HK1145233B - Method and apparatus for implementing a digital signal quality metric - Google Patents

Description

Method and apparatus for implementing a digital signal quality metric

Technical Field

The present invention relates to digital radio broadcast receivers, and more particularly, to a method and apparatus for implementing a signal quality metric for an OFDM digital signal in a digital radio receiver.

Background

Digital radio broadcasting technology delivers digital audio and data traffic to mobile receivers, portable receivers, and fixed receivers. One type of digital radio broadcast, known as in-band on-channel (IBOC) Digital Audio Broadcasting (DAB), uses terrestrial transmitters on existing intermediate frequency (MF) and Very High Frequency (VHF) radio bands. HD Radio developed by iBiquity Digital corporation^TMThe technology is an example of an IBOC implementation for digital radio broadcasting and reception.

The IBOC DAB signal can be transmitted in a hybrid format that includes an analog modulated carrier in combination with a plurality of digitally modulated carriers or in an all-digital format in which the analog modulated carrier is not used. By using the hybrid mode, broadcasters may continuously transmit analog AM and FM signals having higher quality, and more robust digital signals simultaneously, allowing themselves and their listeners to convert from analog to digital radio while maintaining their current frequency allocations.

One feature of digital transmission systems is the inherent ability to transmit digitized audio and data simultaneously. Thus, the technology also allows for wireless data services from AM and FM radio stations. The broadcast signal may include metadata such as the artist, song title, or station call letter (station call letter). Specialized messages regarding events, traffic, and weather may also be included. For example, traffic information, weather forecasts, news, and sports scores may all be scrolled through on the radio receiver display while the user listens to the radio station.

IBOC DAB technology can provide digital quality audio that is superior to existing analog broadcast formats. Because each IBOC DAB signal is transmitted within the spectral mask (spectral mask) of an existing AM or FM channel allocation, it does not require a new spectral allocation. IBOC DAB promotes spectrum economy while enabling broadcasters to supply digital quality audio to listeners' existing base stations.

Multicasting, i.e., the ability to deliver several programs or data streams on one channel in the AM or FM spectrum, enables a broadcast station to broadcast multiple data streams on separate supplemental or sub-channels of the main frequency. For example, the multiple data streams may include alternate music formats, local traffic, weather, news, and sports. The supplemental channels may be accessed in the same manner as conventional broadcast station frequencies using a tuning or seeking function. For example, if the analog modulated signal center frequency is 94.1MHz, the same broadcast on IBOC DAB may include supplemental channels 94.1-1, 94.1-2, and 94.1-3. Highly specialized programming on supplemental channels can be delivered to closely targeted listeners, creating more opportunities for advertisers to integrate their brands with the programming content. As used herein, multicasting includes transmitting one or more programs on a single digital radio broadcast channel or on a single digital radio broadcast signal. The multicast content may include Main Program Services (MPS), Supplemental Program Services (SPS), Program Service Data (PSD), and/or other broadcast Data.

National radio system commissionThe reception of the banners, an organization for standards set sponsored by the national broadcasting institute and the consumer electronics association, adopted the IBOC standard named NRSC-5A in 9 months of 2005. NRSC-5A sets forth the requirements for broadcasting digital audio and auxiliary data on AM and FM broadcast channels, the disclosure of which is incorporated herein by reference. The standard and its references contain a detailed description of the RF/transport subsystem and the transport and service multiplexing subsystem. Copies of the standard may be made from the NRSC websitehttp://www.nrscstandards.org/standards.aspThus obtaining the product. HDradio of iBiquity^TMThe technology is an implementation of the NRSC-5A IBOC standard. Relating to HD Radio^TMFurther information on the technology can be found inwww.hdradio.comAndwww.ibiquity.comis found.

Other types of Digital Radio broadcasting systems include satellite systems such as XM Radio, Sirius and WorldSpace and terrestrial systems such as Digital Radio Mondiale (DRM), Eureka 147 (Brand DAB), DAB Version 2, and FMeXtra. As used herein, the phrase "digital radio broadcast" encompasses digital audio broadcasts, including in-band on-channel broadcasts, as well as other digital terrestrial broadcasts and satellite broadcasts.

Because many applications require an accurate indication of signal quality, including, for example, a seek-search function, resolution of 300kHz spaced interferers, first adjacent interferer sideband selection, and diversity switching, a measure of the quality of the received digital signal is desirable. It is also desirable to obtain this metric quickly and to be efficient and reliable for FM hybrid and all-digital signals. It is also desirable to enable metric computation for existing HD Radio when implementing metric computation^TMAny changes to the receiver hardware or software are minimized.

Disclosure of Invention

In a first aspect, the present invention provides a method for detecting a digital radio signal. The digital radio signal comprises a series of symbols, each symbol consisting of a plurality of samples. The method comprises the following steps: receiving a digital radio signal, forming (develoop) a correlation waveform (correlation waveform) having peaks corresponding to symbol boundaries, normalizing the correlation waveform, and calculating the peak of the normalized correlation waveform, wherein the peak represents the quality of the received digital radio signal.

The digital radio signal may include upper and lower sidebands and the method may be applied independently to each sideband to produce a peak of the normalized correlation waveform for each sideband. The digital signal quality metric may be verified by calculating a peak index difference (peak index delta). The method may include calculating a peak index corresponding to a peak value of the normalized correlation waveform for each of the upper and lower sidebands. Then, a peak index difference representing a difference between peak indexes of the upper and lower sidebands may be determined, and the peak index difference and the peak value of the upper and lower sidebands may be compared with a threshold. The digital signal quality metric may also be verified by calculating the frequency offset difference between the upper and lower sidebands and determining whether the difference meets a certain threshold, thereby indicating whether the detected signal is the desired signal of interest or an adjacent interfering signal.

In another aspect, the present invention provides a receiver for detecting digital radio signals. The digital radio signal comprises a series of symbols, each symbol consisting of a plurality of samples. The receiver includes an input for receiving a digital radio signal, and a processor for calculating a peak corresponding to a symbol boundary of the normalized correlation waveform, wherein the peak represents a quality of the received digital radio signal.

Drawings

Fig. 1 is a block diagram of a transmitter for use in an in-band on-channel digital radio broadcasting system.

FIG. 2 is a schematic diagram of a hybrid FM IBOC waveform.

FIG. 3 is a schematic diagram of an extended hybrid FM IBOC waveform.

FIG. 4 is a schematic diagram of an all-digital FM IBOC waveform.

FIG. 5 is a schematic diagram of a hybrid AM IBOC DAB waveform.

FIG. 6 is a schematic diagram of an all-digital AM IBOC DAB waveform.

FIG. 7 is a functional block diagram of an AM IBOC DAB receiver.

FIG. 8 is a functional block diagram of an FM IBOC DAB receiver.

Fig. 9a and 9b are diagrams of the IBOC DAB logical protocol stack from a broadcast perspective.

Figure 10 is a diagram of the IBOC DAB logic protocol stack from the receiver perspective.

Fig. 11a is a graphical representation of an OFDM signal in the frequency domain.

Fig. 11b is a graphical representation of an OFDM signal in the time domain.

Fig. 11c is a graphical representation of the conjugate product signal peak representing the symbol boundary.

FIG. 11d is a graphical representation of the conjugate product multiplied by the respective amplitude taper (amplitude taper).

FIG. 12 is a block diagram of one implementation of an acquisition module.

Fig. 13a, 13b and 13c are graphical representations of symbol timing for a peak forming module.

Fig. 14 is a flow chart of a first part of a signal acquisition process.

FIG. 15 is a functional block diagram illustrating an acquisition algorithm.

FIG. 16 is a functional block diagram of sideband combining.

Fig. 17 is a diagram illustrating waveform normalization near symbol boundaries.

Fig. 18 is a graph of normalized correlation peaks.

Fig. 19 is a flow chart of a second portion of the signal acquisition process.

Fig. 20 through 24 are graphs of the probability of stopping at a particular frequency for various conditions in a search application for digital signal quality metrics, in accordance with the present invention.

Detailed Description

Fig. 1-13 and the accompanying description herein provide a general description of an IBOC system, including the broadcaster structure and operation, the receiver structure and operation, and the structure of the IBOC DAB waveform. Fig. 14-24 and the accompanying description herein provide a detailed description of the structure and operation of an acquisition module for implementing a digital signal quality metric in accordance with aspects of the present invention.

IBOC system and waveform

Referring to the drawings, FIG. 1 is a functional block diagram of relevant components of a studio site 10, FM transmitter site 12 and Studio Transmitter Link (STL)14 that may be used to broadcast FM IBOC DAB signals. The studio site includes studio automation equipment 34, an overall operations center (EOC)16, and an STL transmitter 48, where the overall operations center (EOC)16 includes an inputter (importer)18, an exporter (exporter)20, and an Exciter Auxiliary Service Unit (EASU) 22. The transmitter site includes an STL receiver 54, a digital exciter 56, and an analog exciter 60, wherein the digital exciter 56 includes an exciter engine (exgine) subsystem 58. Although in fig. 1 the exporter is at the radio station's studio site and the exciter is at the transmitting site, the units may be co-located at the transmitting site.

At the studio site, the studio automation equipment supplies Main Program Service (MPS) audio 42 to the EASU, MPS data 40 to the exporter, Supplemental Program Service (SPS) audio 38 to the exporter, and SPS data 36 to the importer. MPS audio serves as the primary audio program source. In hybrid mode, it preserves the existing analog radio program format in both analog and digital transmission. MPS data, also referred to as Program Service Data (PSD), includes information such as music title, artist, album name, and the like. The supplemental program service may include supplemental audio content as well as program related data.

The importer contains hardware and software for provisioning Advanced Application Services (AAS). A "service" is content delivered to a user via IBOC DAB broadcast, and an AAS may comprise any type of data not classified as MPS, SPS or Station Information Service (SIS). SIS provides station information such as call signs (call signs), absolute times, locations associated with GPS, and the like. Examples of AAS data include real-time traffic and weather information, navigational map updates or other images, electronic program guides, multimedia programs, other audio services, and other content. Content for the AAS may be supplied by a service provider 44, with the service provider 44 providing service data 46 to the importer via an Application Program Interface (API). The service provider may be a broadcaster located at a studio site or a third party service and content provider that is an external source. The importer can establish session connections between multiple service providers. The inputter encodes and multiplexes the service data 46, the SPS audio 38, and the SPS data 36 to produce the outputter link data 24, which is output to the outputter via the data link.

The exporter 20 contains the hardware and software necessary to provision the main program service and SIS for broadcast. The exporter accepts digital MPS audio 26 over an audio interface and compresses the audio. The exporter also multiplexes MPS data 40, exporter link data 24, and the compressed digital MPS audio to produce exciter link data 52. In addition, the output accepts analog MPS audio 28 at its audio interface and applies a preprogrammed delay thereto to produce a delayed analog MPS audio signal 30. This analog audio can be broadcast as a backup channel to the hybrid IBOC DAB broadcast. The delay compensates for the system delay of the digital MPS audio, allowing the receiver to mix between the digital and analog programs without a shift in time. In an AM transmission system, the delayed MPS audio signal 30 is converted by the outputter to a mono (mono) signal and sent directly to the STL as part of the exciter link data 52.

The EASU 22 accepts MPS audio 42 from the studio automation equipment, rate converts it to the appropriate system clock, and outputs two copies of the signal, one digital (26) and one analog (28). The EASU comprises a GPS receiver which is connected to an antenna 25. The GPS receiver allows the EASU to derive a master clock signal and synchronize that signal to the exciter's clock by using a GPS unit. The EASU provides the main system clock used by the exporter. The EASU is also used to bypass (or redirect) the analog MPS audio from passing through the exporter in the event that the exporter has a catastrophic failure and is no longer operational. The bypassed audio 32 may be fed directly to the STL transmitter to remove dead-air events.

The STL transmitter 48 receives delayed analog MPS audio 50 and exciter link data 52. It outputs exciter link data and delayed analog MPS audio over STL link 14, which may be unidirectional or bidirectional. The STL link may be a digital microwave or ethernet link, for example, a standard user datagram protocol or a standard TCP/IP may be used.

The transmitter site includes an STL receiver 54, an exciter 56, and an analog exciter 60. The STL receiver 54 receives exciter link data, including audio and data signals as well as command and control messages, over the STL link 14. The exciter link data is transmitted to the exciter 56, and the exciter 56 generates an IBOC DAB waveform. The exciter includes a main processor, numerical up-converter, RF up-converter, and exciter engine (exgine) subsystem 58. The exciter engine accepts exciter link data and modulates the digital portion of the IBOC DAB waveform. The digital up-converter of the exciter 56 converts the sideband portion of the exciter engine output from digital to analog. The digital to analog conversion is based on a GPS clock, which is common to the GPS based clock of the exporter derived from the EASU. Thus, the exciter 56 comprises a GPS unit and an antenna 57. An alternative method for synchronizing the follower and exciter clocks can be found in U.S. patent application serial No. 11/081,267 (publication No. 2006/0209941a1), the disclosure of which is incorporated herein by reference. The RF up-converter of the exciter up-converts the analog signal to the appropriate in-band channel frequency. The upconverted signal is then passed to a high power amplifier 62 and an antenna 64 for broadcast. In an AM transmission system, the exciter engine subsystem inherently adds backup analog MPS audio to the digital waveform in the hybrid mode; therefore, the AM transmission system does not include the analog exciter 60. In addition, the exciter 56 generates phase and amplitude information, and the analog signal is directly output to the high power amplifier.

IBOC DAB signals can be transmitted in the AM and FM radio bands using a wide variety of waveforms. The waveforms include an FM hybrid IBOC DAB waveform, an FM all-digital IBOC DAB waveform, an AM hybrid IBOC DAB waveform, and an AM all-digital IBOC DAB waveform.

Fig. 2 is a schematic diagram of a hybrid FM IBOC waveform 70. The waveform includes an analog modulated signal 72 at the center of a broadcast channel 74, a first plurality of evenly spaced orthogonal frequency division multiplexing subcarriers 76 in an upper sideband 78, and a second plurality of evenly spaced orthogonal frequency division multiplexing subcarriers 80 in a lower sideband 82. The digitally modulated subcarriers are divided into blocks and the various subcarriers are referred to as reference subcarriers. A frequency partition is a set of 19 OFDM subcarriers containing 18 data subcarriers and one reference subcarrier.

The hybrid waveform includes an analog FM modulated signal plus digitally modulated primary subcarriers. The subcarriers are at evenly spaced frequency locations. The subcarrier locations are numbered from-546 to + 546. In the waveform of fig. 2, the subcarriers are at locations +356 to +546, and-356 to-546. Each elementary primary sideband is composed of 10 frequency blocks. Subcarriers 546 and 546, also included in the primary sidebands, are additional reference subcarriers. The amplitude of each subcarrier may be scaled by an amplitude scaling factor.

Fig. 3 is a schematic diagram of an extended hybrid FM IBOC waveform 90. An extended hybrid waveform is created by adding primary extended sidebands 92, 94 to primary sidebands present in the hybrid waveform. One, two, or four frequency bins may be added to the inner edge of each primary sideband. The extended hybrid waveform includes the analog FM signal plus digitally modulated primary subcarriers (subcarriers +356 to +546 and-356 to-546) and some or all of the primary extended subcarriers (subcarriers +280 to +355 and-280 to-355).

The upper primary extended sideband includes subcarriers 337 through 355 (one frequency partition), 318 through 355 (two frequency partitions), or 280 through 355 (four frequency partitions). The lower primary extension sideband includes subcarriers-337 through-355 (one frequency partition), -318 through-355 (two frequency partitions), or-280 through-355 (four frequency partitions). The amplitude of each subcarrier may be scaled by an amplitude scaling factor.

FIG. 4 is a schematic diagram of an all-digital FM IBOC waveform 100. The all-digital waveform is constructed by disabling the analog signal, substantially extending the bandwidth of the primary digital sidebands 102, 104, and adding lower power secondary sidebands 106, 108 in the spectrum not occupied by the analog signal. The all-digital waveform in the embodiment shown includes digitally modulated subcarriers at subcarrier locations-546 to +546, without an analog FM signal.

In addition to the 10 primary frequency partitions, there are all 4 extended frequency partitions in each of the primary sidebands of the all-digital waveform. Each secondary sideband also has 10 secondary primary (SM) frequency partitions and four secondary extended (SX) frequency partitions. However, unlike the primary sidebands, the secondary primary frequency partitions are mapped closer to the channel center, while the extended frequency partitions are farther from the center.

Each secondary sideband also supports a small Secondary Protected (SP) region 110, 112, which includes 12 OFDM and reference subcarriers 279 and-279. Sidebands are referred to as "protected" because they are in regions of the spectrum that are less likely to be affected by analog or digital interference. The additional reference sub-carrier is placed in the center of channel (0). The frequency bin ordering of the SP region is not applied because the SP region does not contain frequency bins.

Each secondary primary sideband covers subcarriers 1 through 190, or-1 through-190. The upper secondary extended sideband includes subcarriers 191 through 266 and the upper secondary protected sideband includes subcarriers 267 through 278, plus additional reference subcarrier 279. The lower auxiliary extended sideband includes subcarriers-191 through-266 and the lower auxiliary protected sideband includes subcarriers-267 through-278, plus an additional reference subcarrier-279. The full frequency range of the entire full digital spectrum is 396,803 Hz. The amplitude of each subcarrier may be scaled by an amplitude scaling factor and the auxiliary sideband amplitude scaling factor may be user selectable. Any of the four may be selected for application to the auxiliary sidebands.

In each waveform, the digital signal is modulated by using Orthogonal Frequency Division Multiplexing (OFDM). OFDM is a parallel modulation scheme in which a data stream modulates a large number of orthogonal subcarriers that are transmitted simultaneously. OFDM is inherently flexible, easily allowing logical channels to be mapped to different groups of subcarriers.

In the hybrid waveform, the digital signal is transmitted in the form of a hybrid waveform on the fundamental Primary (PM) sideband on either side of the analog FM signal. The power level of each sideband is slightly lower than the total power in the analog FM signal. The analog signal may be mono or stereo and may include an auxiliary Communication automation (SCA) channel.

In an extended hybrid waveform, the bandwidth of the hybrid sidebands may be extended toward the analog FM signal to increase digital capacity. This additional spectrum allocated to the inner edge of each primary sideband is referred to as the primary extended (PX) sideband.

In all-digital waveforms, the analog signal is removed and the bandwidth of the fundamental digital sideband is fully extended as in the extended hybrid waveform. In addition, this waveform allows the lower power digital auxiliary sidebands to be transmitted in the spectrum not occupied by the analog FM signal.

Fig. 5 is a schematic of an AM hybrid IBOC DAB waveform 120. The hybrid format includes a conventional AM analog signal 122 (band limited to about ± 5kHz) along with a DAB signal 124 that is nearly 30kHz wide. The spectrum is contained within a frequency channel 126 having a bandwidth of about 30 kHz. The frequency channel is divided into an upper frequency band 130 and a lower frequency band 132. The upper band extends from the center frequency of the channel to about +15kHz from the center frequency. The lower band extends from this center frequency to about-15 kHz from the center frequency.

The AM hybrid IBOC DAB signal format in one example comprises the analog modulated carrier signal 134 plus OFDM subcarrier locations spanning the upper and lower frequency bands. Encoded digital information representing the audio or data signal (program material) to be transmitted is transmitted on the sub-carriers. The symbol rate is less than the subcarrier spacing due to the guard time between symbols.

As shown in fig. 5, the upper frequency band is divided into a primary portion 136, a secondary portion 138, and a third portion 144. The lower frequency band is divided into a primary portion 140, a secondary portion 142, and a third portion 143. For the purposes of this description, the third portions 143 and 144 may be viewed as including groups of subcarriers referenced 146, 148, 150, and 152 on fig. 5. The subcarriers located near the center of the channel in the third section are referred to as inner subcarriers, and the subcarriers located farther from the center of the channel in the third section are referred to as outer subcarriers. In this example, the power levels of the inner subcarriers in the groups 148 and 150 are shown to decrease linearly with frequency spaced from the center frequency. The remaining subcarrier groups 142 and 152 in the third section have a substantially constant power level. Fig. 5 also shows two reference subcarriers 154 and 156 for system control, whose levels are fixed at different values from the other sidebands.

The power of the subcarrier in the digital sidebands is significantly lower than the total power in the analog AM signal. The level of each OFDM subcarrier within a given primary or secondary portion is fixed at a constant value. The primary or secondary portions may be scaled relative to each other. In addition, status and control information is sent on reference subcarriers located on either side of the main carrier. Separate logical channels, such as an IBOC Data Service (IDS) channel, may be transmitted in the individual subcarriers just above and below the frequency edges of the upper and lower secondary sidebands. The power level of each basic OFDM sub-carrier is fixed relative to the unmodulated primary analog carrier. However, the power levels of the secondary subcarriers, the logical channel subcarriers, and the third subcarriers are adjustable.

Using the modulation format of fig. 5, the analog modulated carrier and the digitally modulated subcarriers are transmitted within the channel mask specified for standard AM broadcasting in the united states. The hybrid system uses analog AM signals for tuning and backup.

Figure 6 is a schematic diagram of subcarrier allocation for an all-digital AM IBOC DAB waveform. The all-digital AM IBOC DAB signal 160 includes first and second groups 162 and 164 of evenly spaced subcarriers, referred to as base subcarriers, that are placed in an upper band 166 and a lower band 168. Third and fourth sets 170 and 172 of subcarriers, referred to as secondary subcarriers and third subcarriers, respectively, are also placed in upper band 166 and lower band 168. The third set of two reference subcarriers 174 and 176 is located closest to the center of the frequency channel. Subcarriers 178 and 180 may be used to transmit program information data.

FIG. 7 is a simplified functional block diagram of an AM IBOC DAB receiver 200. The receiver includes an input 202 connected to an antenna 204, a tuner or front end 206, and a digital down-converter 208 for producing a baseband signal on line 210. Analog demodulator 212 demodulates the analog modulated portion of the baseband signal to produce an analog audio signal on line 214. The digital demodulator 216 demodulates the digitally modulated portion of the baseband signal. The digital signal is then deinterleaved by a deinterleaver 218 and decoded by a Viterbi decoder 220. Service demultiplexer 222 separates the main and supplemental program signals from the data signal. The processor 224 processes the program signal to produce a digital audio signal on line 226. The analog and primary digital audio signals are mixed as shown in block 228 or a supplemental digital audio signal is transmitted to produce an audio output on line 230. A data processor 232 processes the data signals and generates data output signals on lines 234,236 and 238. The data signals may include, for example, Station Information Service (SIS), Main Program Service Data (MPSD), Supplemental Program Service Data (SPSD), and one or more Auxiliary Application Services (AAS).

FIG. 8 is a simplified functional block diagram of an FM IBOC DAB receiver 250. The receiver includes an input 252 connected to an antenna 254, and a tuner or front end 256. The received signal is provided to an analog-to-digital converter and digital down converter 258 to produce a baseband signal at output 260 that includes a series of complex signal samples. The signal samples are complex in that each sample includes a "real" component and an "imaginary" component, the imaginary component of which is sampled orthogonally to the real component. Analog demodulator 262 demodulates the analog modulated portion of the baseband signal to produce an analog audio signal on line 264. The digitally modulated portion of the sampled baseband signal is then filtered by a sideband isolation filter 266 having a bandpass frequency response that includes the subcarriers f present in the received OFDM signal₁-f_nA collection of (a). The filter 268 suppresses the effect of the first neighboring interferer. The complex signal 298 is routed to an input of an acquisition module 296, and the acquisition module 296 acquires or recovers the OFDM symbol timing offset or error, and the carrier frequency offset or error, from the received OFDM symbols as represented in the received complex signal 298. The acquisition module 296 forms the symbol timing offset Δ t and the carrier frequency offset Δ f as well as status and control information. The signal is then demodulated (block 272) to demodulate the digitally modulated portion of the baseband signal. Then, the digital signal is deinterleaved by a deinterleaver 274 and decoded by a Viterbi decoder 276. Service demultiplexer 278 separates the main and supplemental program signals from the data signal. Processor 280 processes the main and supplemental program signals to produce a digital audio signal on line 282. The analog and primary digital audio signals are mixed as shown in block 284 or a supplemental program signal is transmitted to produce an audio output on line 286. A data processor 288 processes the data signals and generates the signals on lines 290, 292 and 294The data output signal of (2). The data signals may include, for example, Station Information Service (SIS), Main Program Service Data (MPSD), Supplemental Program Service Data (SPSD), and one or more Advanced Application Services (AAS).

In practice, many of the signal processing functions shown in the receivers of fig. 7 and 8 may be implemented using one or more integrated circuits.

FIGS. 9a and 9b are diagrams of the IBOC DAB logical protocol stack from the transmitter side. From the receiver perspective, the logical stack will be traversed in the opposite direction. Most of the data transferred between the various entities within the protocol stack is in the form of Protocol Data Units (PDUs). A PDU is a structured data block that is generated by a particular layer (or process within a layer) of the protocol stack. A PDU of a given layer may encapsulate PDUs from the next higher layer of the stack and/or include content data and protocol control information originating in the layer (or process) itself. The PDUs generated by each layer (or process) in the transmitter protocol stack are input to the corresponding layer (or process) in the receiver protocol stack.

As shown in fig. 9a and 9b, there is a configuration manager 330, as a system function, that provides configuration and control information to various entities within the protocol stack. Configuration/control information may include user-specified settings, as well as information generated from within the system, such as GPS time and location. Service interface 331 represents an interface for all services except SIS. The service interface may be different for each of the various types of services. For example, for MPS audio and SPS audio, the service interface may be an audio card. For MPS data and SPS data, the interfaces may have different forms of Application Program Interfaces (APIs). For all other data services, the interface has the form of a single API. The audio codec 332 encodes the MPS audio and SPS audio to produce a core stream (stream 0) and an optional enhancement stream (stream 1) of MPS and SPS audio encoded packets, which are passed to an audio transport 333. The audio codec 332 also relays unused capacity status to other parts of the system, thus allowing for the inclusion of opportunistic data. MPS and SPS data are processed by Program Service Data (PSD) transport 334 to produce MPS and SPS data PDUs, which are passed to audio transport 333. Audio transport port 333 receives encoded audio packets and PSD PDUs and outputs a bitstream containing compressed audio and program service data. SIS transport 335 receives SIS data from the configuration manager and generates SIS PDUs. The SIS PDU may contain station identification and location information, program type, and absolute time and location associated with GPS. The AAS data transport 336 receives AAS data from the service interface and opportunistic bandwidth data from the audio transport and generates AAS data PDUs, which may be based on the quality of the service parameters. The transport and encoding functions are collectively referred to as layer 4 of the protocol stack, and the corresponding transport PDU is referred to as a layer 4PDU, or L4 PDU. Layer 2, which is a channel multiplex layer (337), receives transport PDUs from the SIS transport means, AAS data transport means, and audio transport means and formats them into layer 2 PDUs. The layer 2 PDU includes protocol control information and a payload, which may be audio, data, or a combination of audio and data. The layer 2 PDUs are routed to layer 1 (338) through the correct logical channel, which is the signal path that conducts the L1 PDUs through layer 1 at the specified level of service. There are a plurality of layer 1 logical channels according to a service mode, wherein a service mode is a specific configuration of operational parameters that specify throughput, performance levels, and selected logical channels. The number of active layer 1 logical channels and the characteristics defining them vary with each service mode. Status information is also transferred between layer 1 and layer 2. Layer 1 converts the PDUs and system control information from layer 2 into AM or FM IBOC DAB waveforms for transmission. Layer 1 processing may include scrambling, channel coding, interleaving, OFDM subcarrier mapping, and OFDM signal generation. The output of the OFDM signal generation is a complex baseband time domain pulse that represents the digital portion of the IBOC signal for a particular symbol. The discrete symbols are concatenated to form a continuous time domain waveform, which is in turn modulated to create an IBOC waveform for transmission.

Fig. 10 shows the logical protocol stack from the receiver side. The IBOC waveform is received (560) by the physical layer, layer 1, which demodulates the signal and processes it to separate the signal into logical channels. The number and kind of logical channels will depend on the service mode and may include logical channels P1-P3, PIDs, S1-S5, and SIDs. Layer 1 generates L1 PDUs corresponding to the logical channel and sends the PDUs to layer 2 (565), which layer 2 demultiplexes the L1 PDUs to generate SIS PDUs, AAS PDUs, PSD PDUs for the primary program service and any supplemental program services, as well as stream 0 (core) audio PDUs and stream 1 (optional enhancement) audio PDUs. The SIS PDUs are then processed by SIS transport 570 to produce SIS data, AAS PDUs are processed by AAS transport 575 to produce AAS data, and PSD PDUs are processed by PSD transport 580 to produce MPS data (MPSD) and any SPS data (SPSD). The SIS data, AAS data, MPSD, and SPSD are then sent to the user interface 590. The SIS data, if desired by the user, may be displayed. Likewise, MPSD, SPSD, and any text-based or graphics-based AAS data may be displayed. Stream 0 and stream 1 PDUs are processed through layer 4, which consists of audio transport 590 and audio decoder 595. There may be up to N audio transports corresponding to the number of programs received on the IBOC waveform. Each audio transport generates encoded MPS packets or SPS packets corresponding to each received program. Layer 4 receives control information from the user interface including commands such as storing or playing programs and finding or searching for radio stations broadcasting all-digital or hybrid IBOC signals. Layer 4 also provides status information to the user interface.

As previously described, the digital portion of the IBOC signal is modulated using Orthogonal Frequency Division Multiplexing (OFDM). Referring to fig. 11a, the OFDM signal used in the present invention is characterized by including a plurality of equidistantly spaced subcarriers f₁-f_nA multi-frequency carrier signal. Adjacent subcarriers, such as f₁And f₂Are separated from each other by a predetermined frequency increment such that adjacent subcarriers are orthogonal to each other. By orthogonal, it is meant that the subcarriers do not exhibit crosstalk when properly nyquist weighted. In a hybrid system incorporating the present invention and using digital and analog transmission channels, there is a 70kHz bandwidth for each sidebandIn each sideband, there are 191 carriers. In an all-digital embodiment of the present invention, there are 267 carriers in each sideband with a 97kHz bandwidth for each sideband.

Fig. 11b shows an OFDM symbol 5 in the time domain. The symbols having an effective symbol period or time width T, and a full symbol period T_α. OFDM subcarrier orthogonality requires the generation of a functional interdependence between the effective symbol period T and the frequency of the spacing between adjacent OFDM subcarriers. Specifically, the frequency interval between adjacent subcarriers is limited to be equivalent to the inverse of the effective symbol period of each OFDM symbol 5. That is, the frequency spacing is equal to 1/T. Spanning the effective symbol period T of each OFDM symbol 5 is a predetermined number N of equidistantly spaced time symbol samples (not shown). Moreover, the entire period T of each OFDM symbol 5 is spanned_αIs a predetermined number N_αN (1+ α) equally spaced time symbol samples, α is the amplitude tapering (tapering) factor of the symbol and may be considered herein as a fractional multiplier. During modulation, the OFDM modulator generates a series of OFDM symbols 5, each symbol comprising a symbol period T corresponding to the whole symbol_αPredetermined number N of_αWherein the first alphan samples and the last alphan samples of each symbol are tapered (tapered) and have equal phase. In one embodiment, each full symbol period T is spanned_αA predetermined number N of time samples_αIs 1080, the predetermined number N of time samples spanning each valid symbol period T is 1024, and the number of samples in each of the first alphan samples and the last alphan samples is 56. These values are merely exemplary and may vary depending on system requirements. In addition, during modulation, a cyclic prefix is added so that the beginning and end of each transmitted symbol are highly correlated.

A predetermined amplitude-time profile or envelope 11, 15, 13 is applied to the signal levels of these samples. This amplitude distribution comprises symmetrically increasing or decreasing amplitude tapers 11, 15 at the beginning and end, respectively, of each symbol 5, and at themA flat amplitude distribution 13 extending in between. These rounded or tapered edges provided in the time domain serve to greatly reduce unwanted side lobe energy in the frequency domain, thus providing a more spectrally efficient OFDM signal. Although the entire symbol period T of the symbol 5_αExtending outside the effective symbol period T but the orthogonality between adjacent sub-carriers in the frequency domain (fig. 11a) is not compromised as long as the amplitude taper 11, 15 of the symbol 5 follows the nyquist or raised cosine taper function. More specifically, orthogonality is maintained in the present invention by a combination of root-raised cosine weighting (or amplitude tapering) of the transmitted symbols and root-raised cosine matched filtering of the received symbols.

The beginning and ending portions of an OFDM symbol 5 share the additional important feature that the first alphan OFDM symbol samples across the beginning portion of the OFDM symbol 5 have a phase substantially equal to the last alphan symbol samples across the ending portion of the OFDM symbol 5, where the beginning portion of the OFDM symbol 5 has a duration at and the ending portion of the OFDM symbol 5 also has a duration at. Again, α is the amplitude tapering factor of the symbol and may be considered herein as a fractional multiplier.

Acquisition module structure and operation

One embodiment of the basic acquisition module 296 described in U.S. patent nos. 6,539,063 and 6,891,898 is shown in fig. 12. The received complex signal 298 is provided to the input of a peak forming module 1100. the peak forming module 1100 provides a first stage of signal processing for obtaining the symbol timing offset of the received OFDM signal. The peak forming module 1100 forms a boundary signal 1300 at its output, with a plurality of signal peaks in the boundary signal 1300, each signal peak representing a received symbol boundary position of each received OFDM symbol represented in the received signal 298 input into the peak forming module 1100. Because these signal peaks represent received symbol boundary positions, their time positions indicate received symbol timing offsets. More specifically, because the receiver has no initial or a priori knowledge of the true or actual received symbol boundary positions, such positions also begin to be assumed, or are arbitrarily created to enable receiver processing to function. The acquisition module 296 determines the symbol timing offset Δ t that exists between this a priori assumption and the true received symbol boundary position, thus enabling the receiver to recover and track the symbol timing.

In forming the signal peaks representing the OFDM symbol boundaries, peak forming module 1100 utilizes the cyclic prefix applied by the transmitter, as well as the predetermined amplitude taper and phase characteristics inherent in the beginning and ending portions of each received OFDM symbol. In particular, a complex conjugate product between the current sample and the sample N samples before it is formed. This product formed between the first alphan samples and the last alphan samples in each symbol produces a signal peak corresponding to each symbol that includes the alphan conjugate products so formed.

Mathematically, the formation of the conjugate product is represented as follows. Let D (T) denote the received OFDM signal, and let T_αT denotes the total OFDM symbol duration or period, where 1/T is the OFDM channel spacing and α is the amplitude tapering factor of the symbol. The signal peaks in the boundary signal 1300 appear as a series of pulses or signal peaks in the conjugate product of D (T). D (T-T). As a result of the nyquist amplitude tapering applied to the beginning and end portions of each OFDM symbol, each pulse or signal peak has a half sine wave amplitude distribution of the form: w (t) ═ last¹/₂sin (π T/(α T)), for 0 ≦ T ≦ α T, w (T) ≦ 0, and others.

Also, the periodicity of the signal 1300, i.e., the period of the signal peak train, is T_α. Referring to fig. 11c, a train of signal peaks included in the boundary signal 1300 has an amplitude envelope w (T) and the peaks are spaced apart by T_αThe time interval of (c). Referring to fig. 11d, the product of the overlapping beginning and ending amplitude tapers 11, 15 multiplied by the squared amplitude in the conjugate product results in a half sine wave w (T) having a duration width α T corresponding to α N samples.

Referring again to fig. 12, for each signal sample input to peak forming module 1100, a product sample is output from multiplier circuit 1250 that represents the conjugate product between the input sample and a predecessor sample that is T samples apart from the input sample. The complex conjugate former 1200 produces at its output the complex conjugate of each input sample, which is provided as one input to the multiplier 1250. The conjugate sample at this output is multiplied by the delayed sample output from delay circuit 1150. Thus, a complex conjugate product is formed between the received signal 298 and its delayed replica, which is obtained by delaying the received signal 298 for a predetermined time T using the delay circuit 1150.

Referring to fig. 13a, 13b and 13c, the correlation symbol timing for the peak forming module 1100 is shown. Fig. 13a shows successive OFDM symbols 1 and 2 provided at the input of the peak forming module 1100. Fig. 13b shows a delayed version of OFDM symbols 1 and 2 as output from delay circuit 1150. FIG. 13c shows a graph for N_αEach corresponding set of N (1+ α) product samples (which in one possible embodiment is equal to 1080 samples) forms a signal peak, the series of signal peaks being generated in response to a conjugate multiplication between the received signal of fig. 13a and its delayed version on fig. 13 b.

As a specific example, if the received OFDM symbol period T_αCorresponding to N_α1080 signal samples and an N samples at each of the beginning and end of the symbol correspond to 56 signal samples, then for each 1080 sample OFDM symbol input to peak forming module 1100, a corresponding set of 1080 product samples will appear in boundary signal 1300. In this example, the delay circuit 1150 gives 1024(N) the sample delay so that each sample input to the multiplier 1250 is multiplied by its predecessor sample that is 1024 samples away. The signal peak so formed for each respective group of 1080 product samples includes only the 56 conjugate products formed between the first 56 and last 56 samples of each respective symbol.

The peak forming module 1100 can be implemented in any number of ways as long as the correspondence between the beginning and end of each symbol is utilized in the manner previously described. For example, peak forming module 1100 may operate on each sample as it arrives, such that for each incoming sample, a product sample is provided at its output. Alternatively, the plurality of samples may be stored, such as in the form of a vector, thus creating a current sample vector and a delayed sample vector, which may be input to a vector multiplier to form vector product samples at its output. Alternatively, the peak forming module may be implemented to operate on a continuous, rather than sampled, discrete-time signal. However, in such an approach, it is desirable that the incoming received signal 298 also be continuous rather than a sampled signal.

Ideally, the boundary signal 1300 has readily identifiable signal peaks therein, as shown in fig. 11c and 13 c. In practice, however, each signal peak is substantially indistinguishable from unwanted noise sample products located in adjacent symbols. As the peak forming module 1100 continues to form the product between the samples spanning each received symbol and the predecessor samples delayed therefrom, the boundary signal 1300 includes the desired signal peak and the noise conjugate product. For example, the first α N (56) samples in each symbol are multiplied by the last α N samples therein to produce the desired signal peak α N samples in duration. However, the remaining N (1024) samples are multiplied by N samples obtained by the neighboring sample in response to the delay imparted to it by the delay circuit 1150 (see fig. 13). These additional unwanted products have the effect of filling in noise between the occurrences of the wanted signal peaks. Thus, the noise product corresponding to the OFDM signal may be substantial.

In addition to the aforementioned product noise present in the boundary signal 1300, noise derived from other sources is well known in the digital communications arts. Such noise is imparted to the signal by ambient noise, scattering, multipath and fading, and signal interference during its propagation through the atmosphere. The front end of the receiver also adds noise to the signal.

The latter signal processing stages are partly dedicated to combat the aforementioned derogative influence of noise on the wanted signal peaks in the boundary signal 1300 or, more specifically, to improve the signal-to-noise ratio of the signal peaks present in the boundary signal 1300. A signal enhancement module 1350 is provided at the output of the peak forming module 1100 and includes first and second stage signal enhancement circuits or modules. The first stage signal enhancement circuit is a summing and superposition circuit or module 1400 and the second stage enhancement circuit is a matched filter 1450 provided at the output of the first stage signal enhancement circuit.

The addition-superposition circuit 1400 additively superposes a predetermined number of signal peaks and their surrounding noise products in order to enhance the detectability of the signal peaks by increasing the signal-to-noise ratio of the signal peaks in the boundary signal 1300. To implement this process of additive superposition, a predetermined number of consecutive segments of the boundary signal 1300 are first superimposed or overlapped in time. Each of these superimposed segments, when output from peak forming module 1100, includes a useful portion of the symbol period of the conjugate product sample and includes a desired signal peak surrounded by undesired noise product samples.

After a predetermined number or blocks of signal segments overlap in time, product samples occupying predetermined time positions in the superimposed segment groups are accumulated to form accumulated signal samples for the predetermined positions. In this way, for each predetermined sample position across the superimposed boundary signal segment, an accumulated signal is formed comprising accumulated signal samples.

If, for example, 32 adjacent boundary signal segments are to be superimposed and if each segment includes a useful portion of a symbol period of 1080 samples, the additive superposition circuit 1400 produces 1080 accumulated samples for each adjacent block of 32 segments (1080 samples per segment) that are input thereto. Thus, by adding the superimposed conjugate products of the 32 segments point-by-point, the conjugate products of the 32 segments (each segment comprising 1080 samples, a signal peak, and noise therein) are additively superimposed or "folded" on top of one another. Basically, in this folding process, over 32 adjacent symbols, the product of 32 segments is added point-by-point to the corresponding conjugate product that is further away from one symbol period (or 1080 samples) to produce an accumulated signal segment that includes 1080 accumulated samples. The signal processing process is then repeated for the next adjacent block of 32 boundary signal segments to produce another accumulated signal segment, and so on.

The accumulated signal segments produced by additively superimposing a predetermined number of adjacent segments of the boundary signal 1300 include enhanced signal peaks therein, which exhibit enhanced signal-to-noise ratios for the signal peaks in each of the constituent input boundary signal segments. The reason for this enhancement is that the superposition of the boundary signal segments aligns their respective signal peaks so that, as the segments are accumulated, each signal peak adds to the next signal peak, thus achieving a form of coherent processing gain based on the repeating nature of the boundary signal peaks.

While the aligned repeated signal peaks in the boundary signal segments coherently accumulate to form an enhanced (accumulated) signal peak at the output of the additive superposition module 1400, in contrast, the random nature of the noise conjugate product around the signal peak in each boundary signal segment produces an incoherent addition during the additive superposition process. Because the signal peaks add coherently and the surrounding noise products with zero mean add non-coherently, and are thus averaged, the enhanced signal peaks output from the additive superposition module 1400 exhibit an improved signal-to-noise ratio overall. The processing gain and signal to noise ratio enhancement resulting from the additive superposition module is increased along with the number of boundary signal segments being superposed, resulting in an accumulated signal segment. Counteracting this advantage is a corresponding disadvantageous increase in acquisition delay, since more boundary signal segments are collected to produce accumulated signal peaks. Thus, a certain predetermined number, e.g. 16 or 32, represents a balance between these two competing points of interest in any application, wherein the number on average is ultimately limited by the fading bandwidth.

In mathematical terms, at boundariesThe additive superposition of adjacent segments of the conjugate product present in signal 1300 may be expressed by:where K is the number of superimposed segments, D is the input 298 to the peak forming module 1100, and K is the number of segments, such as, for example, 16. An important aspect of the above-described signal processing is that the symbol timing is preserved at each stage thereof: an OFDM symbol input to the peak forming module 1100, a boundary signal segment input to the add-and-add module 1400, and an accumulated signal segment output therefrom, each having T_αTime period (corresponding to N1080 samples). In this way, the symbol timing offset is preserved throughout the process, as indicated by the position of the signal peak within the signal segment.

In operation, the add-add module 1400, the summing module 1600, and the feedback delay module 1650 together provide an add-add function. That is, the summing block 1600 adds the now input samples to the accumulated results of samples in adjacent symbols, each sample being spaced apart in time by one symbol period T_α(corresponding to 1080 samples). The delay module 1650 gives a one symbol period delay between accumulations. In other words, each accumulation result output by the summation module 1600 is delayed by 1 symbol periodT_αAnd then fed as input to the summation module 1600 where it is added to the next input sample. This process is repeated for all input samples on each input symbol.

In other words, the first accumulated sample in the accumulated signal segment represents the accumulation of all the first samples of all 32 boundary signal segments. In the accumulated signal segment, the second accumulated sample represents the accumulation of all second samples of all 32 boundary signal segments, and so on.

After a predetermined number of signal segments have been accumulated to produce an accumulated signal segment, reset generator 1700 provides a reset signal to delay module 1650. For example, if the predetermined number of boundary signal segments to be accumulated is 32, the reset generator 1700 issues a reset to the feedback delay module 1650 every 32 signal segments. In response to the issuance of the reset, the additive superposition module 1400 accumulates the next predetermined number of adjacent boundary signal segments.

As previously described, the output of the additive superposition module 1400 is an accumulated signal comprising a series of accumulated signal segments, each segment including an enhanced signal peak 1550 therein. In a high noise environment, the enhanced signal peak 1500, while exhibiting an improved signal-to-noise ratio, still is not practically distinguishable from surrounding noise. Therefore, it is desirable to further enhance the signal-to-noise ratio of the enhanced signal peak.

To further enhance the signal-to-noise ratio of enhanced signal peak 1550, the accumulated signal output from add-add module 1400 is input to matched filter 1450. The time impulse response of matched filter 1450 is matched to the shape or amplitude envelope of the enhanced signal peak input thereto and, in one embodiment of the present invention, follows a root raised cosine distribution. Specifically, the impulse response of the matched filter corresponds to the function w (t) shown in fig. 11d, which is determined by pointwise multiplying the first alphan samples of symbol 5 by the last alphan samples of symbol 5. See fig. 11b and 11 d.

Although a non-matched low pass filter may be used to smooth out the noise present in the accumulated signal, matched filter 1450 provides an optimal signal-to-noise improvement over the desired signal, i.e. enhanced signal peak 1500, in a gaussian noise environment. The matched filter 1450 is implemented as a Finite Impulse Response (FIR) digital filter which provides at its output a filtered version of the complex samples input thereto.

Summarizing the signal processing stages briefly from the outset up to the output of the matched filter, peak forming module 1100 generates a plurality of signal peaks whose time positions represent symbol boundary positions that represent the symbol timing offset for each received OFDM symbol. The signal enhancement module 1350 enhances the detectability of signal peaks by first additively superimposing a predetermined number of input signal segments to produce accumulated signal segments having enhanced peaks therein, and then, second, match filtering the accumulated signal segments to produce accumulated, match-filtered signal segments that are optimally ready for subsequent peak detection processing. This process is continuously running to produce a plurality of filtered enhanced signal peaks at the output of the signal enhancement module 1350. The time positions of these filtered enhanced signal segments within the matched filtered, accumulated signal segments output from signal enhancement module 1350 represent symbol boundary positions or OFDM symbol timing offsets.

Viewed individually, and particularly in combination, the additive superposition module and the matched filter advantageously enhance signal peak detectability. Their introduction after the peak forming stage allows for efficient use of OFDM signals which comprise a large number of frequency carriers and which operate in a propagating noise signal environment.

The next stage of signal processing required to determine the symbol timing offset is to detect the time position of the signal peak output from the signal enhancement module 1350. The time position of a signal peak is, in effect, the sample index or sample number of the enhanced signal peak within the filtered accumulated signal segment output from the matched filter.

The filtered complex signal 1750 output from the matched filter 1450 is provided as an input to a peak selector module 1900, the peak selector module 1900 detecting the enhanced filtered signal peak and its time position or sample index. In operation, the squared magnitude generator 1950 of the peak selector 1900 squares the magnitude of the complex signal samples input thereto to generate a signal waveform at its output. The output of the squared magnitude generator 1950 is provided as an input to a maximum finder 2000, which examines the magnitude of the samples input thereto and identifies the time position or sample index corresponding to the signal peak.

This time position of the signal peak is actually provided as a symbol timing offset, which is provided by the acquisition module 296 to the input of a symbol timing correction module (not shown). It will be appreciated that the time position provided as the timing offset Δ t may need to be adjusted slightly to compensate for various processing delays introduced by previous signal processing stages. The initialization delay while loading the filter, etc. may increase the delay that needs to be calibrated out of the last timing offset estimate. However, such delays are typically small and implementation specific.

After the time position of the signal peak has been determined (in order to determine the symbol timing offset), the next stage in the signal processing is to determine the carrier phase error and the corresponding carrier frequency error of the received OFDM signal. The matched filtered, enhanced signal peaks in complex signal 1750 represent the cleanest points or points where the signal-to-noise ratio is greatest, at which the carrier phase error and frequency error are determined. The phase of the complex samples at this peak position gives an indication of the frequency error that exists between the transmitter and the receiver, since the conjugate product at this point, as formed by peak forming module 1100, should produce a zero phase value in the absence of carrier frequency error. The conjugate product at this point of the signal peak, and in fact at every other point in the signal peak, should yield a zero phase value because mathematically, the conjugate product between the symbol samples with equivalent phase (as are the samples at the beginning and end of each received symbol) cancels the phase when there is no carrier frequency error. Any residual phase present at the peak of the signal output from the matched filter is proportional to the carrier frequency error, so once the residual phase is determined, it is simple to calculate the frequency error.

Mathematically, the carrier frequency error Δ f produces a residual phase shift 2 π Δ fT between the samples at the beginning and end portions of the OFDM symbol that form the conjugate product peak. Thus, the frequency error is represented by:wherein G is_maxIs the peak of the matched filter output and Arg represents the argument (phase) of the complex-complex sample-at the signal peak. The Arg function is equivalent to a four quadrant arctangent. Since the arctangent cannot detect angles outside the 2 pi window, the frequency estimate is ambiguous until the channel is spaced by a multiple of 1/T. In any event, this frequency error estimate, along with the timing offset estimate provided by the position of the signal peak, is sufficient to allow symbol demodulation to begin. Subsequent receiver frame boundary processing-but not part of the invention-resolves the frequency ambiguity when performing demodulation.

In fig. 12, the match filtered complex signal 1750 and the time position or sample index are provided as inputs to the phase extractor 2050. Phase extractor 2050 extracts the residual phase from the complex samples representing the enhanced signal peak output from the matched filter. The extracted phase is provided to an input of a frequency generator 2100 that merely scales the extracted phase input thereto to produce a carrier frequency error Δ f, which is then provided by an acquisition module 296 to a frequency correction module (not shown). Thus, the time position of the filtered signal peak provided at the output of matched filter 1450 represents the symbol timing offset and the carrier frequency error is derived from the phase of this signal peak.

FM digital signal quality metric

The above-described methods and apparatus for deriving or recovering symbol timing offset and carrier frequency error from a received OFDM signal provide a basic technique for determining failing symbol timing offset and carrier frequency error. U.S. patent No.6,539,0636,891,898 describes additional techniques for deriving or recovering symbol timing offset and carrier frequency error from a received OFDM signal, any of which may be used to implement a digital signal quality metric in accordance with the present invention. Because the acquisition function described in these patents is time domain processing that occurs near the beginning of the baseband processing chain and prior to OFDM demodulation, it can be exploited to provide an effective digital signal quality metric.

Moreover, the predetermined amplitude and phase characteristics described above and inherent in the beginning and ending portions of the OFDM symbols, i.e., the tapering of the sample amplitudes in the beginning and ending portions of each OFDM symbol and their equivalent phases, are advantageously utilized by existing IBOC systems to efficiently acquire OFDM symbol timing and frequency in the receiver. These features may be used in accordance with the present invention to implement a digital signal quality metric. Thus, in one aspect, the present invention utilizes these symbol characteristics to provide a digital signal quality metric by using a preexisting FM acquisition module.

Preferably, the acquisition algorithm used for the digital signal quality metric consists of two operations: pre-acquisition filtering and acquisition processing. Pre-acquisition filtering is used to prevent spurious acquisitions on a large second adjacent channel (second adjacent channel). Each primary sideband is filtered prior to acquisition processing. In one example, the pre-acquisition filter is an 85-tap Finite Impulse Response (FIR) filter designed to provide 40dB band rejection while limiting its effect on the desired fundamental sideband. Existing pre-acquisition filters can be fully reused without modification when computing the quality metric of the present invention. After the input samples are filtered, they are passed to the acquisition processing function.

The acquisition processing function uses intra-symbol correlations resulting from the transmitter applying a cyclic prefix to each symbol to construct an acquisition peak. As previously described, the peak position indicates the position of the true symbol boundary within the input sample, while the phase of the peak is used to derive the frequency error. Furthermore, frequency diversity can be obtained by independently processing the upper and lower fundamental sidebands of the digital radio signal.

Each symbol comprises a plurality of samples. The inputs to the acquisition process are blocks of upper and lower fundamental sideband samples. In one example, each block consists of 940 real or imaginary samples having a rate of 372,093.75 samples per second.

The acquisition algorithm modified for calculating the digital signal quality metric is shown in fig. 14 and 19. Referring first to fig. 14, 940-sample filtered data blocks are buffered as 1080-sample symbols, as shown in block 370. As previously described, the first 56 and last 56 samples of each transmitted symbol are highly correlated due to the cyclic prefix. The acquisition process reveals this correlation by complex conjugate multiplying each sample in an arbitrary symbol with its predecessor that is 1024 samples away (block 372). To enhance the detectability of the resulting 56-sample peak, the corresponding products of the 16 adjacent symbols are folded on top of one another to form a 1080-sample acquisition block (block 374). In this embodiment, sixteen symbols are used instead of the 32 symbols as described with reference to the previously described acquisition method in order to speed up the calculation of the digital signal quality metric, but fewer symbols, such as 8, may be desirable, and any other suitable number of symbols may be used.

56-sample fold peaks, while visible within the acquisition block, are very noisy. Therefore, block 376 shows that it has been smoothed with a 57-tap FIR filter whose impulse response matches the shape of the peak:for n0, 1.., 1079 where n is the output sample index, x is the matched filter input, y is the matched filter output, and h [ k [, j]Is the filter impulse response, expressed as follows:for k 0, 1

Taking the magnitude squared of the matched filtered output values (block 378), symbol boundary detection may be simplified by converting the complex values to real values. This calculation increases the dynamic range of the input, making the symbol boundary peaks less ambiguous and allowing peak searching in one dimension (two dimensions for I and Q values). The magnitude squared calculation is: y [ n ]]＝I[n]²+Q[n]²For n0, 1.., 1079 where I is the real part of the input, Q is the imaginary part of the input, y is the squared output of the magnitude, and n is the sample index. For each 16-symbol block, the upper and lower sideband match filtered, squared-amplitude output waveforms are used to generate a digital signal quality metric. The acquisition process continues as described above and the quality metric algorithm continues as shown in figure 19 (block 450), as shown in block 380.

The next step in the quality metric algorithm is to compute a normalized correlation peak (block 452- — 458) to obtain improved discrimination of the symbol boundary peaks. Normalizing the correlation peak provides a basis for assessing the quality of the signal and indicates the probability that a digital signal is present. The peak value of the normalized correlation peak may range from zero to one, with a value of 1 indicating the maximum likelihood that a digital signal is present. The peak value of the normalized correlation peak thus provides a digital signal quality metric.

A circuit according to a prior art acquisition algorithm for calculating the correlation peak is shown in block 382 of fig. 15. Input 384 is a 1080-sample symbol received on either the upper sideband or the lower sideband. The input samples are shifted by 1024 samples 386 and the complex conjugate 388 of the shifted samples are multiplied 390 with the input samples. The 16 symbols are folded as shown by block 392 and adder 394. The folded sum is filtered 396 through a root raised cosine matched filter and squared 398 to produce a correlation peak 399. Thus, the acquisition algorithm finds the symbol boundary by multiplying the current input sample with the complex conjugate of the input delayed by 1024 samples. At the beginning of the symbol, the phase of the conjugate product over the next 56 samples is substantially zero for each OFDM subcarrier. The constituent OFDM subcarriers are coherently combined over this time interval rather than over the remaining samples in the symbol. The result is a resolvable correlation peak 399 after the 16 symbols are folded and matched filtered.

Referring again to fig. 19, additional processing steps in accordance with the present invention are shown. The normalized correlation peak is determined by first calculating a normalized waveform for each of the upper and lower sideband waveforms (block 452). This normalized waveform utilizes the amplitude correlation between the first and last 56 samples of the OFDM symbol due to the root raised cosine pulse shaping applied at the transmitter. Referring to FIG. 15, block 400 illustrates the calculation of a normalized waveform 416. The magnitude squared 406 of each input symbol is delayed 386 by 1024 samples and added 404 to the current magnitude squared sample 402. The 16 symbols are shown by block 408 and adder 410And (4) folding. The folded sum is root raised cosine matched filtered 412 and squared sum reciprocal 414 to produce normalized waveform 416. The folding and matched filtering of the normalized waveform is equivalent to those steps performed in existing acquisition algorithms, except that the existing matched filter taps are squared and halved to ensure proper normalization:56 where k is the index of the tap in the matched filter, h k]Is the existing tap of the correlation peak of the conjugate multiplication, and g k]Is the new tap of the normalized waveform. After folding the first 16 symbols and performing matched filtering, the symbol boundaries are apparent. As shown in fig. 17, the location of the symbol boundary is marked by a reduction in the amplitude of the resulting waveform.

Referring again to fig. 19, once the normalized waveform is calculated, the next step is normalization of the correlation peak, block 458. Normalizing the correlation peak 399 with the normalization waveform from block 452 will enhance the correlation peak by reducing the level of all samples except those that coincide with symbol boundaries. Referring again to fig. 15, correlation peak 399 is multiplied 418 with normalized waveform 416 to produce normalized correlation peak 420. Fig. 18 shows an example of normalized correlation peaks in a fairly clean environment, where the x-axis represents the sample number and the y-axis is the normalized correlation value.

Once the correlation peak is normalized, the next step in the quality metric algorithm is to find the peak index P_UAnd P_LAnd peak value Q_UAnd Q_L(FIG. 19, block 460). The peak index is the sample number corresponding to the maximum value of the normalized correlation waveform. PA_UAnd P_LThe peak indices of the normalized correlation waveform for the upper and lower sidebands, respectively. Peak valueIs the maximum value of the normalized correlation waveform, which provides a digital signal quality metric.

The quality estimate from each sideband can be computed independently. The peak of the normalized correlation waveform represents the relative quality of the sidebands: q_U＝x(P_U)Q_L＝x(P_L) Where x is the normalized correlation waveform, Q_UIs the upper sideband quality, and Q_LIs the lower sideband quality. Referring to FIG. 15, a peak index 424 is identified and a peak quality value 422 is calculated for one sideband by block 426.

To validate the digital signal quality metric, optionally, peak index differences can be found and concealed. The peak index difference compares the peak index of the upper and lower sidebands of each 16 symbol block: Δ ═ P_U-P_L|

Because the symbol boundaries are modulo-1080 values, the computed difference must be properly concealed to ensure that the minimum difference value is used: if Δ > 540, Δ ═ 1080- Δ

A peak index difference of zero indicates that the peak index from each sideband is the same, thus representing the maximum guarantee: the normalized correlation peak from each sideband corresponds to the presence of a valid digital signal.

As an additional method for verifying the digital signal quality metric, optionally, a frequency offset difference may be calculated for the upper and lower sidebands. According to the previously described acquisition algorithm, the phase of the complex samples at the peak positions of signal 1750 gives an indication of the frequency error that exists between the transmitter and the receiver, since the conjugate product at this point, as formed by peak forming module 1100, should produce a zero phase value in the absence of carrier frequency error. The conjugate product at this point of the signal peak, and indeed at every other point in the signal peak, should yield a zero phase value because mathematically, the conjugate product between the symbol samples with equivalent phase (as are the samples at the beginning and end of each received symbol) cancels the phase when there is no carrier frequency error. Any residual phase present at the peak of the signal output from the matched filter is proportional to the carrier frequency error and once the residual phase is determined, it is simple to calculate the frequency error. The range of frequency offset measured on either sideband is 1/2FFT binary intervals (bin spacing), which is equivalent to 1(2T) for a channel interval of 1/T, as shown in fig. 11 a. If the frequency offset estimate difference between the upper and lower sidebands is within a certain threshold, such as, for example, ± 1/16FFT binary intervals, then it is unlikely that any adjacent interference will have the same frequency offset (and peak index) as the desired signal of interest. Thus, the frequency offset difference indicates that the detected signal is in fact the desired signal of interest.

Referring to fig. 16, the peak and index from a single sideband (fig. 15, items 422 and 424) are combined to produce a peak difference and quality estimate. The peak correlation value 430 from the upper sideband signal processing represents the upper sideband signal quality. The peak correlation value 432 from the lower sideband signal processing represents the lower sideband signal quality. Optionally, the difference between the peak index 434 from the upper sideband signal processing and the peak index 436 from the lower sideband signal processing is determined by subtracting one index from the other, as shown by subtraction point 438. The absolute value of the difference is determined (block 440) and the signal is packed to ≦ 540 samples (block 442) to generate the peak index difference 444. The signal is packed to 540 samples because the symbol boundary offset is modulo-1/2 symbols, which means that the distance to the nearest symbol boundary is always 540 samples.

Once both peak index differences and quality estimates are calculated, they can optionally be compared to thresholds to implement appropriate decision rules. In addition to optionally estimating the peak index difference and the sum of the quality estimates from the two sidebands, the quality of each of the respective sidebands may be separately compared to a threshold. This allows signal quality assessment to be made even when one of its sidebands is corrupted by interference. In addition, quality state parameters reflecting different sensitivity levels may be used. In one example, the quality state parameter is a 2-bit value that indicates to a master controller of the digital radio receiver the quality of the currently tuned channel. In this example, the quality of the received signal increases as the status bit changes from 00 → 11. This allows the receiver manufacturer the ability to adjust the sensitivity of the quality algorithm by changing the threshold of the quality status bit.

The digital signal quality metric may also be used to generate a visual indication of the quality of the received signal on a display of the receiver. Now, a series of bars, called Digital Audio Availability Indicators (DAAI), indicate the strength of the received digital signal. The status bits of the quality status parameter may be associated with the number and size of bars with such an indicator.

As will be appreciated from reading the above description, the simplicity of the algorithm of the present invention limits the changes required to previously known receivers. The degree of influence on the baseband processor and the main controller of the receiver is as follows.

In processing the first acquisition block, the baseband processor must now calculate a normalized waveform, as shown in fig. 15. This entails computing the magnitude squared of the current 1080-sample input symbol and 1024-sample delayed version, adding the magnitude squared vector, accumulating the sum over 16 symbols, match filtering it, and squaring the resulting vector. In addition to increasing in MIPS (million instructions per second), additional memory must be allocated for latency, accumulation, and FIR filtering operations. Other changes include normalizing the correlation peak via vector division, finding the peak and index of the normalized correlation peak, and calculating the peak index difference. The baseband processor may then apply the decision rule and set the quality state parameters appropriately according to the digital signal quality metric.

The digital signal quality metric may be applied to many areas of interest, such as, for example, FM seek-search functions, resolution of 300kHz spaced interferers, first adjacent interfering sideband selection, and diversity switching. Algorithms are implemented in the reference receiver and tested over a range of carrier-to-noise ratios in a variety of environments to implement the digital search-and-search function. In particular, performance was tested within dB of the digital audio threshold, under Additive White Gaussian Noise (AWGN), AWGN with one sideband, Urban Fast (UF) Rayleigh (Rayleigh) fading, and UF Rayleigh fading with 6-dB first adjacent signal.

At each point, at least 300 reacquisitions are forced. For each attempt, the peak index difference and quality estimate are recorded and the following decision rule is applied: (Q)_U≥T_Q) Or (Q)_L≥T_Q) Or (Q)_L+Q_U≥T_Q+0.2 and Δ ≦ T_Δ).

Then at T_QAnd T_ΔThe probability of stopping is calculated and plotted over a range that allows for judicious selection of those thresholds.

The probability of stopping in various circumstances is shown in FIGS. 20 through 23, with respect to the carrier-to-noise ratio Cd/N0, where T_Δ8 and T_QRanging from 0.4 to 0.6. Within this threshold, the probability of a stop with no input signal is practically zero. On each figure, the digital audio threshold is represented by vertical line 500.

After referring to the graphs in fig. 20-23, the suggested default thresholds are set as follows: t is_Δ8 and T_Q0.5. At all environment and carrier-to-noise ratios, these thresholds yield the following performance: the probability of losing a strong station is best minimized while at the same time minimizing the probability of a false stop for a weak signal. The probability of stopping under various circumstances by using these default thresholds is shown in fig. 24. On each curve, the digital audio threshold is marked with a square.

The curves on fig. 24 indicate that the performance in terms of AWGD is quite good. At high carrier-to-noise ratios, the probability of detection is high. Likewise, at low Cd/No values, the false alarm rate is very low. A steep transition region around the digital audio threshold is desirable. In fading environments, a longer dwell time may be utilized to reduce false alarms, but at the expense of increased frequency band scan duration.

The invention provides a method for detecting FM digital HD Radio^TMA method and apparatus for a fast and accurate look-for and search function for the presence of a signal. The algorithm can be combined with existing analog FM find and search techniques to provide an improved method of generating FM find and search functions (for analog, hybrid, and all-digital signals). The methods described herein may be implemented using a software programmable digital signal processor or a programmable/hardwired logic device, or any other combination of hardware and software sufficient to perform the described functions.

While the invention has been described in terms of its preferred embodiments, those skilled in the art will recognize that various modifications may be made to the disclosed embodiments without departing from the scope of the invention, as set forth in the claims.

Claims (25)

1. A method for detecting the quality of a digital radio signal, the digital radio signal comprising an upper sideband and a lower sideband, the method comprising the steps of:

receiving a digital radio signal comprising a series of symbols;

forming an upper sideband correlation waveform and a lower sideband correlation waveform each having a peak corresponding to a symbol boundary;

normalizing the upper sideband correlation waveform and the lower sideband correlation waveform;

calculating a peak value of the normalized upper and lower sideband correlation waveforms, wherein the peak value represents a quality of the received digital radio signal;

comparing the peak values of the normalized upper and lower sideband correlation waveforms to a first predetermined threshold; and

the sum of the peak values of the normalized upper and lower sideband correlation waveforms is compared to a second predetermined threshold.

2. The method of claim 1, further comprising the steps of:

the peak value of each normalized correlation waveform is compared with a predetermined threshold value.

3. The method of claim 1, further comprising the steps of:

the status flag is set to indicate whether the received digital radio signal exceeds a predetermined quality threshold.

4. The method of claim 1, wherein samples received on the upper and lower sidebands are processed separately.

5. The method of claim 4, further comprising the steps of:

each sideband in the digital radio signal is filtered prior to the step of forming the correlation waveform.

6. The method of claim 5, wherein the filtering step is performed using a finite impulse response filter.

7. The method of claim 1, wherein each correlation waveform is based on amplitudes of samples of leading and trailing portions of an orthogonal frequency division multiplexing symbol.

8. The method of claim 7, wherein the amplitudes of the samples of the beginning and ending portions of the orthogonal frequency division multiplexing symbols are tapered.

9. The method of claim 1, wherein the symbols comprise orthogonal frequency division multiplexed symbols and each correlation waveform is based on a cyclic prefix applied to the symbol.

10. The method of claim 1, further comprising the steps of:

the difference in frequency offset of the correlation waveforms of the upper and lower sidebands is calculated.

11. The method of claim 10, further comprising the step of:

the frequency offset difference is compared to a predetermined threshold.

12. A method for detecting the quality of a digital radio signal, the digital radio signal comprising an upper sideband and a lower sideband, the method comprising the steps of:

receiving a digital radio signal comprising a series of symbols;

forming an upper sideband correlation waveform and a lower sideband correlation waveform each having a peak corresponding to a symbol boundary;

normalizing the upper sideband correlation waveform and the lower sideband correlation waveform;

calculating a peak value of the normalized upper and lower sideband correlation waveforms, wherein the peak value represents a quality of the received digital radio signal;

determining a peak index of the normalized upper sideband correlation waveform and a peak index of the normalized lower sideband correlation waveform; and

a peak index difference is calculated that represents the difference between the peak indices of the normalized upper and lower sideband correlation waveforms.

13. The method of claim 12, further comprising the step of:

comparing a sum of peaks of the normalized upper and lower sideband correlation waveforms to a first predetermined threshold; and

the peak index difference is compared to a second predetermined threshold.

14. A receiver for detecting a digital radio signal, the digital radio signal comprising an upper sideband and a lower sideband, the receiver comprising:

an input for receiving a digital radio signal comprising a series of symbols; and

a processor for calculating a peak value corresponding to a symbol boundary of the normalized upper sideband correlation waveform and the normalized lower sideband correlation waveform, wherein the peak value represents a quality of the received digital radio signal, and the processor compares the peak values of the normalized upper sideband correlation waveform and the normalized lower sideband correlation waveform to a first predetermined threshold and compares a sum of the peak values of the normalized upper sideband correlation waveform and the normalized lower sideband correlation waveform to a second predetermined threshold.

15. The receiver of claim 14, wherein the processor sets a status flag to indicate that the received digital radio signal exceeds a predetermined quality threshold.

16. The receiver of claim 14, wherein the samples received on the upper and lower sidebands are processed separately.

17. The receiver of claim 16, further comprising:

a filter for filtering each sideband of the digital radio signal before the processor calculates a peak of the normalized correlation waveform.

18. The receiver of claim 17 wherein the filter comprises a finite impulse response filter.

19. The receiver of claim 14, wherein each correlation waveform is based on amplitudes of samples of leading and trailing portions of an orthogonal frequency division multiplexing symbol.

20. The receiver of claim 19, wherein the amplitudes of the samples of the beginning and ending portions of the symbol are tapered.

21. The receiver of claim 14, wherein the symbols comprise orthogonal frequency division multiplexed symbols, and each correlation waveform is based on a cyclic prefix applied to the symbol.

22. The receiver of claim 14, wherein the processor calculates a frequency offset difference of the correlation waveforms of the upper and lower sidebands.

23. The receiver of claim 22, wherein the processor compares the frequency offset difference to a predetermined threshold.

24. A receiver for detecting a digital radio signal, the digital radio signal comprising an upper sideband and a lower sideband, the receiver comprising:

an input for receiving a digital radio signal comprising a series of symbols; and

a processor for calculating a peak value corresponding to a symbol boundary of the normalized upper and lower sideband correlation waveforms, wherein the peak value represents a quality of the received digital radio signal, and the processor compares at least one of the peak values of the normalized upper and lower sideband correlation waveforms to a predetermined threshold, and the processor determines a peak index of the normalized upper sideband correlation waveform and a peak index of the normalized lower sideband correlation waveform, and a peak index difference represents a difference between the peak indexes of the normalized upper and lower sideband correlation waveforms.

25. The receiver of claim 24, wherein the processor compares a sum of peak values of the normalized upper and lower sideband correlation waveforms to a first predetermined threshold and compares a peak index difference to a second predetermined threshold.