WO1998040966A1

WO1998040966A1 - System and method for high speed digital data casting concurrent with amplitude modulated broadcasts

Info

Publication number: WO1998040966A1
Application number: PCT/US1998/004948
Authority: WO
Inventors: Timothy F. Dyson; Joseph R. Burns
Original assignee: Data Design & Development Corporation
Priority date: 1997-03-14
Filing date: 1998-03-13
Publication date: 1998-09-17
Also published as: AU6462398A

Abstract

A system and method for transmitting digital information concurrent with a co-existing monophonic amplitude modulated (AM) signal are disclosed. The system allows any AM transmission, in particular commercial AM Radio Broadcasts, to incorporate an orthogonally modulated data signal without significant interference to the ongoing AM broadcast. A high bandwith efficiency modulated data signal is subcarrier modulated (14) onto one or more subcarrier signals. The composite data carrier (22) is then quadrature phase modulated (32, 36) onto the broadcast carrier signal (42, 44, 46) before amplitude modulation. The rate of data transmission, measured in bits-per-second, can be matched to the intended transmission channel (16) and AM modulation characteristics in order to optimize delivery performance and quality of service. In the preferred embodiment, data is phase modulated onto a set of orthogonal base band carrier tones which are then used to angle modulate (54) the RF Carrier. Transmission and receiver methods are disclosed which utilize existing state-of-the-art Compatible Quadrature Amplitude Modulation (C-QUAM) AM Stereo phase modulation Exciter with small modification.

Description

SYSTEM AND METHOD FOR HIGH SPEED DIGITAL DATA CASTING CONCURRENT WITH AMPLITUDE MODULATED BROADCASTS

FIELD OF THE INVENTION

This invention relates to a system and method for transmitting digital information concurrent with a co-existing monophonic amplitude modulated (AM) signal.

DESCRIPTION OF RELATED ART A variety of methods have been proposed for concurrent transmission of two or more information bearing signals along with an amplitude modulated carrier. Most of these methods achieve relatively slow data rates as measured in terms of bits-per-second. The current standard commercial AM Stereo broadcast is one such method for analog information delivery which uses quadrature modulation of the RF Carrier to achieve the equivalent phase modulation. A U.S. Pat. 4,688,255 to Leonard R. Kahn is an example of a proposed method to deliver auxiliary data along with a stereo AM broadcast by additively incorporating an additional information bearing carrier signal. This auxiliary carrier is modulated by any step familiar to those knowledgeable in the state-of-the-art.

All of these previous methods have been restricted to low data rates due to channel bandwidth and interference considerations. In particular, the Kahn method is similarly restricted to comparatively low data rates due to these same concerns and the placement of the auxiliary carrier at the AM Broadcast carrier frequency spacing. This limits the auxiliary data rate for two reasons. The first is the need to keep the signal level of the auxiliary signal plus modulation well below (at least 25 dB) the unmodulated AM Carrier by Federal Communications Commission (FCC) regulation CFR 47, Ch. 1 Sect. 73.44. Secondly, to minimize mutual interference to the AM program broadcast and remain compatible with AM Stereo broadcasts, auxiliary modulation bandwidth must be controlled to remain within the region of the "mixed highs" of the AM broadcast. This leads to practical limitation of the modulation bandwidth of the auxiliary information to less than 2 Khz. Thus, in Kahn, even though the auxiliary carrier is placed at the adjacent station null and is in phase quadrature to the main carrier, modulation sidebands created by information modulation of any appreciable rate will fall within the modulation sidebands of the co-existing AM broadcast. These sidebands will also extend well into the spectrum of any adjacent AM channel. Thus, interference control again limits the achievable data rate.

Further, many of the prior art methods utilize angle modulation which causes an additional problem with the received signal. Angle modulation frequently interferes with the main broadcast reception by inadvertent activation of receiver stereo decoding when no C-QUAM stereo modulation is present. In these cases, the auxiliary data signal is incorrectly decoded as audio and directly interferes with the main transmission audio. Finally, propagation effects such as transmitter antenna and power amplifier non-linearity, multipath, and mis-tuned receivers can introduce inadvertent mutual interference. The latter is controlled by electronic tuning which is the most prevalent today. Multipath is a relatively insignificant effect under most operating conditions. Transmitter non-linearities have generally been adequately addressed at all AM Broadcast stations currently employing AM Stereo (C-QUAM) transmissions.

In all cases of casting data with concurrent AM transmission, data modulation must address the issues of (1) mutual-interference with the AM Broadcast amplitude modulation interfering with the data signal reception (self- induced or "virtual" fading), (2) transmission channel effects due to motion between transmitter and receiver (Doppler, fading) and (3) incidental stereo demodulation activation by the data signal.

The first of these issues, mutual-interference, is addressed in the prior art of

Kahn, by (a) the placement of the additive sidebands at the channel spacing (10 kHz in the US), and (b) by limiting the bandwidth of the data modulation away from the audio signal. The latter tacitly limits the data rate by trading off bandwidth of the modulating waveform for interference control.

Audio interference with the data signal is also a mutual-interference problem. Extreme negative modulation of the carrier envelope by the primary (e.g., voice) transmission reduces the available carrier energy to the receiver. This appears as a virtual fade of the received signal without propagation effects being present. In Kahn, the data modulation rate is varied in direct proportion to the audio signal level. Thus in periods of weak carrier small or no data are present so interference is controlled. This approach also reduces the effective data rate of the method. The second issue, transmission channel effects, is addressed in the prior art by three aspects of the data modulation: embedded timing and synchronization, message packetizing and symbol design, and modulation mechanism. In mobile reception of a data carrying signal, flat fading causes carrier energy to be severely attenuated. In this circumstance, recovery of phase modulated information can be interrupted to the extent that re-acquisition of the carrier would be necessitated. This is a frequent occurrence in prior art AM Transmission.

The last issue, that of inadvertent AM Stereo demodulation, can cause severe interference with the main AM broadcast on those AM Radio Receivers equipped for stereo reception according to the C-QUAM standard. This phenomenon has been reported in the "Report of the National AM Stereophonic Radio Committee", Electronic Industries Assoc, Dec. 1977. One solution to this problem, as shown in the publication "Introduction to the Motorola C-QUAM AM Stereo System", Motorola, Inc., 1992, is the use of a 25 Hz pilot tonal signal which is phase modulated onto the AM Carrier and which acts as the stereo indicator to a receiver. Prior art attempts to insert data below the 100 Hz audio cutoff have caused this inadvertent stereo demodulation and its associated audible interference. In the current invention, the spectrum of the baseband data signal is controlled to specifically null the lower spectral region at and around 25 Hz, eliminating all intentional modulation energy in this lower spectrum.

As pointed out in the discussion of Kahn above, most prior art uses angle modulation to carry auxiliary data. However, the prior art in Kahn is substantially a linearly superposed additional amplitude modulated carrier, albeit one in phase quadrature with the AM Carrier. Being an additive additional signal, it carries most of the disadvantages of amplitude modulation and the current state-of-the-art known as In-band, On-channel (IBOC) modulation. In particular, data rates are restricted by the need to limit amplitude and bandwidth with respect to the primary modulation in order to control self-interference.

SUMMARY OF THE INVENTION The current invention overcomes many of the above noted problems of the prior art and thereby achieves higher data rates by using strictly non-linear modulation of the AM Carrier coupled with high efficiency subcarrier modulation made possible by accurate synchronization and timing recovery.

Briefly described, the present invention enables concurrent data and main program transmission with a minimum of interference at substantially higher data rates than in the prior art. This higher efficiency is the result of maximizing the available baseband bandwidth and controlling the phase modulation index in order to minimize mutual interference. The invention relates to many aspects of the transmission and reception of encoded digital information. At the transmission site, a Data Inserter is used which encodes, packetizes, channel encodes and subcarrier modulates the data signal into a quadrature phase modulated baseband format. A RF Exciter (or modified AM Stereo Exciter) then uses quadrature modulation to effect the constant envelope phase modulation of the AM Broadcast carrier with the baseband signal including embedded synchronization and identification mechanisms. The information signal, in one embodiment, is Error Correction coded, m-ary DPSK modulated onto an orthogonal set of 128 subcarrier signals using Gray Coding. The symboling rate is 9500/128 symbols per second per subcarrier.

The current invention organizes all data into self-contained transmission packets of finite extent. This allows continuous re-acquisition and re- synchronization of the received signal fostering rapid recovery when severe fading or loss of signal occur. Packet synchronization and symbol timing recovery can be accomplished as described in US Patent 5,220,584 wherein phase transitions of a reference signal are detected and synchronized to. In the preferred embodiment, these issues are addressed by incorporating a uniquely formulated spread spectrum sequence into each transmission packet. This design relies on the sequence having an autocorrelation function with high peak-to-sidelobe ratio. Thus, matched filtering to this common wide bandwidth sequence allows large processing gain so that highly reliable packet detection and resynchronization is assured. It is well-known that propagation characteristics of the AM, or Medium

Frequency, radio channel varies widely across the United States. Therefore, quality is not uniform in all places or at all times. In response to this variation, the invention allows the flexibility to "gear-shift" the baseband digital modulation density to match the outgoing data rate to the available channel. When propagation conditions deteriorate, modulation type can be changed to lower the data rate and thereby improve the reception quality as measured in terms of the received bit error rate (BER). The current invention encodes the modulation type as part of the transmitted packet immediately following the synchronization signal. Both of these signals are designed as spread spectrum waveforms which allow highly sensitive recovery and detection at the receiver through the use of properly designed matched filter receivers. Such filters can be designed and implemented in many alternative ways by those skilled in the art.

The current invention, like nearly all other methods previously proposed, uses angle modulation of the AM Carrier frequency to convey the information bearing signal. However, in concert with CFR 47, Ch.1 Sect. 73.127 of the FCC,

AM Stereo Modulation is not employed at the time of auxiliary data transmission. In this case, the carrier phase is available for auxiliary data modulation. The current invention phase modulates the AM Carrier and controls both the spectrum of the resulting sidebands and the modulation index so as to minimize interference with the concurrent AM Broadcast. Since phase modulation of the AM Carrier is orthogonal to any amplitude modulation, envelope detection receivers are largely unaffected by the presence of the auxiliary data. In the preferred embodiment, the information bearing signal phase modulates the AM Carrier by employing quadrature amplitude modulation to effect the trigonometric double angle identity. This embodiment allows the data signal to utilize existing AM Stereo Exciters after minor modification, to provide carrier modulation.

In another aspect of the invention, a receiver and data demodulator are disclosed which make maximum advantage of the angle modulated nature of the data signal. The receiver performs phase-locked loop (PLL) tracking of the carrier/IF to persist through carrier fades, self-synchronize to the transmitted subcarrier and recover symbol timing in order to synchronously demodulate the auxiliary data.

These and other features of the invention will be more fully understood by reference to the following drawings. BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1A is a schematic diagram of the auxiliary data insertion and modulation process according to the preferred embodiment of the invention.

FIG. 1 B is a schematic diagram of a conventional C-QUAM Stereo Exciter. FIG. 2 is a graphical presentation of the data encoding and modulation structures used for digital modulation.

FIG. 3A is a schematic diagram of the invention's use of multiple orthogonal subcarrier modulation method.

FIG. 3B depicts the preferred embodiment's efficient method of digitally generated orthogonal subcarrier modulation. FIG. 4 is a graphic diagram of the invention's phase modulated signal spectrum as produced in the preferred embodiment.

FIG. 5 depicts the preferred embodiment's transmission packet structure incorporating timing recovery and phase reference information. FIG. 6 is a schematic of the preferred embodiment of the quadrature modulator arrangement to constant envelope phase modulate a carrier signal

FIG. 7 is a schematic diagram of the radio receiver design according to the preferred embodiment of the invention.

FIG. 8 is a diagram of the process of digital demodulation in the invention's preferred embodiment.

FIG. 9 is a depiction of the complex demodulator used in the preferred embodiment of the invention.

FIG. 10 is a schematic representation of the method of frequency division demultiplex used in the preferred embodiment of the receiver-demodulator aspect of the invention.

FIG. 11 is a graphic illustration of byte level de-interleaver in the preferred embodiment.

DETAILED DESCRIPTION OF THE INVENTION During the course of this description like numbers will be used to identify like elements according to the different figures which illustrate the invention. 1. Transmitter

1.1 Digital Subcarrier Generation and Modulation

The preferred embodiment of data signal subcarrier modulation and carrier phase modulation is illustrated in FIG. 1A. Baseband modulation is accomplished by first applying an 10 Error Correction Encoding of the incoming information source. This error correction encoding is optional and may be omitted in alternative embodiments. The preferred method is an efficient block code such as a Reed- Solomon code for which methods are well known to those skilled in the art. Redundantly encoded data is next 12 Byte Interleaved using a block interleave based on the packet structure. Each transmission packet, depending on the selected modulation type 16, consists of a fixed number of encoded data blocks.

FIG 2 illustrates the data structures as the data is organized and processed for packet transmission. In the preferred embodiment, M encoded data blocks 72, used to form a transmission packet, is shown in Fig. 2 along with the associated data bytes. Byte interleaving 74 re-organizes the encoded data in preparation for digital modulation 14. Modulation results in a set 78 of K Data Frames, each containing N channel coded symbols. In the preferred embodiment, the channel coding is differentially encoded m-ary PSK, so a phase reference frame 79 is included at the head of the data. Packet framing 20 concatenates the unique Synchronization sequence 18, the Modulation type identification sequence 16 and the assembled data frames to form a transmission packet. Data frames are generated at a sample rate equivalent to four times the symbol rate. In one embodiment, as illustrated in the example in FIG. 2, each data frame is generated from 128 16-PSK symbols. The invention uses multiple orthogonal subcarriers for the baseband digital modulation. FIG. 3A depicts the Digital Modulation 14 in more detail. Referring to FIG. 3A, interleaved bytes are mapped 80 to the appropriate symbol (Trellis Coded, PSK, etc.) and these symbols are loaded into a serial-to-parallel buffer 82 to form the basis of a data frame. Each parallel symbol is integrated 84 with the corresponding symbol in the immediately previous frame. Each symbol of a frame 86 modulates one of N orthogonal tonals which are subsequently summed 88 to form the baseband data frame time waveform. Thus the digital modulation is a form of frequency division multiplex. Baseband frame samples are interpolated 94 to the required output rate and then a guard band is appended. This guard band circularly extends each data frame by a time span sufficient to support uncorrupted decoding. In normal practice the guard band is equal to the expected multipath differential delay. In the current invention, multipath effects are not manifest in a linear manner. The guard band is still used in the invention however to counteract the symbol smearing effects of timing jitter in the receiver/demodulator. The preferred embodiment of the digital modulator is seen in FIG. 3B. An inverse fast-Fourier Transform (IFFT) 108 performs the subcarrier modulation and summation of the differentially encoded symbols of buffer 106. 104 represents a unit delay in feedback of the integrated symbol buffer. Therefore, T is the reciprocal of the orthogonal subcarrier spacing, or the time span of the N input symbols. The maximum signaling rate is therefore (N/T) symbols-per-second. Assuming m-bit symbols, the maximum transmission rate allowed is m(N/T) bits-per-second. The need for the guard band limits the actual rate to (T/(T+g)) times the maximum possible rate, where 'g' is the length in seconds of the guard band inserted.

The preferred embodiment is the most efficient modulation method for several reasons. First, the IFFT is well known to those skilled in the art to be computationally optimal for the process employed. Second, it provides an equally efficient and effect method of performing sample rate interpolation. By performing an IFFT of length L > N, setting the unassigned input points to zero identically, the resulting time series will represent the sin(Nx)/sin(x) interpolated series having the interpolation ratio (UN). Having L a highly composite value (e.g., power of 2, 3 or 4 or product of primes) results in maximum efficiency. FIG. 4 depicts the spectrum of the composite phase modulated AM Carrier. The IFFT facilitates the selective nulling of the frequency region immediately surrounding the carrier frequency.

Assuming L > 2N, the N subcarrier tones may be selected to be any subset of the first half of the L tonal positions of the IFFT input. In the preferred embodiment, 128 orthogonal tonals are used and L=512. The input symbols for a data frame are aligned in the input buffer at frequency "bins" 3 through 130. The first 100 Hertz to either side of the AM Carrier are therefore nulled in the sidebands generated by the phase modulation. FIG. 5 shows the final transmission packet structure in the invention's preferred embodiment.

1.2 RF Carrier Modulation

As depicted in FIG 1A, successive transmission packets are concatenated and filtered to the required bandwidth (<10,000 Hz in the preferred embodiment) using a complex-valued, half-band, low pass finite-impulse response (FIR) filter 22. To those skilled in the art such filters are well known to have linear phase response and permit efficient implementations. Continuously filtered transmission packet data samples (at 38,043.478 Hz sampling rate in the preferred embodiment) are quadrature phase modulated by taking the 28 SIN and 24 COS of each sample and then converting the two orthogonal sequences to analog form using a dual, synchronous analog-to-digital converter 26. The imparted timing jitter in the conversion process must be carefully controlled to a small fraction of the symboling rate in order to limit imparted phase noise to the waveform synchronism.

Baseband in-phase (I) and quadrature (Q) analog waveforms from the 52 Data Inserter are applied to the "L" and "R" inputs to a modified C-QUAM compatible Stereo Exciter, such as Broadcast Electronics Model AX-10 AM Stereo Exciter. Required modification to the standard C-QUAM Exciter, FIG. 1 B, entails configuring the L+R 58 and L-R 60 quadrature modulation circuits to remove the 10% L-to-R and R-to-L audio mixing 56 common to stereo and converting the (L+R +dc) input to the Data Inserter "I" output, and the (L-R) input to the Data Inserter "Q" output. The 25 Hz stereo pilot is also removed from the (L-R) modulator input path. For example, FIG. 6 is a schematic of the quadrature modulation circuit that results from this modification to the AX-10 Exciter. This modification assumes differential inputs from the Data Inserter 52 in the preferred embodiment, wherein, I and l-not are input to amplifiers 320 and 322 respectively. Q and Q-not are input to associated amplifiers 324 and 326, respectively. 320 and 322 incorporate 1 1 K potentiometers for gain balancing between the in-phase and quadrature channels to control modulation phase error in the IF Carrier. Clocking source 328 clocks at a 50% duty cycle at the chosen IF frequency (nominally 250khz in the preferred embodiment). 328 outputs four orthogonally phased clocks to chopping modulator switch 330 where the differential I and Q signals are quadrature modulated by the IF signal. Summation junction 334 adds the now in phase differential waveforms to produce the quadrature IF signals:

I {cos(ω_ct)+a₁ cos(3ω_ct)+a₂ cos(5ω_ct)+ ...} Q {sin(ω_ct)+a₁ sin(3ω_ct)+a₂ sin(5ω_ct)+ ...}

Filter 332 serves to remove the unwanted odd order harmonics and sums the quadrature components to produce the desired phase modulated IF Carrier. The output of this modulator circuit is now passed to a limiter-amplifier 40 and to additional filtering 42, as required. The signal is now passed to AM Modulator 44 and the RF Up-converter 46 to place the fully modulated signal at the desired AM Broadcast frequency.

2. Receiver-Demodulator 2.1 Radio FIG. 7 provides a simplified schematic of the invention's phase modulation radio receiver. The receiver is designed to operate in the AM Broadcast band in the preferred embodiment and as such builds upon the available C-QUAM integrated circuit technology. In the preferred embodiment, Motorola part numbers MC13030, MC145157-2 from the C-QUAM Radio chip set are utilized. FIG. 7 is one approach of those equivalent others that could be arranged by those skilled in the art. Referring to FIG. 7, the radio is electronically tuned to the desired broadcast frequency by a PLL circuit used to adjust the voltage controlled oscillator 408 to the required mixing frequency, F_RF + 10.7Mhz, used in mixer 404. Phase Detector/Comparator 412 provides an output signal proportional to the phase difference of the reference tuning setting and the gain adjusted VCO 408 output signal. Loop filter 410 is set to allow relatively quick response frequency adjustment. Mixer 404 output is amplified into the Bandpass Filter 416 and Limiter 418 combination to produce the constant envelope phase modulation. Reference Oscillator 432 provides the phase referencing signal for the second stage down conversion and for the Demodulator Analog-to-digital conversion clock. The limiter 418 output is mixed with the PLL 436 controlled secondary IF signal. This signal is used to phase lock the second IF output to the reference frequency phase. PLL 436 is also used to track minor variations in the carrier frequency, hence the reference to carrier-follower design. IF Filter 422 is designed to allow a bandpass of not more than 28,000 Hz centered at the nominal IF frequency, 76,086.95 Hz or one-fourth the conversion clock phase reference frequency. Since the nominal spectral bandwidth of the significant part of the phase modulated IF signal is approximately 2.2 times the baud rate of the digital modulation (i.e., about 21 ,250 Hz), this IF is a reasonable compromise between noise rejection, adjacent channel rejection and carrier tracking.

The reference frequency out of divide down circuit 414 is at a ratio of 23:1 from the input reference oscillator frequency. PLL loop filter 428 is designed to provide a loop bandwidth of 100 Hz or less. This should be compared to the phase modulated IF spectrum, as depicted in FIG. 4, and insures that the tracking loop does not respond to data modulation of the carrier. Loop design is set to freewheel from the last state whenever the input to phase comparator 430 drops below the locking threshold. This insures that under conditions of RF drop-out due to fading or large negative amplitude modulation from the audio signal, the receiver will recover lock on the RF Carrier (i.e., IF) without excessively long re-acquisition time.

2.2 Digital Demodulator

2.2.1 1F Phase Demodulation

FIG. 8 is a depiction of the process of recovering the base band data modulation from the phase of the IF signal. The figure depicts one method (i.e., a digital method) of performing phase demodulation of the IF signal: analog-to-digital 440 conversion followed by a complex demodulation 442 and arctangent 444 computation. Several other methods are available to phase demodulate the IF signal as could be devised by those skilled in the art. The preferred embodiment is as depicted for the purposes of efficient use of the requisite implementation Digital Signal Processing hardware. The Complex Demodulator 442 of the preferred embodiment is as depicted in FIG. 9. IF waveform is sampled synchronously to the carrier phase by virtue of PLL 436 as shown in FIG. 7. The sampling rate (F_s) of the preferred embodiment is four times the IF frequency, F_s = 304,347.82 Hz. The first step in the complex demodulation process is the complex bandshift to baseband of the IF signal. Because of the chosen relationship between the IF frequency and F_s , the two mixers 480 are created by a table of 4 sin/cos look-up values and then two multiplies per incoming sample. The resulting two data channels are low pass filtered (nominally) to the signaling baud rate, approximately 9500 Hz in the preferred embodiment. The sampling rate of these channels is also decimated to maximize efficiency of the invention. Those skilled in the art will note that sampling rate decimation is often most efficiently accomplished through staged filtering and rate reductions. The preferred embodiment use two such stages and the use of half-band FIR digital filters 482 and 486 leads to the most efficient method of filtering and simultaneous down-sampling to target rate of four times the baud rate: 38043.47 Hz in the preferred embodiment. Decimation is performed as an integral part of the FIR filtering computation and occurs in a decimation by 4 followed by decimation by 2 in the second FIR filter.

Final phase demodulation occurs by computing the angle associated with the phasor represented by the in-phase and quadrature phase complex demodulator output. The arctangent computation 444, implemented by a quadrant based successive approximation/table look-up procedure. The arctangent look-up table is designed to limit the maximum angular error to less than one degree of phase. 2.2.2 Packet Detection and Frame Synchronization

A critical aspect of data recovery is to accurately determine when a transmission packet has arrived and to synchronize to the data framing structure, as indicated in FIG. 2. The invention facilitates packet detection and data synchronization by incorporating a known polyphase sequence in every transmission packet. This sequence is optimally detected in the receiver- demodulator by applying matched filter 448 to the phase demodulator output data. The matched filter (MF) output is sent to threshold detector 450 which compares the local maximum MF output power to the surrounding RMS power over the processing block. The comparison 446 implements the maximum likelihood test for sequence presence at fixed probability of false detection. The preferred embodiment utilizes a sequence with a bandwidth-time product of 128 (21 dB processing gain).

When packet presence detection 446 has occurred, data frame synchronization is achieved by sample selection of the central N samples in each anticipated data frame offset from the packet detection MF peak time index. The guard band inserted at the modulator now comes into play in that any timing alignment jitter due to noise or weak signal is offset by the timing guard band on each data frame. Cumulative timing error across a transmission packet, due to synchronization timing error, clock drift, differential sampling skews or Doppler is also addressed by the use of guard bands. The length of each guard band must be selected to at least twice the total expected timing drift over a packet duration divided by the number of data frames per packet. As an example, a design for commercial AM Broadcast channels and 50 ppm (parts per million) clock-oscillators would allow for approximately 9 msec/second of timing jitter at the receiver. This approach also allows a trade-off of transmission packet duration against timing recovery overhead.

2.2.3 Frequency Demux and Soft-symbol Recovery

The following disclosure refers back to subcarrier modulation description and, in particular, the timing references of FIG. 5. Subcarrier signal recovery (FIG 10) is accomplished through application of a fast Fourier transform to perform frequency demux. Beginning at the (phase) reference frame, (T_h + T_m + g/2) seconds after the leading edge of packet sync, L samples are loaded into FFT input buffer 500. An L-point FFT 504 is performed once every T_r = (T + g) seconds and the phasors associated with frequency bins assigned to the subcarriers, number three through 130 in the preferred embodiment, are selected from the FFT output and loaded into the differential decoder 502 history buffer. The current output phasors are also passed to the decoder where they are complex conjugated and multiplied with the immediately previous FFT output phasors at the corresponding frequency bins to effect differential decoding 506. The first FFT output phasors following packet synchronization represent the reference frame. The first output from the differential decoder following packet synchronization will be the complex conjugated reference frame for use in channel assessment.

2.2.4 Symbol Decode/Frame Recovery

As depicted in FIG. 8, Symbol decode 464 is effected in each embodiment according to the digital modulation format employed at the modulator. The decoding operation in all cases maps the differentially decoded phasors representing soft- symbols to a corresponding bit pattern representing the transmitted encoded binary data. Methods of accomplishing this are varied and well-known to those skilled in the art. Frame recovery 466 assembles the decoded bits from each of the N symbols per data frame every T_r seconds. Because the modulation type is encoded in each packet header, symbol decode 464 and frame recovery 466 will produce a variable amount of binary data according the bandwidth efficiency used at the transmitter. The table below codifies the relationship between modulation density, in bits-per-symbol (bpsy), and the number of Reed-Solomon blocks per transmission packet assuming a fixed packet length of 27 data frames, as in the preferred embodiment:

TABLE 1 bpsy bit/frame bytes/pkt blks/pkt

2 256 864 12

3 384 1296 18

4 512 1728 24

5 640 2160 30 m 128m 432m 6m

Frame recovery aggregates decoded symbols into bytes. Bytes are organized into blocks of 72 and arranged in the de-interleave buffer in accordance with TABLE 1. 2.2.5 Error Correction

De-interleaver 468 in the preferred embodiment is designed to work with an interleave pattern which disperses code symbols in both time and frequency as illustrated in FIG 11. Those skilled in the art will be able to devise suitable alternative interleave functions provided interleaving spans at least one transmission packet. The selected algorithm is defined as follows:

If memory byte address order of the frame recovery output buffer is defined as: a(k) = k ; k=0, 1 , 2, ... 72M where M is the appropriate value from Table 1. Viewing the interieaver input as an Mx72 byte matrix 508, consecutive encoder symbols (i.e., bytes) are mapped according to 510. This mapping is defined as follows: Define the parameters: m > minimum effective code length (odd and relatively prime to 72 and M)

N = block length of encoder (In the preferred embodiment, N=72) M = encoder blocks per packet k = 0, 1 MN-1 p = k m mod(MN) row(p) = row index of interleave result for input byte address 'p' col(p) = column index of interleave result for input byte address 'p' then row and col are given by col(p) = k mod(N) row(p) = p mod(mN)

The de-interleaver re-assembles the original byte order by placing consecutive frame recovery output bytes at memory locations separated by the distance 'nϊ. That is, if n = 0,1 , ... , MN-1 represent the consecutive output order of bytes from frame recovery, then the corresponding de-interleaved destination address is given by p = n m mod(MN).

Error correction 470 performs Reed-Solomon error detection and correction processing is applied to the de-interleaved transmission packet buffer 514. Reed- Solomon (RS) decoding is well-known and has several efficient alternative available implementations, as known by those skilled in the art. In the preferred embodiment coding blocks are organized around 50 byte data blocks. Accordingly, the preferred embodiment performs RS Decoding of the RS(72,50) coded blocks using the well known Berlekamp-Massey and Forney Algorithms. The preferred embodiment of the invention uses a table driven methodology based on 8-bit symbols in the Golais field associated with the primitive polynomial:

p(z) = z⁸ + z⁷ + z² + z + 1

and the code generating polynomial: g(z) = z²² + g₂₁z²¹ + ... + g₁z + g_o

created in the usual manner, as understood by those skilled in the art.

The output from RS Decoding results in 50 byte blocks in the preferred embodiment. Therefore, 300m bytes per transmission packet are delivered to the intended application by the invention. This equates to a payload rate of 6126.16m bits-per-second or a conservative minimum of 6126 bps. In particular, Table 2 provides the generator for the preferred embodiment.

Table 2

Generator Polynomial Coefficients Used In Preferred Embodiment

Zec 's coef # log Binary Code

³21 36 - 01010111

'20 21 -01011111 g 19 167-11100110

22 - 10111110 228 - 10111000

9,7 "l6 147-11010101

50 - 00110110 θι g 13-00111101 g_i3 194-01001001

9₁₂ 74 - 10111011 73 - 10011110

011 g ^a10 97 - 01100001 g_g 240 - 00110011 g₈ 82 - 10110111 g₇ 142-00100010 g₆ 7-10000000 g₅ 111-11100000 g₄ 183-11111111 g₃ 96-11110011 g₂ 228 - 10111000

As noted above, the present invention overcomes many shortcomings of the prior art to thereby more efficiently transmit digital data concurrent with AM monophonic broadcasts. As an example, in the preferred embodiment, interference with the audio is addressed by isolating all data modulation to a subcarrier based signal which in turn phase modulates the carrier. The subcarrier is permitted to fill the audio band to the fullest extent possible with a limit at the low frequency end at 100 Hz to avoid inadvertent activation of the stereo decoder in AM Stereo receivers. Thus a minimum of 9 K Hz of bandwidth is available for data modulation, nearly four times that possible in Kahn or any other previous approach. While the invention has been described with reference to the preferred embodiment thereof, it will be appreciated by those of ordinary skill in the art that various modifications can be made to the structure and function of the individual parts of the system without departing from the spirit and scope of the invention as a whole.

Claims

WHAT IS CLAIMED IS:

1. A data insertion apparatus (52) for high rate digital data casting concurrent with AM monophonic broadcasts comprising: a. a converting means for converting binary data into subcarrier waveforms

(14); and, b. a subcarrier modulating means for modulating said subcarrier waveforms onto an AM transmitter's carrier phase (54).

2. The apparatus of Claim 1 wherein said subcarrier modulating means comprises a means for quadrature phase modulation (32, 36) of the transmitter's carrier.

3. The apparatus of claim 2, wherein said rate is matched to the AM transmitter's modulation characteristics and intended transmission channel (22).

4. The apparatus of claim 3, wherein said subcarrier modulation means comprises a means for C-QUAM compatible signaling (328, 330, 332, 334).

5. The apparatus of claim 4, wherein said subcarrier modulation means comprises a means for modulating a set of 2^m orthogonal subcarriers, m being a non-negative integer, using block error correction coding of interleaved data (80, 82).

6. The apparatus of claim 4, wherein said subcarrier modulation means comprises a means for modulating a set of 2^m orthogonal subcarriers, m being a non-negative integer, using differentially encoded Trellis Coded PSK modulation (80) of interleaved data.

7. The apparatus of claim 4, wherein said subcarrier modulation means comprises a means for reducing timing jitter using symbol guard bands (96, 114).

8. An AM Band receiver apparatus for receiving high rate digital data casting concurrent with AM Monophonic broadcasts comprising decoder circuits (436) wherein the phase of the AM transmission is carrier phase demodulated (442, 444).

9. The apparatus of claim 8 wherein the decoder circuits comprise a carrier locked-loop receiver (402-418, 436).

10. A baseband subcarrier demodulator apparatus wherein digital data is recovered and decoded from the phase demodulated AM Carrier signal, said system comprising: a. a frame synchronization and symbol timing recovery means (446) comprising a spread spectrum code symbol sequence (452) and a matched filter detection processor (450); b. a subcarrier frequency recovery means (436) comprising a hard limiter (418), a phase detection means (430), short time averager (422) and second order tracking loop means (436) to estimate the residual frequency shift, to thereby bandshift the subcarrier back to its nominal transmitted frequency; and, c. a subcarrier decoding means for decoding of the subcarrier comprising differential phase detection (460), soft symbol Trellis Decoding (464), de- interleaving (468) and block error correction (470).

11. The apparatus of claim 10 wherein the spread spectrum code symbol sequence is a pseudo-random sequence (452).

12. The apparatus of claim 10 wherein the spread spectrum code symbol sequence is a Barker code sequence (452).

13. The apparatus of claim 10 wherein the spread spectrum code symbol sequence is a polyphase code sequence (302).

14. A data insertion method for high rate digital data casting concurrent with AM monophonic broadcasts comprising the steps of: a. converting binary data into subcarrier waveforms (10-14); and, b. phase modulating said subcarrier waveforms onto an AM transmitter's carrier (54).

15. The method of Claim 14 wherein said phase modulating step comprises quadrature phase modulation of the transmitter's carrier (24, 28, 32, 36).

16. The method of claim 15, wherein said rate is matched to the AM transmitter's modulation characteristics and intended transmission channel (20, 22,

304).

17. The method of claim 16, wherein said phase modulation step comprises

C-QUAM compatible signaling (Fig. 6).

18. The method of claim 17, wherein said phase modulation step comprises modulating a set of 2^m orthogonal subcarriers (106), m being a non-negative integer, using block error correction coding of interleaved data.

19. The method of claim 17, wherein said phase modulation step comprises modulating a set of 2^m orthogonal subcarriers, m being a non-negative integer, using differentially encoded (102) Trellis Coded PSK modulation (100) of interleaved data.

20. The method of claim 17, wherein said phase modulation step comprises reducing timing jitter using symbol guard bands (96, 114, 306-310).

21. A method for receiving high rate digital data casting concurrent with AM Monophonic broadcasts (Fig. 7) comprising the step of demodulation of the carrier phase of the AM transmission (442, 444).

22. The method of claim 21 wherein said demodulation step comprises use of a carrier locked-loop receiver (Fig. 7).

23. A baseband subcarrier demodulator method wherein digital data is recovered and decoded from the phase demodulated AM Carrier signal, said method comprising the steps of: a. recovering frame synchronization and symbol timing data, which data comprises a spread spectrum code symbol sequence (452); b. detecting subcarrier frequency information utilizing a hard limiter (418), a phase detection step (430), a short time averager (442) and a second order tracking loop (436) to estimate the residual frequency shift, to thereby bandshift the subcarrier back to its nominal transmitted frequency; and, c. decoding subcarrier information, using differential phase detection (460), soft symbol Trellis Decoding (464), de-interleaving (468) and block error correction (470).

24. The method of claim 23 wherein the spread spectrum code symbol sequence is a pseudo-random sequence (452).

25. The method of claim 23 wherein the spread spectrum code symbol sequence is a Barker code sequence (452).

26. The method of claim 23 wherein the spread spectrum code symbol sequence is a polyphase code sequence (452).