DIRECT SAMPLING RECEIVER
FIELD OF THE INVENTION
The present invention relates generally to a direct sampling receiver with no downconversion, and more particularly, to a Global Positioning System (GPS) receiver for direct L-band sampling.
BACKGROUND OF THE INVENTION
The Global Positioning System (GPS) is part of a satellite based navigation system developed by the United States Department of Defense. It provides global coverage with precise navigation in all weather conditions. In a fully operational GPS, the entire surface of the earth is covered by up to twenty-four satellites dispersed in six circular orbits with four satellites in each orbit. Each orbital plane intersects the equator at 55 degrees.
Each GPS satellite transmits two spread spectrum L-band carrier signals termed LI and L2. Carrier frequency LI is allocated to 1575.42 MHz and carrier frequency L2 is positioned at 1227.60 MHz. The LI signal from each satellite is binary phase shift key (BPSK) modulated by two pseudo-random-noise or PRN codes, the coarse acquisition code (C/A-code) and the precision code (P-code). The L2 signal from each satellite is BPSK modulated by only the P-code. The use of two carrier signals LI and L2 to carry the P-code permits correction of the propagation delays of the signals through the ionosphere.
The C/A-code is a relatively short coarse grained code having a clock or chip rate of 1.023 MHz. The C/A-code has a length of 1023 chips and thus repeats every millisecond. The P-code is a relatively long fme-grained code having an associated clock or chip rate of
10.23 MHz. The full P-code repeats every 259 days, with each satellite transmitting a unique portion of the full P-code. The portion of the P-code used for a given GPS satellite has a length of precisely one week before this code portion repeats. Accordingly, high precision navigation receivers use C/A code for rapid satellite signal acquisition and then switch to the P-code for obtaining the precise navigational solution.
Superimposed on each code, at 50 bits per second, is the satellite data containing the satellite ephemeris, the almanac for all other satellites, clock error parameters, and satellite status. In operation, the GPS receiver slews the code it generates to match the one received from the chosen satellite. When the match occurs, the receiver notes the code from the satellite is time delayed by an amount corresponding to the range from the satellite to the receiver. This range is called the pseudo-range because it contains an offset corresponding to the error in the receiver's clock. After obtaining four or more pseudo-ranges from four
or more satellites, the receiver, with the knowledge of their ephemeris, can solve four simultaneous equations to obtain its position (x,y,z) and the time t. The GPS has opened a wealth of consumer and military applications where precise positioning is required or highly desirable. Today, the GPS receiver is being utilized in automobile navigation, surveying, precision farming, etc. However, many applications are not being implemented because of the excessive cost, size and power consumption of the currently available GPS receivers. For example, the size and power consumption of GPS receivers prevent their integration into hand-held cellular telephones which would otherwise be a natural platform. The automobile navigation system is another example where the full potential of GPS is not being realized. The sales of this equipment in the automotive industry has been slow and relegated primarily to luxury vehicles because of its relative high cost. Another potential market which has not yet been fully exploited pertains to military applications. The current price, size, weight and power consumption of militarized GPS equipment makes it difficult to broaden its application by integrating it with other command, control, and communication systems.
The conventional GPS receiver employs an antenna capable of receiving the L-band GPS signal. The L-band signal is downconverted in an analog fashion to an intermediate frequency (IF) by a mixer. The mixer operates by multiplying the received L-band signal with a locally generated oscillator (LO) signal. The IF signal is then filtered to reduce its noise bandwidth and eliminate unwanted sidebands. Subsequently, the IF signal, now at a much lower frequency (say, tens of MHz) than the L-band signal, is typically sampled by a low speed analog-to-digital converter (ADC), which produces the baseband signal in digital form. The baseband signal retains all the information, modulation, and bandwidth of the original L-band signal but its frequency only spans from zero to the signal bandwidth. The baseband signal is then digitally code-correlated, demodulated, and signal processed to obtain the navigational solution. The complexity of the analog downconversion circuitry, which requires off-chip resonators for LO generation and IF filters for band shaping and image rejection, has contributed to the excessive cost, size and power consumption of the GPS receiver.
To make the GPS receiver economically viable for wider applications and integration with other systems, the architecture must be simplified. The consolidation of a GPS receiver into a single integrated circuit (IC) would be an attractive approach. However, heretofore, the integration of the GPS receiver onto a single IC has not been possible due to the requirements for the analog downconverter and the IF circuitry. To this end, a GPS receiver capable of sampling directly at GPS L-band frequencies would eliminate the need for LO generation, mixing and IF filtering in the GPS receiver. The problem associated with L-band
sampling, however, is that the tight phase jitter requirements for the L-band ADC dictate that it be done with a high speed technology such as gallium-arsenide (GaAs). Unfortunately, while the GaAs approach can potentially provide for the RF amplification and the sampling, it cannot provide the density required for integration of the remaining digital signal processor (DSP), the memory, and the microprocessor onto a single IC, and therefore, is not a viable solution.
Accordingly, there is a current need for an innovative approach for consolidating a GPS receiver into a single IC. The innovative approach must present a low cost solution which can provide the high density circuitry that silicon CMOS technology provides for the DSP, the memory, and the microprocessor and at the same time allow the high speed performance demanded by the GPS receiver.
SUMMARY OF THE INVENTION
According to one embodiment of the present invention, a direct sampling receiver employs a sampling circuit for downconverting a radio frequency (RF) signal to a baseband signal by generating a discrete-time signal comprising the baseband signal from periodic sampling of the RF signal. The sampling circuit is preferably an analog-to-digital converter (ADC) which produces a digital baseband signal at its output. The digital baseband output is demodulated and processed. Noise level reduction is achieved by employing a filter in front of the sampling circuit for removing alias frequencies from the desired RF signal.
In one embodiment, the direct sampling receiver is used for GPS applications. The GPS receiver RF section is implemented with Silicon-Germanium (SiGe) technology which provides AGC RF amplification of the initial L-band signal, alias frequency filtering of the AGC amplifier output, and direct downconversion to digital baseband without the need for analog IF circuitry. The manufacturing compatibility of the SiGe technology with silicon CMOS technology, which is preferably used for the baseband DSP, the memory, and the microprocessor, makes the consolidation of the entire GPS receiver onto a single IC possible. Off-chip pre-fϊltering is also be desirable in certain applications.
An attractive feature of the single IC GPS receiver using SiGe/siliconCMOS technology is that high density, small size, and low power consumption can be achieved at low cost. As a result, GPS application can now be integrated with a wide range of other hand-held systems or devices where battery life, battery weight and system cost have been of such paramount importance in the past that integration with a bulky and power hungry GPS module would not be possible. The integration of GPS integration with other handheld systems can greatly enhance the performance and utility of the latter. For example, the integration of GPS into cellular phones will allow new functions to be performed such as automatic user location
attachment to a 911 message to avoid error caused by user stress. Automobile navigation systems will also become commonplace if GPS function is provided by a low cost IC which can reside in, for example, the radio. The single IC GPS receiver will also find unlimited applications in the military arena. It is envisioned that every soldier will have GPS access integrated into common field equipment or as a small stand-alone communication-navigation pack with self contained batteries.
In summary, the single IC GPS receiver will enable new uses, expand current uses, and create the opportunity for integrating position sensing with other existing user functions, as well as a host of other equipment not yet envisioned.
BRIEF DESCRIPTION OF THE DRAWINGS
These and other features, aspects, and advantages of the present invention will become better understood with regard to the following description, appended claims, and accompanying drawings where:
FIG. 1 is a functional block diagram of a single IC GPS receiver in accordance with an embodiment of the present invention;
FIG. 2 is a functional block diagram an RF section of a single IC GPS receiver with off-chip pre-filtering in accordance with an embodiment of the present invention;
FIG. 3A is a graph showing how undesired alias frequencies fold into the baseband in accordance with an embodiment of the present invention.
FIG. 3B is a graph showing the bandpass characteristic of a noise shape filter in accordance with an embodiment of the present invention; FIG. 4 is a graph showing the bandpass characteristic of an anti-aliasing filter in accordance with an embodiment of the present invention; and
FIG. 5 is a functional block diagram of a dual channel high precision application of a GPS receiver in accordance with an embodiment of the present invention.
DETAILED DESCRIPTION
The direct sampling receiver is well suited for use as a navigation or survey receiver for GPS applications. Navigation receivers are capable of receiving C/A code modulated GPS frequency at LI and typically can determine the receiver's position on the earth's surface to within- 10 meters rms. Survey receivers are dual channel devices capable of receiving both C/A code and P-code modulated GPS frequency at LI . High precision survey receivers receive the P-codes modulated on both frequencies LI and L2. The use of both GPS frequencies allows extra path delays of the GPS signals through the ionosphere to be removed from the navigational solution. The following discussion pertains to a single LI channel GPS
receiver, however, it will be understood that this technology can readily be applied to the L2 channel of a high precision survey receiver, as well as other communication systems. FIG. 1 is a functional block diagram of a GPS receiver integrated into a single IC 10a using Silicon-Germanium (SiGe) technology developed by IBM. SiGe technology supports the fabrication of extremely high speed (ft of 65 GHz) bipolar transistors which allows direct sampling at GPS L-band frequencies. As a result, a baseband signal can be generated without analog downconversion to an IF frequency. SiGe technology has the further advantage of being able to integrate with standard CMOS technology used for digital signal processing thus allowing a single chip implementation of the entire GPS receiver system. In addition, the SiGe Bi-CMOS implementation is estimated to consume only one-quarter of the power of the more esoteric and expensive technologies such as GaAs.
An RF section 12 at the front end of the GPS receiver 10 directly samples the GPS L- band frequencies at input 12a to the IC 10a to produce a digital baseband signal. This is achieved by periodic sampling of the GPS L-band frequencies to produce a discrete-time baseband signal. The L-band sampling capability of the RF section 12 is the key component for integrating the GPS receiver 10 onto a single IC. The baseband signal is digitally demodulated by a conventional twelve channel digital signal processor (DSP) 14. Each channel demodulates a particular satellite signal and preferably includes a PRN code correlator, a code tracking loop, a carrier tracking Costas loop, bit sync and demodulation, and data I/O circuitry (not shown). The basic DSP channel is then repeated twelve times so that multiple GPS satellites can be demodulated simultaneously. For a navigation receiver about six thousand CMOS gates are needed and for a survey receiver about eleven thousand gates are needed for each GPS channel.
A conventional microprocessor 16 with sufficient processing power to control the DSP 14 provides data acquisition from each active satellite DSP channel and forms the navigation solution. The microprocessor 16 is supported by memory 18 comprising RAM, ROM and nonvolatile RAM. The microprocessor 16 supplies the synchronous and asynchronous communication signals to I/O ports 19 for interfacing through IC output 19a with external devices. The RF section 12, DSP, microprocessor 16, memory 18 and I/O ports 19 are implemented with SiGe Bi-CMOS technology onto the single IC 10a.
A functional block diagram of the RF section 12 of the GPS receiver interfaced with off-chip circuitry is shown in FIG. 2. Interfacing with the GPS receiver is an antenna 20 capable of receiving the PRN code modulated GPS signal (LI). The LI signal from the antenna is pre-filtered and amplified before being applied to the RF section 12 of the GPS receiver. A cascade circuit positioned in front of the GPS receiver IC 10a comprising a bandpass filter 22, a low-noise amplifier (LNA) 24 and a noise shaping filter 26 is used for
this function. The pre-filtering process is performed off-chip in order to optimize the noise figure of the overall system by the use of the discrete LNA 24. It will be understood that the pre-filtering process could be consolidated onto the receiver IC in environments and applications where optimal noise figure performance is not required.
The signal from the antenna 20 is first applied to a bandpass filter 22. The bandpass filter is centered at the GPS carrier frequency (LI) and has a noise bandwidth of 40-50 MHz. The bandpass filter 22 eliminates interference from cellular telephone transmissions and other high power signals that might otherwise overload or saturate the LNA 24. The bandpass filter 22 has its output connected to the LNA 24. The LNA 24 provides good noise figure performance with high gain. The output of the LNA 24 is connected to the noise shaping filter 26. The noise shaping filter rejects out-of-band noise that has been amplified up by the LNA 24. Without this filter 26, the out-of-band noise would be converted down to the desired baseband frequency by the GPS receiver 10.
The noise shaping filter 26 is in part necessitated by the direct conversion of the L- band carrier to a digital baseband signal by the GPS receiver. The importance of the noise shaping filter 26 becomes apparent by examining the relationship between the input RF frequency frf, the sampling frequency f g and the resultant digital baseband frequency f ^ which can be expressed by the following equation
frf = nfs ± fb (1)
where n is any integer. Therefore, we observe the unfortunate fact that an infinite number of RF frequencies will be downconverted to the same baseband frequency f^. In fact, if we set the baseband bandwidth to fs/2, the Nyquist limit, then the entire RF spectrum will be converted into baseband: the + sign in Eq. (1) is responsible for folding a section of width fg/2 above nfs into the baseband and the - sign for the fs/2 section below nfs into the baseband with a frequency inversion as shown in FIG. 3A. Thus, a noise shaping filter is required to prevent this from happening. The characteristic of the noise shaping filter is thus dictated by the sampling frequency fs and the desired baseband frequency bandwidth in accordance with Eq (1). In other words, the filter is designed such that for a required baseband bandwidth, a range of RF frequencies corresponding to only one n and only one sign (+ or -) in Eq. (1) will pass through the filter. For purposes of designing the noise shaping filter 26, let the P-code chip rate of 10.23
MHz be f, then the LI carrier frequency 1575.42 MHz is 154f, the PRN code-spreaded main lobe width is 2f and the bandwidth including the first sidelobes is 4f as shown in FIG. 3B. In order to satisfy the Nyquist sampling rate and obtain both the in-phase (I) and quadrature
(Q) channels, the minimum sample frequency for fs is 4f or 40.92 MHz. Thus, according to Eq. (1), if we set fs at 41.189 MHz, slightly above the minimum required sampling frequency, 38fs will fall at the lower edge of the main lobe of LI as shown in FIG 3B. As a result,
38fs, by virtue of Eq. (1), will then convert the entire LI main lobe into the baseband, as further illustrated by the frequency domain representation of FIG. 3B.
The sampling clock for fs is generated by an oscillator 28 controlled by an off-chip crystal resonator 30. Referring to FIG. 3B, each of the alias sample frequencies 33 (nfs) which occurs every 41.189MHz (fs) across the bandwidth 27 of the LNA can downconvert noise in its frequency neighborhood of ± 2f into the baseband according to Eq. (1). Therefore, a noise shaping filter is required with bandpass characteristics shown by 29 to allow only the LI GPS signal (and in-band noise) through and to be converted into the baseband. As shown, only the LI signal which satisfies Eq (1), (38fs + f^ with f^ going from 0 to 20.46 MHz), is in the passband of 29. Other frequencies satisfying Eq. (1) are suppressed by the noise shaping filter response 29.
Referring back to FIG. 2, the output of the noise shaping filter 26 is applied to the RF section 12 of the GPS receiver 10. An AGC amplifier 32 is connected to the noise shaping filter 26. Preferably, the AGC amplifier provides low noise amplification with 45 dB of gain and 50 dB of AGC range. The AGC amplifier outputs a signal having about 90 dB gain, 45 dB is from the antenna, which makes the signal amplitude sufficient for analog-to-digital sampling. This AGC function allows the GPS receiver 10 to continue operating in the presence of high levels of intentional or incident jamming. It has been shown that one embodiment of the GPS receiver 10 will support jamming-to-signal (J/S) ratios of 60 dB which is very important in commercial applications where incidental jamming occurs with surprising regularity or in military applications where the J/S ratio must be optimized.
The AGC amplifier 32 is a high gain broadband device with .a moderate noise figure. The out-of-band noise, which has been suppressed down to thermal noise level by the noise shaping filter 26, is now amplified up again by the AGC amplifier 32. While the out of band noise level is significantly below the in-band signal plus noise, the aliasing nature of Eq. (1) will again fold in a broad bandwidth of noise into the baseband in the manner discussed previously pertaining to FIG. 3A. This will severely degrade the signal to noise ratio. As shown in FIG. 4, an AGC amplifier having a 4 GHz bandwidth 31 will produce around one- hundred alias sampling frequencies 33 spaced apart by 41.189 MHz (fs) with about two hundred alias bands, each 20.46 MHz (2f) wide. This would result in an elevated noise level at the output of RF section 12 that would prevent the demodulation of the L-band carrier and the recovery of the PRN code.
The effective noise figure of the RF section 12 increases by the number of aliased bands which are downconverted to baseband. In decibels, the effective noise figure of this portion of the RF section is
Neff = Nagc + 101°glθ(2Bagc fs) (2)
where Ne^- is the effective noise figure of the RF section 12, Naεc is the noise figure of the AGC amplifier, Ba„c is the AGC amplifier bandwidth and fs is the sampling rate. It can be seen from Eq. (2) that if the undersamping ratio (Bagc/fs) is large, then the effective noise figure of the RF section 12 will also be large. The total noise figure of the receiver is given by
NFt= NFlna + (NFeff 1)/G, na
where NFt is the total receiver noise factor, NF |na is the composite noise factor of the amplifiers and filters which precedes the AGC amplifier, NFeff is the effective noise factor
Neff derived from Eq. (2) of the RF section 12 given as a linear factor (NFeff = 10 , and Gjna is the composite linear gain of the amplifiers and filters which preceed the AGC amplifier.
In the circumstance where NFeff is large due to the large undersampling ratio, and Gjna is insufficient to reduce this in the above equation, then the total receiver noise factor will be increased due to the noise contribution of the RF section 12. In practice this is highly undesirable. To alleviate this problem, an anti-aliasing filter 34 is connected to the output of the
AGC amplifier 32 as shown in FIG. 2. The anti-aliasing filter 34 is preferably designed to pass both L-band GPS carriers with a bandpass 35 between 1200 and 1600 MHz. This approach allows the same GPS receiver IC to be used in either the LI or L2 channel, yet reject the vast majority of alias frequencies. The anti-aliasing filter 34 reduces the number of alias frequencies bands downconverted to baseband to approximately twenty from two hundred, which results in a 10 dB improvement in the RF section 12 noise figure. This improvement can be used to reduce the required LNA gain resulting in a lower cost, lower power consumption receiver.
The signal at the output of the anti-aliasing filter 34 is coupled to a 2-bit sign- magnitude analog-to-digital converter (ADC) 36. Most commercial GPS receivers use a one- bit sampler which is insufficient for good sampling performance in the presence of jamming. An overflow bit from the ADC 34 output is used by the DSP 14 to control the gain of the AGC amplifier 32. Specifically, the gain of the AGC amplifier 32 is controlled by monitoring
the number of overflow bits detected by the DSP 14. If an excessive number of overflow bits are detected, the gain of the AGC amplifier 32 is reduced to keep it out of saturation. FIG. 5 shows a GPS configured as a high accuracy survey receiver. Two identical GPS receiver IC's 10a are used which communicate with each other through a conventional dual port memory 38. A pre-filter (not shown) positioned in front of each GPS receiver IC 10a determines whether the GPS receiver is an LI or L2 channel device. Cross-correlation products are generated by passing LI samples to the L2 channel, and L2 samples to the LI channel through the dual port memory 38. This process is performed by conventional handshaking techniques known in the art. This dual channel approach for the GPS single chip receiver design has the benefit of accommodating both high accuracy and standard accuracy applications.
It is apparent from the foregoing that the present invention satisfies a long felt need for an innovative approach for integrating a GPS receiver into a single IC. It will be understood that the features of the present invention may be embodied in other specific forms and used in a variety of communications and other electronic applications, without departing from the spirit or essential attributes of the present invention. It is, therefore, desired that the present embodiment be considered in all respects as illustrative and not restrictive, reference being made to the appended claims rather than the foregoing description to indicate the scope of the invention.