones at different frequencies within the speech range of human speech.

The IBS modem 28 enables the voice signals 22 and digital data 29 to be transmitted over the same digital voice channel using the same cellular telephone circuitry. This avoids the user having to use a separate wireless modem to transfer digital data and allows the cell phone user to talk and send data within the same digital wireless call.

The present invention modulates the digital data 29 into synthesized voice tones. This prevents the cell phone vocoder 18 from filtering or corrupting the binary values associated with the digital data 29. The same cellular telephone transceiver and encoding circuitry is used to transmit and receive voice signals and digital data. This enables the IBS modem 28 to be much smaller, simpler, and more energy efficient than a stand-alone wireless modem. In some embodiments, ISB modem 28 is implemented entirely in software, using only existing hardware components in cellular telephone 14.

One or more servers 40 are located at any of various locations within wireless network 12, PSTN network 42, or IP network 46. Each server 40 includes one or more IBS modems 28 that encode, detect, and decode the digital data 29 transmitted and received over the digital voice channel 34. The decoded digital data is either processed at the server 40 or sent to another computer, such as computer 50.

Referring to fig. 2, the first transmission part of the IBS modem 28 includes an IBS encoder 52 and a digital-to-analog converter (D/a) 54. IBS encoder 52 is typically implemented using a Digital Signal Processor (DSP). Data source 30 represents any device that requires wireless transmission or reception of digital data. For example, the data source 30 may be a laptop computer, a palmtop computer, or a Global Positioning System (GPS) (see fig. 15).

The data source 30 outputs the digital bit stream 29 to the IBS encoder 52. The IBS encoder 52 converts the digital data 29 into specially formatted IBS packets for transmission over the digital wireless voice channel. The IBS encoder 52 then converts the bits from the IBS packets into digital data tones, which are then fed to the D/a converter 54.

The IBS modem 28 outputs binary values that each represent the amplitude and phase components of the audio tones. The D/a converter 54 converts these digital values to analog audio tones 26, which analog audio tones 26 are then output to the auxiliary audio port 15 on the cellular telephone 14. The cellular telephone 14 then processes the analog audio tones 26. An analog-to-digital (A/D) converter 16 in the cellular telephone 14 encodes the synthesized analog audio tones 26 into digital values. The speech encoder 18 encodes the digital representation of the synthesized tones 26 into encoded digital data 32 and outputs the encoded data to the transceiver 19, and the transceiver 19 transmits the encoded digital data 32 over a digital speech channel 34.

The preferred voltage of the synthesized audio tone 26 output from the D/a converter 26 is about 25 millivolts (positive and negative peak-to-peak amplitude). This voltage level may prevent the audio tone 26 from saturating the voice channel circuitry in the cellular telephone 14.

Since the digital data 29 is input through the existing hands-free audio port 15 in the cell phone 14, the IBS modem 28 can be installed as an after-market device that can connect any data source 30 to the cell phone 14. Data source 30 may communicate digital data 29 in any digital format. For example, digital data 29 may be sent over an RS-232 interface, a Universal Serial Bus (USB) interface, or any other serial or parallel interface.

Fig. 3 shows an alternative embodiment of the IBS modem 28. The IBS modem 28 in fig. 3 is located within the cell phone 14 and is implemented in software by using the existing cell phone processor or by using some combination of its own components and existing cell phone components. In this embodiment, cellular telephone 14 may include a data port 56 that receives digital data 29 from external data source 30. In an alternative embodiment, the digital data source 30 is internal to the cellular telephone 14. For example, the data source 30 may be a Global Positioning System (GPS) chip that includes a GPS receiver (not shown) for receiving global positioning data from GPS satellites (fig. 14).

The IBS encoder 52 in fig. 3 described above is typically implemented in software using a DSP, and the same DSP used to implement the speech encoder 18 can be used. The D/a converter 54 outputs the synthesized audio tones representing the digital data 29 to the internal a/D converter 16 of the cellular phone 14. The IBS encoder 52 in an alternative embodiment not only synthesizes the digital data 29 into audio tones, but also quantizes the digital frequency values. The IBS encoder 52 then directly inputs the quantized data 55 to the speech encoder 18. In yet another embodiment of the present invention, IBS encoder 52 is implemented entirely in software within the same DSP that implements speech encoder 18.

The speech coder 18 uses a particular coding scheme associated with the wireless communication network 12 (fig. 1). For example, the speech encoder 18 may be a VCELP encoder that converts speech signals to digital CDMA signals. The a/D converter 16, D/a converter 54 and transceiver 19 are existing cellular telephone components known to those skilled in the art.

It is important to note that ISB encoder 52 is capable of transferring digital data 29 using the same cellular telephone circuitry that transferred the voice signal. The IBS encoder 52 prevents any signal approximation, quantization, encoding, modulation, etc. by the a/D converter 16, speech encoder 18, or transceiver 19 from corrupting or filtering any binary bits of the digital data 29.

Fig. 4 illustrates in detail the IBS encoder 52 shown in fig. 2 and 3. Data buffer 58 holds binary bit stream 29 from data source 30. The packetizer 60 partitions the bits in the buffer 58 into bytes containing the IBS packet payload. The packet formatter 62 adds a packet preamble and postamble that help prevent corruption of IBS packet payloads. The IBS modulator 64 then modulates the bits in the IBS packet using two or more different frequencies 66 and 68 to produce digital data tones 69.

Preventing corruption of digital data in a voice channel

Cellular telephone vocoders increase bandwidth in the speech channel by utilizing predictive coding techniques that attempt to describe the speech signal without having to transmit all of the information associated with human speech. If any unnatural frequencies or tones (i.e., frequencies representing digital data) are generated in the voice channel, these frequencies may be rejected by the speech encoder 18 (FIG. 2). For example, if the amplitude of a digital data tone is greater than the amplitude of a normal speech signal, or the same digital data tone is generated for an excessively long period of time, the speech encoder 18 may filter out the high amplitude or extended frequency signal. Depending on the manner in which the digital data tones are encoded, the digital bits represented by these non-natural audio tones may be partially or entirely removed from the voice channel.

IBS encoder 52 encodes digital data 29 in a manner such that the speech encoder does not filter out or corrupt the tones representing the digital data. IBS encoder 52 does this by controlling the amplitude, time period, and pattern of the synthesized audio tones used to represent the binary bit values.

Referring to fig. 5, the packet formatter 62 (fig. 4) adds a packet preamble 73 and header 75 to the IBS packet 70. The packet preamble 73 includes a preamble pattern 72 and a sync pattern 74. Checksums 78 and post-packet synchronization codes 79 are appended to the back end of IBS packets 70.

Fig. 6 shows synthesized digital data tones 69 output from the IBS modulator 64 (fig. 4). The IBS modulator 64 (fig. 4) converts the digital bits in the IBS packet 70 into one of two different tones. The first tone is generated at the f1 frequency and represents a binary "1" value, and the second tone is generated at the f2 frequency and represents a binary "0" value. In one embodiment, the f1 frequency is 600Hz and the f2 frequency is 500 Hz.

It has been determined that the most efficient frequency range for producing tones representing binary bit values is somewhere between 400Hz and 1000 Hz. The IBS modulator 64 includes sine and cosine tables that are used to generate digital values representing different amplitude and phase values at the f1 and f2 frequencies.

In one embodiment of the invention, digital data is output over the voice channel 34 at a baud rate of 100 bits/second. This baud rate has been found to be effective in preventing corruption of digital audio data by a wide variety of different cellular telephone speech encoders. The sine wave of each of the f1 and f2 tones begins and ends at a zero amplitude point and lasts 10 milliseconds. For each digital data tone, 80 samples are produced.

Referring to fig. 7, an Automatic Gain Controller (AGC) 80 is one type of encoding function used in cellular telephone 14. AGC 80 may be software located in the same DSP that implements vocoder 18. AGC 80 scales instantaneous energy changes in the speech signal. There are cases when no speech signal is input to the AGC 80 for a period of time, followed by a series of audio tones 82 containing the beginning of the IBS packet 70. AGC 80 scales up a first set of tones 82 located at the beginning of IBS packet 70. After the end of the IBS packet 70, the AGC 80 also looks ahead at the zero signal level 84 and will scale up the tones 83 at the end of the IBS packet 70 as part of its predictive scaling scheme. This scaling prevents excessive amplification of the signal or noise when transient energy changes occur in the voice channel.

As previously shown in fig. 6, the "1" and "0" bits of the IBS packet 70 are represented by tones f1 and f2, respectively. If these tones are amplified by AGC 80, the digital bits represented by these frequencies may be discarded in the encoding. For example, speech encoder 18 may treat the amplified tones as noise and filter them out of the speech channel. To prevent inadvertent filtering of tones representing digital data, the IBS packet 70 in fig. 5 includes preamble bits 72 and postamble bits 79. The preamble bits 72 and the postamble bits 79 do not contain any digital data bits 29 from the data source, but include a number of victim bits that are not necessary for detection or encoding of the IBS packet 70. Thus, the tones produced for these victim bits in the preamble and postamble can be scaled or filtered by the AGC 80 without affecting any digital data contained in the IBS packet payload 76.

The bit patterns in the preamble 72 and sync pattern 74 are specifically formatted to further prevent corruption of the packet payload 76. A random sequence and/or an alternating sequence of binary bits "1" - "0" are used in the preamble 72 and/or the sync pattern 74. These alternating or random bit patterns prevent the adaptive filter (fig. 2) in the cell phone vocoder 18 from filtering out the tones representing the remaining bits in the IBS packet 70.

Referring to fig. 8, the adaptive filter adapts around the frequency currently being transmitted through the wireless network. For example, if a long period of the same f1 tone is currently being transmitted, the adaptive filter used in the cell phone may adapt around the f1 spectrum as shown by filter 86.

Another short tone at another frequency f2 may immediately follow the long-period f1 tone. If the filter 86 is too slow to accommodate, the first few f2 tones are filtered from the voice channel. If the filtered f2 tones represent bits in an IBS bit stream, these bits are lost.

To prevent the adaptive filter in the cell phone from dropping binary bits, portions of the preamble 73 include a random or alternating pattern of "1" to "0" bits. This pre-prepares the adaptive filter shown by filter 88. The preamble 73 (fig. 5) is intended to include a portion of the same bit sequence that may or may not be generated in the packet payload 76. For example, IBS encoder 52 may look ahead at the bit pattern in payload 76. Encoder 52 may then place a subset of the bits into a portion of the preamble to represent a sequence of bits in the packet payload.

This prepares the adaptive filter for the same duration and with the same f1 and f2 frequencies that may follow similar sequences in the IBS packet payload 76. Thus, the adaptive filter is less likely to filter out tones that actually represent the digital data being transmitted.

Fig. 9 is a block diagram of a receiving circuit 91 that receives voice and data signals in the radio channel 34. The IBS modem 28 also includes an IBS decoder 98 that detects and decodes the digital data tones transmitted in the voice channel 34. Receive circuitry 91 is located at CTSS 38 (fig. 1) that receives wireless transmissions from cell site 36 (fig. 1). The same receiving circuit 91 is also located in the cellular phone 14.

As previously described in fig. 2 and 3, the decoder portion of the IBS modem 28 can be located outside the cell phone 14 or can be located within the cell phone 14. Dashed line 104 represents IBS modem 28 located outside the cell phone and dashed line 106 represents built-in IBS modem 28 located inside the cell phone. The IBS modem 14 may be located at any telephone location in the PSTN network 42 or the IP network 46 (fig. 1). The receive circuit 91 may be different when the IBS modem 28 is coupled to a landline. However, the IBS modem 28 operates on the same principle by transmitting and receiving synthesized tones over a voice channel over a telephone line.

Signals in the radio channel 34 are received by the transceiver 90. The speech encoder 92 decodes the received signal. For example, the speech coder 92 may decipher signals transmitted in TDMA, CDMA, AMPS, etc. The D/a converter 94 converts the digital voice signal into an analog signal. The analog voice signal is then output from the audio speaker 17.

If the IBS modem 28 is outside the receive circuit 91, the a/D converter 96 re-converts the analog signal to a digital signal. The IBS decoder 98 re-demodulates any tones representing the digital data into digital IBS packets. The packet disassembler 100 disassembles the packet payload from the IBS packets 70 and stores the decoded digital data in the data buffer 102.

Fig. 10 is a state diagram illustrating how the IBS decoder 98 in fig. 9 operates. The IBS decoder 98 repeatedly samples and decodes the voice signal received from the radio channel 34. State 110 searches for tones in the speech signal that represent digital data. The IBS decoder 98 enters the active state 112 if the signal-to-noise ratio (SNR) of the tones in the frequency range of the digital data tones is greater than a predetermined value. The active state 112 collects tone samples. If at any time within the active state 112, the SNR is below the active threshold, or times out before enough tone samples are collected, the IBS decoder 98 returns to the search state 110 and begins searching for digital data tones again.

After collecting many samples, the IBS decoder 98 looks for bits that identify the preamble 73 (fig. 5) in the IBS packet 70. If the preamble 73 is detected, the IBS decoder 98 proceeds to a clock recovery state 114. The clock recovery state 114 is synchronized with the synchronization pattern 74 (fig. 5) in the IBS packet 70. The IBS decoder 98 then demodulates the packet payload 76 in state 116. If no preamble 73 is found, the IBS decoder 98 returns to the search state 110 to begin searching again for the start of the IBS packet 70.

The IBS decoder 98 demodulates all of the packet payload 76 and then performs a checksum 78 as a final confirmation that the valid IBS packet 70 has been successfully demodulated. Control then returns to the search state 110 to begin searching for the next IBS packet 70.

Fig. 11 is a detailed diagram of the search state 110 of the IBS decoder 98. The search state 110 uses in-band and out-of-band filtering. "in-band" as used in the following description refers to tones in the frequency range of a tone (500 Hz) representing a binary "1" value of digital data and a tone (600 Hz) representing a binary "0" value of digital data.

The first bandpass filter 118 measures (in-band) the energy of the signal in the voice channel in the frequency range of about 400Hz-700 Hz. The second band pass filter 120 (out-of-band) measures the energy of signals in the voice channel that are outside the range of 400Hz-700 Hz. The signal-to-noise ratio (SNR) between the in-band energy and the out-of-band energy is calculated in block 122. If a tone representing digital data is present in the voice channel, the energy measured by the in-band filter 118 will be much greater than the energy measured by the out-of-band filter 120.

If the SNR is below the selected threshold in comparator 124, the signal in the voice channel is determined to be either the actual voice signal or noise. If the SNR is above the threshold, the IBS decoder 98 determines that the tones represent in-band digital data. When digital data is detected, the IBS decoding 98 proceeds to the active state 112 (fig. 10), which begins searching for the start of the IBS packet 70.

Fig. 12 shows the active state of the IBS decoder 98. The search state 110 notifies block 130 when an in-band tone is detected in the voice channel. The samples of the audio tone are windowed in block 132 with a number of samples associated with a single binary bit. In one embodiment, 80 samples of the digital data tone are taken and padded with zeros, which are then associated with a Discrete Fourier Transform (DFT).

The first DFT has coefficients representing the 500Hz tone and is applied to the windowed data in block 134. If the samples contain a 500Hz tone ("0" binary bit value), the first DFT produces a high correlation value. The second DFT represents the 600Hz tone and is applied to the windowed data in block 136. If the windowed samples contain a 600Hz tone ("1" binary bit value), the second DFT produces a high correlation value. Block 138 selects a binary "0" or binary "1" bit value for the windowed data based on which of the 500Hz DFT or 600HzDFT yields the largest correlation value.

In decision block 140, the IBS decoder 98 continues to demodulate the tones until the preamble of the IBS packet 70 is detected. The IBS decoder 98 then proceeds to the clock recovery state 114 (fig. 13) to synchronize with the sync pattern 74 (fig. 5) in the IBS packet 70. If more bits need to be demodulated before the preamble 73 can be verified, then the decision block 140 returns to block 132 and the next 80 samples of the digital data tone are windowed and demodulated.

Fig. 13 depicts the clock recovery state 114 of the IBS decoder 98. After detecting the preamble 73 in the IBS packet 70 in the active state 112, the clock recovery state 114 demodulates the next series of bits associated with the sync pattern 74 (fig. 5). The clock recovery state 114 aligns the tone samples with the center of the associated filter described in the active state 112. This improves the accuracy of the decoder when demodulating the IBS packet payload 76.

Decision block 142 looks for sync patterns 74 in IBS packets 70. If after demodulating the next tone, no sync pattern 74 is found, decision block 142 in block 148 shifts the window for sampling sync pattern 74 by one sample. Decision block 150 then rechecks sync pattern 74. If sync pattern 74 is found, decision block 144 determines the power ratio of the sync pattern detected. This power ratio represents the confidence that the demodulator is synchronized to the sync pattern. The power ratio is compared to the power ratio obtained for different window shift sample positions. If the power ratio is greater than the power ratio of the previous sample location, the power ratio is saved as the new maximum power ratio in block 146.

If the power ratio of sync pattern 74 is less than the previously measured power ratio, then in block 148 the decoder shifts the sampling window by one sample position. The power ratio of the shift window is then determined and compared to the current maximum power ratio at decision block 144. The window is moved until the maximum power ratio of sync pattern 74 is found. The window offset value at the maximum power ratio is used to align the demodulator correlation filter with the center sample of the first bit 77 (fig. 5) in the IBS packet header 75.

The IBS decoder 89 then jumps to the demodulation state 116 (fig. 10) where the determined window offset is used to demodulate the remaining 500Hz and 600Hz tones representing the packet payload bits 76 and checksum bits 78. The demodulation state 116 correlates the f1 and f2 tones with the DFT in the same manner as in the active state (fig. 12). The checksum bit 78 is then used as a final check to verify that a valid IBS packet has been received and accurately deciphered.

Fig. 14 shows the IBS modem 28 located in the battery pack connected to the cell phone 14. The hands-free voice channel lead 200 couples the IBS modem 28 to a voice channel 202 in the cellular telephone 14. A switch 204 couples either the voice signal from the microphone 17 or the digital data tones from the IBS modem 28 to the voice channel 202.

The switch 204 is controlled by an on-screen menu (not shown) in the cellular telephone 14, or by a button 206 extending out of the rear end of the battery pack 208. The switch 204 may also be controlled by one of the keys on the keypad of the cellular telephone 14.

The button 206 may also be used to initiate other functions provided by the IBS modem 28. For example, a Global Positioning System (GPS) includes a GPS receiver 210 located in the battery pack 208. The GPS receiver 210 receives GPS data from GPS satellites 212. In an emergency situation, the cellular telephone operator need only press the button 206. Pressing the button 206 automatically causes the GPS receiver 210 to collect GPS data from GPS satellites 212. At the same time, the switch 204 connects the IBS modem 28 to the voice channel 202 of the cell phone 14. Thereby activating the IBS modem 28. Once the GPS data is collected in the IBS modem 28, the data is formatted, encoded, and output by the IBS modem 28 to the voice channel 202 of the cell phone 14.

The user 23 may press the button 206 at any time after manually calling a certain telephone number. After establishing a voice channel with another endpoint, the user 23 presses the button 206. The switch 204 is connected to the IBS modem 28 and the IBS modem 28 is activated. The GPS data (or other digital source) is then sent to the endpoint via the IBS modem 28 in the form of digital data tones over the established voice channel. After the data is successfully transferred, the user presses the button 206 again, causing the switch 204 to reconnect to the audio receiver 17.

Fig. 15 shows different types of data sources that may be connected to the IBS modem 28. The palm computer 212, GPS receiver 214, or computer 216, etc. may be coupled to the IBS modem 28. The IBS modem 28 converts the bit output from the above-described device into digital data tones that are then output over a radio channel 34 in the wireless network. Since data may be transmitted to another endpoint through cellular telephone 14, no separate wireless modem is required for any of devices 212, 214, or 216.

Implementation of in-band signaling modem in sound card

IBS modems can be implemented in standard computer sound cards. See fig. 16, such as Stillwater, Ok 74075; a Sound card such as Sound blast card produced by Creative Labs, inc., 1523 cimaron plata is included in the computer 250. The speaker output 253 of the sound card 252 outputs audio tones to a hands-free port 257 on a cellular telephone 258. The microphone input on the sound card 252 is connected to the speaker output of the cellular telephone 258.

The computer includes a processor 254 that converts the digital data into an audio format used by the sound card 252 to output synthesized audio tones. Cellular telephone 258 encodes and transmits the audio tones over a voice channel of a wireless communication network. Cell site 261 receives the transmitted audio tones and forwards the audio tones through PSTN network 263. Computer 262 is connected to telephone line 260 at the destination location of the telephone call. Another sound card 264 and a processor 266 in the computer 262 condition the audio tones into digital data. The digital data represented by the audio tones is displayed on the computer 262. Sound cards may be used for data encoding, decoding, or both. A sound card may be used at either computer 250 or computer 262, or at both computer 250 and computer 262.

Referring to fig. 16 and 17, in block 270, data files, GPS data, data entered by the user using a keyboard, or any other digital data are grouped and formatted into IBS packets by the computer 250. Packets and packet formatting are depicted in fig. 4 and 5. In block 272, the binary bit values in the IBS packets are converted to the digital format used by the sound card 252 (fig. 16) to generate the synthesized audio tones. For example, binary "1" bit values in an IBS packet are converted to a digital format representing a first f1 frequency tone and binary "0" bit values are converted to a second frequency tone. The f1 and f2 tones are generated in a manner similar to that described in fig. 6.

In block 274, the sound card outputs block tones representing binary bit values in a similar manner as the IBS encoder 52 and digital-to-analog converter 54 described in fig. 3. The cellular telephone 276 encodes the audio tones in block 276 and transmits the encoded audio tones over a voice channel in block 278 in a wireless communication network.

Referring to fig. 16 and 18, a cellular telephone call is established with respect to a destination telephone number. In block 280, the user picks up the ringing telephone line, or computer 262 (fig. 16) located at the destination end of the cellular telephone call, is programmed to detect the ringing signal from telephone line 260. If a ring signal is detected, the user or computer 262 generates an "off-hook" signal on telephone line 260 in block 282. In block 284, the sound card 264 functions similar to an analog-to-digital converter by converting the audio tones on the telephone line 260 to digital data. Similar to the IBS decoder 98 described in fig. 9-13, the sound card 264 and the processor 266 (fig. 16) together decipher the IBS audio tones. The digital representation of the detected IBS tones is then displayed on the screen of the computer 262 in block 290.

In one example, the user wishes to locate the cell phone 258. The user instructs the computer 262 (fig. 16) to dial the telephone number of the cellular telephone 258. The computer 262 uses the sound card 264 to send IBS tones that instruct the cell phone 258 to return GPS location data. The computer 250 may have a GPS receiver or the cell phone 258 may have a separate GPS receiver. If the GPS receiver and IBS modem are within the cell phone 258 as shown in fig. 2-9, the computer 250 need not be connected to the cell phone 258.

The GPS data is converted to IBS tones by the sound card 252 as shown in fig. 17 or by an internal IBS modem as described in fig. 2-9. IBS tones representing GPS data are sent back to telephone line 260 over the wireless communication channel and PSTN network 263. A sound card 264 in the computer 262 monitors the phone line 260 for IBS audio tones. When detected, the IBS tones are converted to digital GPS data and displayed to the user on the screen of the computer 262 by the processor 266. The mapping program in computer 262 may then convert the GPS longitude and latitude values to country, city, and street addresses.

Synchronization

Fig. 19 shows an alternative technique for demodulating and synchronizing IBS modems in the IBS decoder 300. At interface 301, IBS audio tones are received over a voice channel of a wireless communication network. The received tones are converted from analog to digital form by a/D converter 302. The IBS signal detector 304 detects the presence of IBS audio tones in the same manner as described in fig. 11.

An alternative synchronization technique begins with the decoder 300 using multipliers 306 and 308 to synchronize the data signalsThe IBS signal is tuned to the composite baseband. Multiplier 306 effectively shifts any IBS tones at the first and second IBS frequencies f1 and f2 to DC. The first baseband signal is called S_A', the second baseband signal being referred to as S_B'. Matched filter bank 310 applies a matched filter to the baseband signal having the desired pulse shape for the two audio tones representing binary "1" and binary "0" values. S output from the matched filter bank 310_AThe signal representing a binary 1 value, S_BThe signal represents a binary 0 value. Taking into account the possibility of S_AOr S_BThe matched filter bank may also add filtering to the known characteristics of the wireless communication channel in the signal.

The matched filter is selected to match the pulse shaping applied to the modulator. Pulse shaping is selected in consideration of the best trade-off between signaling bandwidth, bit rate and intersymbol interference. The pulse shaping filter is applied to a set of stages of a digital oscillator of the modulator.

The IBS synchronizer 312 aligns the modulator with the synchronization pattern attached to the front end of the IBS packet. From S_AAnd S_BA sample segment 316 of the signal and a sample start time T_BAre input together into the synchronous demodulator 314. The demodulator 314 outputs a power value 320 to the IBS synchronizer 312 indicating how synchronized the demodulator is with the start bit in the synchronization mode. The IBS synchronizer 312 uses start times T relative to the individual samples_BThe power value 320 determines the optimal start time (T) for demodulating the remaining bits in the IBS packet_B). The IBS packet modulator 322 then uses the optimal start time T_BDemodulation from S_AAnd S_BThe binary bit value of the signal.

Fig. 20 depicts the sync demodulator 314 and IBS packet demodulator 322 of fig. 19 in more detail. First integrator 324 couples S_AThe first sample segment of the signal is integrated. The integrator starts at a sample start time T_BAnd integrates N samples representing the duration of one IBS bit (baud time). Detector 326 feeds the magnitude of the integrated value to summer 332. In a similar manner, an integrator328 pairs starting at the sample start time T_BSignal S of_BIs integrated. The detector 330 will detect S_BThe magnitude of the integrated segment of the signal is fed to adder 332. The output of the summer 332 is the power signal 320 that is sent back to the synchronizer 312. The IBS packet demodulator 322 (fig. 19) further includes a base station according to S_AAnd S_BThe amplitude of the signal, a comparator 334 that produces either a binary 1 value or a binary 0 value.

For further explanation, fig. 21 shows the signal S output from the matched filter bank 310_AAnd S_BIs shown in the figure. S_AAnd S_BSeveral samples 336 of the signal represent the bit duration of one IBS tone. In the example shown in fig. 21, 5 samples are taken for each bit duration T. For each integration, the sample start time T_BOne sample is moved. The start sample of the first integral starts at a sample start time T_b1. As shown in fig. 21, the sample start time T_b1S, which does not correspond to a value representing a binary "1_ASignals or S representing binary "0" values_BThe signals are consistent. For T_b1The sync demodulator 314 in fig. 20 produces a power output value of 0.0.

When using the sample start time T_B2The demodulator 314 produces an output value of-2.0. Sample start time T_B3Representative and signal S_BThe start of the "0" tone in (1) synchronizes the best samples. At the synchronization start time T_B3The output power is-3. When the sample starts time T_B4And T_B5Further away from the optimal synchronization position, the magnitude of the output power decreases. Fig. 22 shows the magnitude of the power distribution for different sample start times. The maximum power value being at the start time T of the sample_B3. Thus, the IBS synchronizer 312 uses the best sample start time T_b3(FIG. 19).

Referring to fig. 20 and 21, starting at a sample time T_b3The first sampling segment 338 of fig. 20 produces an output value of-3 from the adder 332 of fig. 20. For any adder value less than zero, comparator 334 in FIG. 20 produces a binary "0". To pairAt sample value 340 in the second segment, the output of adder 332 produces an output value of + 3. Since the output value of the second sample segment is greater than 0, comparator 334 generates a binary "1" value. The IBS packet demodulator 322 (fig. 19) continues to decode S for the remaining IBS bit stream_AAnd S_BTones in the signal.

Fig. 23 shows a variation of the synchronization scheme described in fig. 19-22. IBS tones are detected in block 341. For audio pitch frequencies f representing binary bit "1" values_AAnd an audio tone f representing a binary bit "0_BMultiplier 342 shifts the IBS tones to the base tones. To f_AAnd f_BEach single sample t (x) of the signal is baseband shifted.

Instead of summing the samples for the entire baud, a running sum of the latest baud values is obtained in block 344 using the new samples t (x). For example, at a sample rate of 20 samples per bit, the 21 st sample T (N + 1) is removed from the running sum and the next sample T (x) is added to the running sum. Two running sums of magnitudes for pitch a and pitch B are obtained in block 345, respectively, and the two magnitudes are compared by comparator 346. A binary "1" or a binary "0" value is output from the comparator 346 depending on which of the a-tone or B-tone samples has the largest magnitude. In a correlation block 347, the binary bit values output from the comparator 346 are correlated with a known sync pattern. Selected sample start time T_BThe latest sample determined to produce the largest correlation value with the sync pattern. Then starting time T according to the selected sample_BThe remaining bits in the IBS packet are demodulated.

Multi-channel in-band signaling modem

Fig. 24 shows an encoder section 350 of a multi-channel inband signaling (MIBS) modem. Data source 351 produces a binary bit stream. The MIBS encoder 350 generates a plurality of inband signaling channels within the same voice channel. Data buffer 352 holds a stream of binary bits from data source 351. The packet assembler 353 assembles the bits in the buffer 352 into a packet payload and adds a preamble and a postamble to the packet payload, thereby forming an IBS packet as described above in fig. 4.

The encoder 350 includes two modulators 356 and 362 that each produce different audio tones representing bits in an IBS packet. Modulator 356 modulates a binary "1" value with f1 frequency 360 and a binary "0" value with f2 frequency 358. The modulator 362 modulates the other bits of the IBS packet having a binary "1" value with the f3 frequency 364 and a binary "0" value with the f4 frequency 366. The f1 and f2 output from modulator 356 are referred to as a first inband signaling channel and the f3 and f4 tones output from modulator 362 are referred to as a second IBS channel. The tones output from the two modulators 356 and 362 are combined together by a summer 368 and then output to a D/a converter 370 and other cellular phone circuits 14 (fig. 2). The cellular telephone circuit 14 encodes the tones in both IBS channels and transmits the tones over the voice channel of the cellular telephone network.

The individual modulators 356 and 366 are similar in operation to the IBS modulator 64 shown in fig. 4. Any number of IBS channels may be generated in the IBS modulator 24. For example, a third IBS channel may be provided by adding a third IBS modulator that modulates bits of a third portion of the IBS packet to the tones using frequencies f5 and f 6. The output of the third IBS channel may be fed to a summer 368. However, for simplicity, only a two-channel IBS modem with two corresponding IBS modulators 356 and 362 are shown in fig. 24.

The IBS channel controller 354 controls how the transmit and receive IBS modems use the multiple IBS channels. For example, the first IBS channel may only be used by the first IBS modem for IBS packets and the second IBS channel may only be used by the first IBS modem for receiving IBS packets. A second IBS modem on the other end of the transmission transmits using the second IBS channel and receives using the first IBS channel. The IBS channel controller 354 adds control bits to the IBS packets that negotiate the use of multiple IBS channels between the two communicating IBS modems. The different structures of the IBS modem are explained in more detail below in fig. 26 and 27. The controller 354 also controls which portions of the IBS packets are modulated by modulators 356 and 362. For example, the modulators may modulate all other IBS packets or each modulator may modulate a different portion of the same IBS packet.

Fig. 25 shows a decoder 375 of the MIBS modem. Audio tones from the voice channel are decoded by receive circuitry 372 and fed to a/D converter 374. The first filter 376 filters signals that are outside the frequency range of two tones in the first IBS channel and the second filter 378 filters signals that are outside the frequency range of the two tones in the second IBS channel. The frequency range of filter 376 is from f1- Δ f to f2 +/Δ f, and the frequency range of filter 378 is from f3- Δ f to f4 +/Δ f. Filters 376 and 378 are shown between decoders 380 and 382, respectively. However, filters 376 and 378 may be implemented anywhere in the same DSP during decoding.

The first IBS channel decoder 380 detects and demodulates two tones in the first IBS channel into binary bit values, and the second IBS channel decoder 382 detects and demodulates two tones in the second IBS channel into binary bit values. The decoders 380 and 382 detect, synchronize and demodulate IBS tones in the same manner as previously described with respect to the decoder 98 in fig. 19 or the decoder 300 in fig. 19. The packet assembler 386 assembles the bits output from the two decoders 380 and 382 into IBS packets that are then output to the data buffer 388.

An IBS channel controller 384 in the receiving IBS modem synchronizes the two decoders 380 and 382 and determines which decoder demodulates which portions or which IBS packets. The controller 384 also conducts a communication protocol with the transmitting IBS modem that negotiates which IBS modem is transmitting IBS packets over which IBS channels and which IBS modem is receiving IBS packets over which IBS channels.

The filter 376 and decoder 380 for the first IBS channel and the filter 378 and decoder 382 for the second IBS channel may be implemented in software in the same DSP. On the other hand, one DSP may be used for each individual channel encoder and decoder in each MIBS modem.

In a "MIBS" modem, frequencies f1& f2 are preferably far from frequencies f2 and f 3. One of the advantages of MIBS is interference mitigation and the ability to adapt to changes in cellular phone performance across multiple manufacturers by dynamically changing frequencies as performance deteriorates. When a modem is detecting an error, a robust low baud rate control signal can be sent to select the new frequency.

Fig. 26 shows one possible configuration of two multi-channel inband signaling (MIBS) modems 390 and 396. Two IBS channels 398 and 400 are transmitted from the MIBS modem 390 over the voice channel of the wireless communication network and then may reach the MIBS modem 396 over the landline telephone network. The two MIBS modems shown in fig. 26 operate in half duplex mode, wherein one of the two IBS modems simultaneously transmits IBS packets over the first IBS channel 398 and the second IBS channel 400.

After the first IBS modem 390 completes transmission 392 of IBS packets over the two IBS channels, the second IBS modem 396 is allowed to transmit 394 back to the modem 390 over the two IBS channels 398 and 400. The MIBS modem 390 sends information in one of the IBS packets indicating to the MIBS modem 396 that the transmission 392 is complete.

Fig. 27 shows an alternative structure where a first IBS channel 398 is dedicated to transmitting IBS packets from the MIBS modem 390 and a second IBS channel 400 is dedicated to transmitting packets from the MIBS modem 396. Thus, both MIBS modems 390 and 396 may transmit and receive packets simultaneously. This full duplex configuration may provide faster communication for certain types of IBS transmissions.

The MIBS modem 390 may transmit different portions of the same IBS packet over the two IBS channels 398 and 400 or may alternatively transmit different IBS packets over the two IBS channels. As with other architectures, one IBS channel may be used to transmit IBS packets and a second IBS channel may be dedicated to signaling and protocol communications between two MIBS modems. As with other alternative structures, bits from multiple portions of the same IBS packet are interleaved in two IBS channels or the same IBS packet is redundantly transmitted over both IBS channels. The information in the two IBS channels can be reconstructed based on the application associated with the IBS packet data.

The request to reconstruct the IBS channel may be encoded into an IBS packet header. For example, the IBS channel controller 354 (fig. 24) in the MIBS modem 390 may send the IBS packets containing the reconfiguration request in the IBS packet preamble 73 (fig. 5) to the MIBS modem 396. A reconfiguration request from modem 390 may request a first IBS channel 398 and a second IBS channel 400, followed by a request to assign a third IBS channel 401 with a slower baud rate to the MIBS modem 396 for sending an acknowledgement message back to modem 390. The MIBS modem 390 then waits for an acknowledgement of the fabric request from modem 396.

The IBS channel controller 384 (fig. 25) in the MIBS modem 396 reads the reconstruction request in the IBS packet preamble. The controller 384 then outputs an acknowledgement back through the encoder of the MIBS modem 396. The encoder formats the acknowledgement into the preamble of the acknowledging IBS packet, then modulates the acknowledging IBS packet, and sends the acknowledging IBS packet back to the MIBS modem 390 over one or more currently assigned IBS channels. A controller in the modem 396 then reconfigures the encoder to receive IBS packets over the first and second IBS channels 398 and 400 and transmit the packets over the low baud rate third channel 401.

Upon receipt of the acknowledgement from modem 396 at modem 390, the controller instructs the encoder and decoder in modem 390 to transmit over the first and second IBS channels and receive over the low baud rate third channel. The two modems 390 and 396 then transmit and receive IBS packets according to the new channel configuration.

Multi-carrier in-band signaling modem

Figure 28 shows a multi-carrier in-band signaling modem in accordance with another aspect of the present invention. The multi-channel modems depicted in fig. 24-27 produce two different audio tones, one tone representing a binary "1" value and the other tone representing a binary "0" value. The two tones are generated in a time-sequential tone stream to represent a binary bit stream.

The multi-carrier IBS modem in fig. 28 simultaneously generates multiple audio tones, each tone representing a different bit position in a four-bit portion of the IBS packet. The particular audio tone associated with one of the four bit positions represents a binary "1" value (or a binary "0" value). If no audio tones are generated for a particular bit time (baud), the IBS decoder assumes a binary bit value of "0" associated with that bit position.

Referring to fig. 28, a bit stream is input to a data buffer 402 for transmission over a voice channel of a wireless communication network. Packet formatter 404 formats the bits into IBS packets. The first portion of one of the IBS packets contains "1010" bits. Packet formatter 404 inputs each of the four bits into four modulators 406, 408, 410, and 412, respectively. The first bit "1" of the four bits is referred to as bit B1, the second bit "0" is referred to as bit B2, the third bit "1" of the four-bit sequence is referred to as bit B3, and the fourth bit "0" is referred to as bit B4.

Modulator 406 receives bit B1, modulator 408 receives bit B2, modulator 410 receives bit B3, and modulator 412 receives bit B4. Since bit B1 is a binary "value," modulator 406 produces a tone of frequency f1 during the first baud period. Since the B2 bit is a binary "0" value, the modulator 408 does not generate the f2 tone for the first baud period. Accordingly, the modulator 410 generates the f3 tone during the first baud period, and the modulator 412 does not generate the f4 tone during the first baud period. These modulators operate in substantially the same manner as the IBS modulator 64 of fig. 4, except that frequency tones are generated for binary "1" values and no tones are generated for binary "0" values.

f1 and f3 are combined by adder 414. Digital-to-analog converter 416 converts the digital signal to an analog signal that is fed into cellular telephone transmitter circuitry 418. Transmit circuitry 418 communicates the audio tones over a voice channel of a cellular telephone network.

Fig. 29 shows a decoder of a multicarrier IBS modem. The receive circuit 420 receives IBS tones from a voice channel of the cellular communication network. The a/D converter 422 converts the audio tones into digital signals. The four band pass filters 424, 426, 428 and 430 are centered around the frequencies of tones f1, f2, f3 and f4, respectively. The tone representing binary bit B1 passes through bandpass filter 424, while other tones, such as tone f3, are filtered by bandpass filter f 1. The decoder 432 determines the tone f1 in a manner similar to the IBS decoder described in fig. 11-13 for only a single tone. Since the decoder 432 detects the f1 tone, a binary "1" value is generated that represents bit B1 in a four-bit sequence.

Since the decoder 434 does not detect any f2 tones, a binary "0" value is generated that represents bit B2 in the four-bit sequence. The decoder 436 detects the f3 tone, thus producing a binary "1" value representing bit B3. The decoder 438 generates a binary "0" value that represents bit B4 because the multicarrier decoder did not generate any f4 tones. The packet assembler 440 receives the four bits B1-B4 and places them into the appropriate IBS packet locations in the data buffer 442.

Having described and illustrated the principles of the present invention in terms of its preferred embodiments, it should be apparent that the invention can be modified in arrangement and detail without departing from such principles. All such modifications and variations are within the spirit and scope of the following claims.

Claims (9)

1. A multi-channel inband signaling modem for communicating digital data over a voice channel of a telecommunications network, comprising:

an input to receive digital data;

a first modulator for converting first binary bit values in a first portion of the digital data into first audio tones having a first frequency and for converting second binary bit values in the first portion of the digital data into second audio tones having a second frequency;

a second modulator for converting the first binary bit values in the second portion of the digital data into a third audio tone having a third frequency, and for converting the second binary bit values in the second portion of the digital data into a fourth audio tone having a fourth frequency; and

an output for outputting audio tones through a voice channel of a digital wireless telecommunications network.

2. A modem according to claim 1, comprising:

a first decoder for monitoring the first and second audio tones and converting any detected first audio tones back to first binary bit values and any detected second audio tones back to second binary bit values; and

a second decoder for monitoring the third and fourth audio tones and converting any detected third audio tones back to the first binary bit values and any detected fourth audio tones back to the second binary bit values.

3. A modem according to claim 2 including a controller which controls when the first and second modulators generate audio tones and when the first and second decoders monitor the audio tones.

4. The modem of claim 3 wherein the controller conducts a configuration session with another multi-channel in-band signaling modem.

5. A modem as defined in claim 3 wherein the controller controls which bits of the digital data are converted to audio tones by the first and second modulators.

6. A modem according to claim 2, comprising:

a first filter coupled to the first decoder that filters signals outside of the frequency range of the first and second audio tones; and

a second filter coupled to the second decoder that filters signals outside the frequency range of the third and fourth audio tones.

7. The modem of claim 1 wherein the audio tones are fed into the same analog-to-digital converter in the cellular telephone that processes human voice signals.

8. The modem of claim 2 wherein the first and second modulators and the first and second decoders are located in a cellular telephone.

9. A modem according to claim 1 including a digital-to-analog converter for converting audio tones to analog signals, said analog signals being input to the audio input of the telephone, subsequently encoded by audio processing circuitry in the telephone, and transmitted by said audio processing circuitry over the voice channel of the telecommunications network.