Enhanced in-band signaling for data communications over digital wireless telecommunication networks

Description of equivalent WO 0199295 A2

ENHANCED IN-BAND SIGNALING FOR DATA

COMMUNICATIONS OVER DIGITAL WIRELESS TELECOMMUNICATIONS NETWORKS

Technical Field

This invention is related to wireless teleconiiriiinications and more specifically to a system that transmits digital data over the audio channel of a digital wireless network "in-band. "

Background of the Invention A cellular telephone allows a user to talk to another user without being tethered to a "land line. " The cell phone includes circuitry that samples the audio signals from the user's voice. These voice signals are converted into a digital form using an A-D converter. The digitized voice signals are encoded by a voice coder

(vocoder) and then modulated onto a carrier frequency that transmits the voice signals over a cell network. The voice signals are sent over the wireless cellular network either to another phone in the wireless cell network or to another phone in a land-line phone 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 the cellular and landline phone networks. These telephone components can implement different coding schemes for encoding and decoding the voice signals. These telecommunication components are designed to efficiently transmit voice signals over wireless and landline voice communication channels. For example, a digital vocoder uses predictive coding techniques to represent the voice signals. These predictive coders filter out noise (non- voice signals) while compressing and estimating the frequency components of the voice signals before being transmitted over the voice channel.

A problem arises when voice communication equipment, such as the vocoder, are used for transmitting digital data. The vocoders may interpret signals representing digital data as a non- voice signal. The vocoder might completely filter out or corrupt those digital data signals. Therefore, digital data can not be reliably transmitted over the same digital audio channel used for transmitting voice signals.

It is sometimes necessary for a user to transmit both audio signals and digital data to another location at the same time. For example, when a cellular telephone user calls "911" for emergency assistance, the user may need to send digital location data to a call center while at the same time verbally explaining the emergency conditions to a human operator. It would be desirable to transmit this digital data through a cell phone without having to use a separate analog wireless modem.

Accordingly, a need exists for transmitting digital data over a voice channel of a digital wireless communications network. Summary of the Invention

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

The foregoing and other features and advantages of the invention will become more readily apparent from the following detailed description of preferred embodiments of the invention, which proceeds with reference to the accompanying drawings. Brief Description of the Drawings

FIG. 1 is a diagram showing a wireless communications network that provides in-band signaling (IBS) according to the invention.

FIG. 2 a detailed diagram of a cellular telephone coupled to an IBS modem according to one embodiment of the invention.

FIG. 3 is another embodiment of the IBS modem according to the invention.

FIG. 4 is a detailed diagram of an IBS modem encoder.

FIG. 5 is a schematic diagram of a IBS packet.

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

FIG. 7 is a diagram showing how digital data is corrupted by an Automatic Gain Controller.

FIG. 8 is a diagram showing how a digital wireless network can filter out digital data tones. FIG. 9 is a detailed diagram of receiving circuitry coupled to an IBS modem decoder.

FIG. 10 is a state diagram for the IBS decoder shown in FIG. 9.

FIG. 11 is a block diagram showing a search state in the IBS decoder.

FIG. 12 is a block diagram showing an active state in the IBS decoder. FIG. 13 is a block diagram showing a clock recovery state in the IBS decoder.

FIG. 14 is a schematic diagram of a cellular phone with the IBS modem located in a detachable battery pack.

FIG. 15 are schematic diagram showing different data sources coupled to a cellular telephone through a IBS modem. FIG. 16 is a schematic diagram showing implementation of the IBS modem using a sound card.

FIGS. 17 and 18 are block diagrams showing how the sound card in FIG. 16 operates.

FIG. 19 is a block diagram of a synchronization circuit for the IBS modem. FIG. 20 is a detailed diagram of the synchronization circuit in FIG. 19. FIG. 21 is a timing diagram showing how the synchronization circuit in FIG. 19 operates.

FIG. 22 is a graph showing how the synchronization circuit identifies the optimum synchronization start time. FIG. 23 is an alternative implementation of the synchronization circuit.

FIG. 24 is an encoder diagram for a multichannel IBS modem.

FIG. 25 is an decoder diagram for a multichannel IBS modem.

FIGS. 26 and 27 show different channel configurations for the multichannel IBS modem shown in FIGS. 24 and 25. FIG. 28 is an encoder diagram for a multicarrier IBS modem.

FIG. 29 is an decoder diagram for a multicarrier IBS modem.

Detailed Description of Preferred Embodiment

Referring to FIG. 1, a wireless communications network 12 includes a cell phone 14 that receives voice signals 22 from a user 23. A voice coder (vocoder) 18 in the cell phone 14 encodes the voice signals 22 into encoded digital voice signals 31 that are then transmitted over a wireless digital radio channel 34 (cell call). The cell phone 14 transmits the encoded voice signals 31 to a cellular communications site (cell site) 36 that relays the cell call to a Cellular Telecommunications Switching System (CTSS) 38. The CTSS 38 either connects the cell call to another cell phone either in the wireless cellular network 12, to a landline phone on a PSTN network 42 as a circuit switched call or routes the cell call over a packet switched Internet Protocol (IP) network 46 as a Voice Over IP (VoIP) call. The cell call can also be routed from the PSTN network 42 back to the cellular network 12 or from the PSTN network 42 to the IP network 46, or visa versa. The cell call eventually reaches a telephone 44 that corresponds with a destination phone number originally entered at the cell phone 14.

Additional data could be inserted at any point in the cellular network 12, such as in PSTN network 42 and IP network 46 and the signal remodulated for transmission over wireline or cellular networks. Such data could be system related such as routing information, toll or tariff information, etc. An In-Band Signaling (IBS) modem 28 enables cell phone 14 to transmit digital data 29 from a data source 30 over the 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 encoding components in the cellular network 12 and landline network 42, such as vocoder 18, from corrupting the digital data. The encoding and modulation scheme used in the IBS modem 28 allows digital data 29 to be transmitted through the same voice coder 18 used in the cell phone 14 for encoding voice signals 22. Any appliance such as a vending machine, etc could be enhanced by this technology. Synthesized tones are defined as signals that represent digital data that also have signaling characteristics that enable the signals to be encoded and decoded by a voice codec without losing the digital data information in the signal. In one example, Frequency Shift Keying (FSK) signals are used to created the synthesized tones at different frequencies within the audio range of human speech. The IBS modem 28 enables voice signals 22 and digital data 29 to be transmitted over the same digital audio channel using the same cell phone circuitry. This prevents a user from having to transmit digital data using a separate wireless modem and enables a cell phone user to talk and send data during the same digital wireless call. The invention modulates the digital data 29 into synthesized audio tones. This prevents the cell phone vocoder 18 from filtering or corrupting the binary values associated with the digital data 29. The same cell phone transceiver and encoding circuitry is used for transmitting and receiving both voice signals and digital data. This enables the IBS modem 28 to be much smaller, less complex and more energy efficient than a standalone wireless modem. In some embodiments, the IBS modem

28 is implemented entirely in software using only the existing hardware components in the cell phone 14.

One or more servers 40 are located at any of various locations in the 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 radio channel 34. Decoded digital data is either processed at the server 40 or routed to another computer, such as computer 50.

Referring to FIG. 2, a first transmitting portion of the IBS modem 28 includes an IBS encoder 52 and a Digital to Analog converter (D/A) 54. The IBS encoder 52 is typically implemented using a Digital Signal Processor (DSP). The data source 30 represents any device that requires wireless transmission or reception of digital data.

For example, the data source 30 can be a laptop computer, a palm computer or a

Global Positioning System (GPS) (see FIG. 15).

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

The IBS modem 28 outputs binary values that each represent an amplitude and phase component of an audio tone. The D/A converter 54 converts these digital values into analog audio tones 26 that are then output to an auxiliary audio port 15 on the cell phone 14. The analog audio tones 26 are then processed by the cell phone

14. An Analog to Digital (A/D) converter 16 in the cell phone 14 encodes the synthesized analog audio tones 26 into digital values. The vocoder 18 encodes the digital representations of the synthesized tones 26 into encoded digital data 32 and outputs the encoded data to a transceiver 19 that transmits the encoded digital data 32 over the radio channel 34.

The preferred voltage of the synthesized audio tones 26 output from the D/A converter 26 is around 25 millivolts peak to peak. This voltage level was discovered to prevent the audio tones 26 from saturating the voice channel circuitry in cell phone

14.

Because the digital data 29 is fed through the existing auxiliary hands free audio port 15 in 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. The data source 30 can transmit digital data 29 in any digital format. For example, the digital data 29 can be sent over an RS-232 interface, 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 inside the cell phone 14 and is implemented in software using the existing cell phone processor or using some combination of its own components and the existing cell phone components. In this embodiment, the cell phone 14 may include a data port 56 that receives the digital data 29 from the external data source 30. In an alternative embodiment, the digital data source 30 is internal to the cell phone 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 as mentioned above is typically implemented in software using a DSP and may use the same DSP used for implementing the vocoder 18. The D/A converter 54 outputs the synthesized audio tones representing digital data 29 to the internal A/D converter 16 of the cell 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 outputs the quantized data 55 directly into the vocoder 18. In still another embodiment of the invention, the IBS encoder 52 is implemented entirely in software in the same DSP that implements the vocoder 18

The vocoder 18 uses a specific encoding scheme associated with the wireless communications network 12 (FIG. 1). For example, the vocoder 18 could be a NCELP encoder that converts voice signals into digital CDMA signals. The A/D converter 16, D/A converter 54 and transceiver 19 are existing cell phone components known to those skilled in the art.

It is important to note that the IBS encoder 52 enables the digital data 29 to be transmitted using the same cell phone circuitry that transmits voice signals. The IBS encoder 52 prevents any signal approximation, quantization, encoding, modulation, etc. performed by the, A/D converter 16, vocoder 18, or transceiver 19 from corrupting or filtering any bits from the digital data 29. FIG. 4 is a detailed diagram of the IBS encoder 52 shown in FIG. 2 and FIG. 3. A data buffer 58 stores the binary bit stream 29 from the data source 30. A packetizer 60 segments the bits in buffer 58 into bytes that comprise a IBS packet pay load. A packet formatter 62 adds a packet preamble and postamble that helps prevent corruption of the IBS packet pay load. An IBS modulator 64 then modulates the bits in the IBS packet with two or more different frequencies 66 and 68 to generate digital data tones 69.

Preventing Corruption of Digital Data in Noice Channels Cell phone voice coders increase bandwidth in voice channels by using predictive coding techniques that attempt to describe voice signals without having to send all the information associated with human speech. If any unnatural frequencies or tones are generated in the voice channel (i.e., frequencies representing digital data), those frequencies might be thrown out by the voice coder 18 (FIG. 2). For example, if the amplitude of the digital data tones are greater than that of normal voice signals or the same digital data tone is generated for too long a time period, the voice coder 18 might filter out that high amplitude or extended frequency signal. Depending on how the digital data tones are encoded, the digital bits represented by those unnatural audio tones may be partially or entirely removed from the voice channel. The IBS encoder 52 encodes the digital data 29 in a manner where voice coders will not filter or corrupt the tones representing digital data. The IBS encoder 52 does this by controlling the amplitudes, time periods and patterns 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 a header 75 to the front of a IBS packet 70. The packet preamble 73 includes a preamble pattern 72 and a sync pattern 74. A checksum 78 and a packet postamble 79 are attached to the backend of the IBS packet 70.

FIG. 6 shows the 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. A first tone is generated at an f 1 frequency and represents a binary " 1 " value and a second tone is generated at a f2 frequency and represents a binary "0" value. In one embodiment the fl frequency is 600 Hertz and the f2 frequency is 500 Hertz (Hz).

It has been determined that the most effective frequency range for generating the tones that represent the binary bit values are somewhere between 400 Hertz and 1600 Hertz. The IBS modulator 64 includes Sine and Cosine tables that are used to generate the digital values that represent the different amplitude and phase values for the fl and f2 frequencies.

In one embodiment of the invention, the digital data is output on the radio channel 34 at a baud rate of 100 bits/second. This baud rate has been found to be effective in preventing corruption of the digital audio data by a wide variety of different cellular telephone voice coders. The sine waves for each fl and f2 tone begin and end at a zero amplitude point and continue for a duration of 10 milliseconds. Eighty samples are generated for each digital data tone. Referring to FIG. 7, an Automatic Gain Controller (AGC) 80 is one encoding function used in the cell phone 14. The AGC 80 may be software that is located in the same DSP that implements the voice coder 18. The AGC 80 scales instantaneous energy changes in voice signals. There are situations when no voice signals have been fed into the AGC 80 for a period of time followed by a series of audio tones 82. that comprise the beginning of a IBS packet 70. The AGC 80 scales the first group of tones 82 at the beginning of the IBS packet 70. The AGC 80 also looks ahead at the zero signal levels 84 after the end of the IBS packet 70, and will scale the tones 83 at the end of the IBS packet 70 as part of its prediction scaling scheme. This scaling prevents the over amplification of signal or noise when instantaneous 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. fl and f2, respectively. If these tones are scaled by the AGC 80, the digital bits represented by those frequencies might be dropped during encoding. For example, the vocoder 18 may see the scaled tones as noise and filter them from the audio channel. To prevent the unintentional filtering of tones that represent digital data, the IBS packet 70 in FIG. 5 includes preamble bits 72 and postamble bits 79. The preamble bits 72 and postamble bits 79 do not contain any of the digital data bits 29 from the data source but include a certain number of sacrificial bit(s) that are not needed for detecting or encoding the IBS packet 70. The tones that are generated for these sacrificial bits in the preamble and postamble can be scaled or filtered by the AGC 80 without effecting any of the digital data contained in the IBS packet pay load 76.

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

Referring to FIG. 8, adaptive filters adapt around the frequencies that are currently being transmitted over the wireless network. For example, if a long period of the same f 1 tone is currently being transmitted, an adaptive filter used in the cell phone may adapt around that fl frequency spectrum as shown by filter 86.

Another short tone at another frequency f2 may immediately follow the long period of fl tones. If the filter 86 is too slow to adapt, the first few f2 tones may be filtered from the voice channel. If the filtered f2 tone represent bits in the IBS bit stream, those bits are lost.

To prevent adaptive filters in the cell phone from dropping bits, some portion of the preamble 73 includes a random or alternating "1" - "0" bit pattern. This preconditions the adaptive filter as shown by filter 88. The preamble 73 (FIG. 5) tries to include a portion of the same bit sequence that is likely or does occur in the packet pay load 76. For example, the IBS encoder 52 can look ahead at the bit pattern in the pay load 76. The encoder 52 can then place a subset of bits in a portion of the preamble to represent the sequence of bits in the packet pay load.

This preconditions the adaptive filter for the same fl and f2 frequencies, in the same duration and in a similar sequence that is likely to follow in the IBS packet pay load 76. Thus, the adaptive filter is less likely to filter out the tones that actually represent the digital data that is being transmitted.

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

As described above in FIGS. 2 and 3, the decoder part of the IBS modem 28 can be external to the cell phone 14 or can be inside the cell phone 14. Dashed line

104 shows an IBS modem 28 external to a cell phone and dashed line 106 shows an internal IBS modem 28 internal to a cell phone. IBS modems 14 can be located at any telephone location in the PSTN network 42 or IP network 46 (FIG. 1). The receiving circuitry 91 may be different when the IBS modem 28 is coupled to a landline. However, the IBS modem 28 operates under the same principle by transmitting and receiving synthesized tones over the voice channel of the phone line.

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

If the IBS modem 28 is external to the receiving circuitry 91, then a A/D converter 96 converts the analog signals back into digital signals. The IBS decoder 98 demodulates any tones representing digital data back into a digital IBS packets. A packet disassembler 100 disassembles the packet payload from the IBS packets 70 and stores the decoded digital data in a data buffer 102.

FIG. 10 is a state diagram explaining how the IBS decoder 98 in FIG. 9 operates. The IBS decoder 98 repeatedly samples and decodes the audio signals received from the radio channel 34. State 110 searches for tones in the audio signal that represent digital data. If the Signal to Noise Ratio (SNR), for tones within the frequency range of the digital data tones, are greater than a preselected value, the IBS decoder 98 goes into an active state 112. The active state 112 collects tone samples. If at any time during the active state 112 the SNR falls below an active threshold value or a timeout is reached before enough tone samples are collected, the IBS decoder 98 returns to the search state 110 and begins again to search for digital data tones.

After a number of samples are collected, the IBS decoder 98 looks for bits that identify the preamble 73 in the IBS packet 70 (FIG. 5). If the preamble 73 is detected, the IBS decoder 98 moves to clock recovery state 114. The clock recovery state 114 synchronizes with the synchronization pattern 74 in the IBS packet 70 (FIG.

5). The IBS decoder 98 then demodulates the packet payload 76 in state 116. If the preamble 73 is not found, IBS decoder 98 goes back to the search state 110 and starts searching again for the begimiing of an IBS packet 70.

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

FIG. 11 is a detailed diagram for the search state 110 of the IBS decoder 98. The search state 110 uses in band and out of band filtering. "In band" is used in the following discussion to refer to tones within the frequency range of the two tones that represent the digital data binary "1" value (500 Hz) and the digital data binary "0" value (600 Hz).

A first band pass filter 118 (in band) measures energy for signals in the audio channel within the frequency range of about 400 Hz to around 700 Hz. A second band pass filter 120 (out of band) measures the energy in the audio channel for signals outside of the 400Hz - 700 Hz range. A Signal to Noise Ratio (SNR) is calculated in block 122 between the in band energy and the out of band energy. If tones representing the digital data exist in the audio channel, the energy measured by the in band filter 118 will be much greater then the energy measured by the out of band filter 120. If the SNR is below a selected threshold in comparator box 124, signals in the audio channel are determined to be actual voice signals or noise. If the SNR is above the threshold, the IBS decoder 98 deterπiines the tones represent in band digital data. When digital data is detected, the IBS decoder 98 moves into the active state 112 (FIG. 10) to begin searching for the beginning of an IBS packet 70.

FIG. 12 shows the active state 112 for the IBS decoder 98. Block 130 is notified by the search state 110 when an in band tone is detected in the audio channel. Samples of the audio tones 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, padded with zeros, and then correlated with Discrete Fourier Transforms (DFTs).

A first DFT has coefficients representing a 500 Hz tone and is applied to the windowed data in block 134. The first DFT generates a high correlation value if the samples contain a 500 Hz tone ("0" binary bit value). A second DFT represents a 600 Hz tone and is applied to the windowed samples in block 136. The second DFT generates a high correlation value if the windowed samples contain a 600 Hz tone ("1" binary bit value). Block 138 selects either a binary "0" or binary "1" bit value for the windowed data depending on which of the 500 Hz DFT or 600 Hz DFT yields the largest correlation value. The IBS decoder 98 in decision block 140 continues to demodulate the tones until the preamble of the IBS packet 70 has been detected. The IBS decoder 98 then moves to clock recovery state 114 (FIG. 13) to synchronize with the sync pattern 74 in the IBS packet 70 (FIG. 5). If more bits need to be demodulated before the preamble 73 can be verified, decision block 140 returns to block 132 and the next 80 samples of the digital data tones are windowed and demodulated.

FIG. 13 describes the clock recovery state 114 for the IBS decoder 98. After the preamble 73 in the IBS packet 70 is detected in the active state 112, the clock recovery state 114 demodulates the next string of bits associated with the sync pattern 74 (FIG. 5). The clock recovery state 114 aligns the tone samples with the center of the correlation filters described in the active state 112. This improves decoder accuracy when demodulating the IBS packet payload 76.

Decision block 142 looks for the sync pattern 74 in the IBS packet 70. If after demodulating the next tone, the sync pattern 74 is not found, decision block 142 offsets the window used for sampling the sync pattern 74 by one sample in block 148. Decision block 150 then rechecks for the sync pattern 74. If the sync pattern 74 is found, decision block 144 deterrnines the power ratio for the detected sync pattern. This power ratio represents a confidence factor of how well the demodulator is synchronized with the sync pattern. The power ratio is compared with the power ratios derived for different window shifted sampling positions. If the power ratio is greater then a previous sampling position, then that power ratio is saved as the new maximum power ratio in block 146.

If the power ratio for the sync pattern 74 is less then the previously measured power ratio, the decoder in block 148 offsets the sampling window by one sample position. The power ratio is then determined for the shifted window and then compared to the current maximum power ratio in decision block 144. The window is shifted until the maximum power ratio is found for the sync pattern 74. The window offset value at the maximum power ratio is used to align the demodulator correlation filters with the center sample of the first bit 77 (FIG. 5) in the IBS packet header 75. The IBS decoder 89 then jumps to demodulate state 116 (FIG. 10) where the identified window offset is used to demodulate the remaining 500 and 600 Hz tones that represent the packet payload bits 76 and check sum bits 78. The demodulation state 116 correlates the fl and f2 tones with DFTs in the same manner as in the active state (FIG. 12). The check sum bits 78 are then used as a final check to verify that a valid IBS packet has been received and accurately decoded.

FIG. 14 is a diagram of the IBS modem 28 located in a battery pack connected to the cellular telephone 14. A hands free audio channel pin 200 couples the IBS modem 28 to the voice channel 202 in the cell phone 14. A switch 204 couples either voice signals from the microphone 17 or digital data tones from the IBS modem 28 to the voice channel 202. The switch 204 is controlled either through a menu on a screen (not shown) in the cell phone 14 or by a button 206 that extends out of the back end of the battery pack 208. The switch 204 can also be controlled by one of the keys on the keyboard of the cell phone 14. The button 206 can also be used to initiate other functions provided through 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 a GPS satellite 212. A cell phone operator simply pushes button 206 during an emergency situation. Pressing the button 206 automatically enables the GPS receiver 210 to collect GPS data from GPS satellite 212. At the same time, the switch 204 connects IBS modem 28 on the voice channel 202 of the cell phone 14. The IBS modem 28 is then activated. As soon as the GPS data is collected in the IBS modem 28, the data is formatted, encoded and output by IBS modem 28 to the voice channel 202 of the cell phone 14. The user 23 can push the button 206 anytime after manually calling up a phone number. After the audio channel is established with another endpoint, the user 23 pushes button 206. 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 as digital data tones through the IBS modem 28 to an endpoint over the established audio channel. After the data has been successfully transmitted, the user presses button 206 again reconnecting switch 204 to the audio receiver 17.

FIG. 15 shows the different types of data sources that can be connected to the IBS modem 28. Any one of a palm computer 212, GPS receiver 214 or a computer 216, etc. can are coupled to the IBS modem 28. The IBS modem 28 converts the bits output from the device into digital data tones that are then output over the radio channel 34 in the wireless network. Because data can transmitted to another endpoint through the cell phone 14, none of the devices 212, 214 or 216 need a separate wireless modem.

Implementation of Inband Signaling Modem in a Sound Card The IBS modem can be implemented in a standard computer sound card. Referring to FIG. 16, a sound card 252, such as a Sound Blaster card manufactured by Creative Labs, Inc., 1523 Cimarron Plaza; Stillwater, Ok 74075 is included in a computer 250. A speaker output 253 of the sound card 252 outputs audio tones to a hands free port 257 on a cell phone 258. A microphone input 259 on the sound card 252 is connected to the speaker output of the cell phone 258.

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

Referring to FIGS. 16 and 17, data files, GPS data, data entered by the keyboard by a user, or any other digital data is packetized and formatted by computer 250 into IBS packets in block 270. Packetization and packet formatting is described in FIGS. 4 and 5. The binary bit values in the IBS packets are converted in block 272 into a digital format used by the sound card 252 (FIG. 16) for generating synthesized audio tones. For example, binary "1" bit values in the IBS packet are converted into a digital format representing a first fl frequency tone and binary "0" bit values are converted into a second f2 frequency tone. The fl and f2 tones are generated similar to the manner described in FIG. 6.

The sound card in block 274 outputs analog tones representing the binary bit values in a manner similar to the IBS encoder 52 and the digital to analog converter 54 described in FIG. 3. The cell phone in block 276 encodes the audio tones and transmits the encoded audio tones over the voice channel in the wireless communications network in block 278. Referring to FIGS. 16 and 18, the cellular phone call is established with a destination phone number. In block 280, either a user picks up the ringing phone line or the computer 262 (FIG. 16) at the destination end of the cellular phone call is programmed to detect a ringing signal from the telephone line 260. If a ring signal is detected, either a user or the computer 262 in block 282 generates an "hook-off" signal on the telephone line 260. The sound card 264 in block 284 acts like an analog to digital converter by converting the audio tones on the telephone line 260 into digital data. The sound card 264 in conjunction with the processor 266 (FIG. 16) decodes the IBS audio tones similar to the IBS decoder 98 described in FIGS. 9-13. The digital representations of detected IBS tones are then displayed on the screen of computer 262 in block 290.

In one example, a user wants to find the location for cell phone 258. The user directs computer 262 (FIG. 16) to dial the phone number for cell phone 258. The computer 262 uses the sound card 264 to send IBS tones that direct cell phone 258 to send back GPS location data. The computer 250 may have a GPS receiver or the cell phone 258 may have a standalone GPS receiver. If the GPS receiver and the IBS modem are internal to the cell phone 258 as shown in FIGS 2-9, the computer 250 does not need to be connected to the cell phone 258.

The GPS data is converted into IBS tones either by the sound card 252 as described in FIG. 17 or through an internal IBS modem as described in FIGS. 2-9. The IBS tones representing the GPS data are transmitted back over the wireless telecommunications channel and the PSTN network 263 to the telephone line 260. The sound card 264 in computer 262 monitors the phone line 260 for the IBS audio tones. When detected, the IBS tones are converted back into digital GPS data and displayed by processor 266 to the user on the screen of computer 262. A mapping process in the computer 262 may then convert the GPS longitude and latitude values into a state, city and street address.

Synchronization FIG. 19 shows an alternative technique for demodulating and synchronizing the IBS modem in the IBS decoder 300. The IBS audio tones are received over the voice channel of the wireless communications network at interface 301. The received tones are converted from analog to digital form by A/D converter 302. The IBS signal detector 304 detects the presence of the IBS audio tones in the same manner as described in FIG. 11. The alternative synchronization technique begins with the decoder 300 tiining the IBS signals to complex basebands with multipliers 306 and 308. Multiplier 306 effectively moves any IBS tones at the first and second IBS frequencies f 1 and f2 to DC. This first baseband signal is referred to as S_A' and the second baseband signal is referred to as S_B' . A matched filter bank 310 applies matched filters to the baseband signals having the expected pulse shapes for the two audio tones representing the binary " 1" and binary "0" values. The S_A signal output from the matched filter bank 310 represents a binary 1 value and the S_B signal represents a binary 0 value. The matched filter bank can also add filtering to account for known characteristics of the wireless communications channel that may exist in the S_Aor S_B signals. The matched filter is selected to match the pulse shaping applied to the modulator. The pulse shaping is selected for the best trade-off between signaling bandwidth, bit rate and inter symbol interference. The pulse shaping filter is applied to the integrated phase of the modulator's numerical oscillator.

An IBS synchronizer 312 aligns the modulator with the synchronization pattern attached to the front of the IBS packet. Segments 316 of samples from the S_A and S_B signals are input to synchronization demodulator 314 along with a sample start time T_B. The demodulator 314 outputs a power value 320 to the IBS synchronizer 312 that indicates how closely the demodulator is synchronized with the beginning bit in the synchronization pattern. The IBS synchronizer 312 uses the power values 320 for each sample start time T_B to determine the optimum synchronization start time (*T_B) for demodulating the remaining bits in the IBS packet. IBS packet modulator 322 then uses the optimum start time *T_B to demodulate the binary bit values from the S_A and S_B signals.

FIG. 20 is a more detailed description of the sync demodulator 314 and the IBS packet demodulator 322 in FIG. 19. A first integrator 324 integrates the first segment of samples for the S_A signal. The integrator starts at sample start time T_B and integrates N number of samples representing the duration T of one IBS bit (Baud time). A rectifier 326 feeds the magnitude of the integration value into an adder 332. In a similar matter, an integrator 328 integrates the segments of samples for signal S_B starting at sample start time T_B. A rectifier 330 feeds the magnitude of the integrated segment of the S_B signal into adder 332. The output of adder 332 is a power signal 320 that is fed back to the synchronizer 312. The IBS packet demodulator 322 (FIG. 19) also includes a comparator 334 that generates either a binary 1 value or a binary 0 value according to the magnitudes of the S_A nd S_B signals. To explain further, FIG. 21 shows a representation of the signals S_A and S_B that are output from the matched filter bank 310. A number of samples 336 of the S_A or S_B signal represent the bit duration T of one IBS tone. In the example shown in FIG. 21, five samples are taken for each bit duration T. The sample start time T_B is shifted one sample for each integration. A starting sample for the first integration starts at sample start time T_bl. As seen in FIG. 21, the sample start time T_M is not aligned with the S_A signal representing a binary "1" value or the S_B signal representing a binary "0" value. The sync demodulator 314 in FIG. 20 generates a power output value of 0.0 for T_bl

When sample start time T_B2 is used, the demodulator 314 generates an output value of -2.0. The sample start time T_B3 represents the sample with the best synchronization with the beginning of the "0" tone in signal S_B. At synchronization start time T_B3 the output power is -3. As the sample start times T_B4 and T_B5 move further away from the best synchronization position, the magnitude of the output power decreases. FIG. 22 shows the magnitude of the power distribution for the different sample start times. The maximum power magnitude is identified at sample start time T_B3. Thus, the optimal sample start time T_b3 is used by the IBS synchronizer 312 (FIG. 19).

Referring to FIGS. 20 and 21, a first sampling segment 338 starting at sample time T_b3 generates an output value from adder 332 in FIG. 20 of -3. The comparator 334 in FIG. 20 generates a binary "0" value for any adder value less than zero. The output of adder 332 for a second segment of sample values 340 generates an output value of +3. Because the output value for 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 the tones in the S_A and S_B signals for the remainder of the IBS bit stream.

FIG. 23 shows a variation of the synchronization scheme described in FIGS. 19-22. The IBS tones are detected in block 341. The IBS tones are shifted to baseband by the multipliers 342 for both the audio tone frequency f_A representing a binary bit "1" value and for the audio tone f_B representing a binary bit "0". The baseband shift is done for each individual sample T(x) of the f_A and f_B signals

Instead of summing an entire baud of samples, a running sum of the latest baud value is taken using the new sample T(x) in block 344. For example, with a sample rate of 20 samples per bit, the 21^st sample T(N+ 1) is deleted from the running sum and the next sample T(x) is added to the running sum. The magnitude of the two running sums for tone A and tone B are each taken in blocks 345 and compared by comparator 346. A binary "1" or binary "0" value is output from comparator 346 depending upon which of the A tone or B tone samples has the largest magnitude value. The binary bit values output from comparator 346 are correlated with the known sync pattern in the correlation block 347. The selected sample start time *T_B is identified as the last sample that generates the largest correlation value with the synchronization pattern. The remaining bits in the IBS packet are then demodulated according to the selected sample start time * T_B.

Multichannel Inband Signaling Modem FIG. 24 shows the encoder portion 350 of a Multichannel Inband Signaling (MIBS) modem. A data source 351 generates a binary bitstream. The MIBS encoder

350 generates multiple inband signaling channels within the same voice channel. A data buffer 352 stores the binary bit stream from the data source 351. A packet assembler 353 assembles the bits in buffer 352 into a packet payload and adds a preamble and postamble to the packet payload to form IBS packets as described above in FIG. 4. The encoder 350 includes two modulators 356 and 362 that each generate different audio tones that represent the bits in the IBS packets. Modulator 356 modulates binary "1" values using an fl frequency 360 and modulates binary "0" values using an f2 frequency 358. Modulator 362 modulates other bits in the IBS packets having binary "1" values using an f3 frequency 364 and modulates binary "0" values using an f4 frequency 366. The fl and f2 tones 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 an adder 368 and then output to the D/A converter 370 and other cell phone circuitry 14 (FIG. 2). The cell phone circuitry 14 encodes and transmits the tones in the two IBS channels over an audio channel of the cellular telephone network.

Each of the individual modulators 356 and 366 are similar in operation to the IBS modulator 64 shown in FIG. 4. Any number of IBS channels can be generated in the IBS modem 24. For example, a third IBS channel could be provided by adding a third IBS modulator that modulates bits for a third portion of the IBS packets into tones using frequencies f5 and f6. The output of the third IBS modulator would be fed into the adder 368. However, for simplicity, only a two channel IBS modem with two corresponding IBS modulators 356 and 362 are shown in FIG. 24. An IBS channel controller 354 controls how the multiple IBS channels are utilized by the transmitting and receiving IBS modems. For example, a first IBS channel may only be used by a first IBS modem for transmitting IBS packets and a second IBS channel may only be used by that first IBS modem for receiving IBS packets. A second IBS modem on the opposite end of the transmission then uses the second IBS channel for transmission and uses the first IBS channel for reception. The

IBS channel controller 354 adds control bits into the IBS packets that negotiate use of the multiple IBS channels between the two communicating IBS modems. The different configurations for the IBS modems are described in further detail below in FIGS. 26 and 27. The controller 354 also controls what portions of the IBS packets are modulated by modulators 356 and 362. For example, the modulators may modulate every other IBS packet or each modulator may modulate different portions of the same IBS packets.

FIG. 25 shows the decoder 375 of the MIBS modem. The audio tones from the audio channel are decoded by receiving circuitry 372 and fed into an A/D converter 374. A first filter 376 filters signals outside a frequency range of the two tones in the first IBS channel and a second filter 378 filters signals outside the frequency ranges of the two tones in the second IBS channel. The frequency range of filter 376 is from fl-Δf to f2+Δf and the frequency range of filter 378 is from f3-Δf to f4+Δf. The filters 376 and 378 are shown before the decoders 380 and 382, respectfully. However, the filters 376 and 378 can be implemented in the same DSP anywhere in the decoding process.

A first IBS channel decoder 380 detects and demodulates the two tones in the first IBS channel into binary bit values and a second IBS channel decoder 382 detects and demodulates the two tones in the second IBS channel into binary bit values. The decoders 380 and 382 detect, synchronize, and demodulate the IBS tones in the same manner as previously described for decoder 98 in FIG. 9 or decoder 300 in FIG. 19.

A packet assembler 386 assembles the bits output from the two decoders 380 and 382 into IBS packets that are then output to a data buffer 388.

The IBS channel controller 384 in the receiving IBS modem synchronizes the two decoders 380 and 382 and determines which decoders demodulate what 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 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 can be implemented in software in the same

DSP. Alternatively, one DSP can be used for each individual channel encoder and decoder in each MIBS modem.

It is preferred in the "MIBS" modem for frequencies fl & f2 to be far apart from frequencies f2 and f3. One advantage of MIBS is interference mitigation and the ability to adapt to variations in cell phone performance across manufacturers by dynamically changing frequencies when performance is bad. A robust low baud rate control signal can be sent to choose a new frequency when one modem is detecting errors.

FIG. 26 shows one possible configuration for two Multichannel Inband Signaling (MIBS) modems 390 and 396. The two IBS channels 398 and 400 are transmitted from MIBS modem 390 over the voice channel of a wireless communications network and then possibly through a landline telephone network to the MIBS modem 396. The two MIBS modems shown in FIG. 26 operate in a half duplex mode where one of the IBS modems transmits IBS packets over both the first IBS channel 398 and the second IBS channel 400 at the same time.

After the first IBS modem 390 has completed a transmission 392 of IBS packets over the two IBS channels, the second IBS modem 396 is allowed to begin a transmission 394 back to modem 390 over the two IBS channels 398 arid 400. The MIBS modem 390 sends information in one of the IBS packets indicating to the MIBS modem 396 that the transmission 392 is completed.

FIG. 27 shows an alternative configuration where the first IBS channel 398 is dedicated to transmitting IBS packets from MIBS modem 390 and the second IBS channel 400 is dedicated to transmitting packets from MIBS modem 396. Thus, both MIBS modem 390 and 396 can transmit and receive packets at the same time. This full duplex configuration can provide faster communications for certain types of IBS transmissions.

The MIBS modem 390 may transmit different potions of the same IBS packets over the two IBS channels 398 and 400 or may alternate transmission of different IBS packets over the two IBS channels. In other configurations, one IBS channel may be used for transmitting IBS packets and the second IBS channel may be used exclusively for signaling and protocol communications between the two MIBS modems. In other alternative configurations, portions of bits from the same IBS packets are interleaved in the two IBS channels or the same IBS packets are transmitted over both IBS channels for redundancy. The information in the two IBS channels can be reconfigured according to the application associated with IBS packet data. A request to reconfigure the IBS channels can be encoded into the IBS packet header. For example, the IBS channel controller 354 (FIG. 24) in MIBS modem 390 may send an IBS packet to MIBS modem 396 that contains a reconfiguration request in the IBS packet preamble 73 (FIG. 5). The reconfiguration request from modem 390 may request both the first IBS channel 398 and the second IBS channel 400 and then request allocation of a third IBS channel 401, with a slower baud rate, to MIBS modem 396 for transmitting acknowledge messages back to modem 390. MIBS modem 390 then waits for an acknowledge of the configuration request from modem 396. The IBS channel controller 384 (FIG. 25) in MIBS modem 396 reads the reconfiguration request in the IBS packet preamble. The controller 384 then outputs an acknowledge back through the encoder of MIBS modem 396. The encoder formats the acknowledge into the preamble of a reply IBS packet that is then modulated and transmitted back to MIBS modem 390 over one or more of the currently allocated IBS channels. The 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 packets over the third low baud rate channel 401.

When the acknowledge from modem 396 is received at modem 390, the controller directs th encoder and the decoder in the modem 390 to transmit over the first and second IBS channels and receive from the low baud rate third channel. The two modems 390 and 396 then transmits and receive IBS packets according to the new channel configuration.

Multicarrier Inband Signaling Modem FIG. 28 shows a Multicarrier Inband Signaling modem according to another aspect of the invention. The multichannel IBS modem described in FIGS. 24-27 generates two different audio tones, one tone representing a binary "1" value and a second tone representing a binary "0" value. The two tones are generated in a sequential tone stream over time to represent a binary bit stream.

The multicarrier IBS modem in FIG. 28 generates multiple audio tones at the same time, where each tone represents a different bit location in a four bit portion of the IBS packet. The particular audio tone associated with one of the four bit locations represents a binary "1" value (or alternatively a binary "0" value). If the audio tone is not generated for a particular bit time (baud), the IBS decoder assumes the binary bit value associated with that bit location is "0". Referring to FIG. 28, a stream of bits are input to data buffer 402 for transmission over the audio channel of a wireless communications network. A packet formatter 404 formats those bits into an IBS packet. A first portion of one of the IBS packets contains the bits " 1010" . The packet formatter 404 outputs each one of the four bits into a different one of the four modulators 406, 408, 410 and 412. The first bit "1" of the four bit sequence is referred to as bit Bl, 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 Bl, modulator 408 receives bit B2, modulator 410 receives bit B3, and modulator 412 receives bit B4. Because bit Bl is a binary "1" value, modulator 406 generates a tone at frequency fl during the first baud period. The modulator 408 does not generate an f2 tone for the first baud period because the B2 bit is a binary "0" value. Accordingly, modulator 410 generates a f3 tone during the first baud period and modulator 412 does not generate a f4 tone during the first baud period. The modulators work in essentially the same manner as the IBS modulator 64 in FIG. 4 except that a frequency tone is generated for the binary "1" values and no tone is generated for the binary "0" value.

The fl and f3 tones are combined together by summer 414. A digital to analog converter 416 converts the digital signal into an analog signal that is fed into cell phone transmit circuitry 418. The transmit circuitry 418 transmits the audio tones over the voice channel of the cellular telephone network.

FIG. 29 shows the decoder for the multicarrier IBS modem. Receive circuitry 420 receives the IBS tones from the voice channel of the cellular communications network. An A/D converter 422 converts the audio tones into a digital signal. Four bandpass filters 424, 426, 428 and 430 each are centered about the frequency for the tones fl, f2, f3, and f3, respectively. The tone representing the binary bit Bl passes through bandpass filter 424 while other tones, such as tone f3, are filtered by the bandpass filter fl. Decoder 432 identifies the tone fl in a manner similar to the IBS decoder described in FIGS. 11-13, only for a single tone. Because the fl tone was detected by decoder 432, a binary "1" value is generated representing bit Bl in the four bit sequence.

Because no f2 tone will be detected by decoder 434, a binary "0" value is generated for bit B2 in the four bit sequence. Decoder 436 detects an f3 tone and accordingly generates a binary "1" value for bit B3. Decoder 438 generates a binary "0" value for bit B4 because no f4 tone was generated by the multicarrier encoder. A packet assembler 440 receives the four bits B1-B4 and places them into the appropriate IBS packet location in the data buffer 442.

Having described and illustrated the principles of the invention in a preferred embodiment thereof, it should be apparent that the invention can be modified in arrangement and detail without departing from such principles. I claim all modifications and variation coming within the spirit and scope of the following claims.