Session B1: RECEIVER DESIGN 1
Paper #1

GPS RADIO IP FOR WIRELESS COMMUNICATIONS:
P. Paddan, P. Naish, M. Phocas, Parthus (UK) Ltd

GPS is poised to play a critical role offering commercial opportunities in wireless communications through the E911 directive and 'location-based m-commerce' services. The NavStream 3000 platform provides solutions for the critical areas for implementing GPS in mobile phones - low cost, low power consumption, accuracy, sensitivity and noise immunity.

The GPS receiver typically comprises two functions: the radio front end and the baseband digital processor. CMOS digital technology improvements now readily (almost) permit the design of the baseband to be portable across silicon vendors and to be incorporated in a system level realisation. The challenge for RF is to move in the same direction and become an intellectual property (IP) component in the convergence of GPS with wireless communication applications. In this paper we will present the results of our first attempts at a RF IP design approach with the RF3000.

We have chosen SiGe BiCMOS processes as the best technology to meet the conflicting requirements of high sensitivity and low cost/space. RF3000 downconverts the L1 band carrier at 1575.42MHz to a frequency and format suitable for subsequent digital processing by our GPS3000 IP processor.

Key technical requirements on the RF subsystem are to achieve the usual demanding GPS requirements of image rejection and blocking immunity simultaneously with the ability to co-operational function within a mobile phone handset, both in a hostile RF sense and also to utilise the scarce resources available, of power, space, bandwidth and MIPS (clock-cycles) to the interferers generated by the mobile telephone protocols.

We have developed GPS RF models at multiple abstraction levels using Agilent's ADS tool including the baseband correlators and code/carrier mixers. With this approach we are able to define, optimise and specify the performance of each block for silicon implementation. This makes the process of translating the design to another process much easier as the overall performance of the system becomes the target rather than the result of the design process.

It has been necessary to design RF3000 with the ability to operate from the same reference clock as the host protocol thereby permitting a common clock strategy to be used. An on-chip phase locked loop synthesises suitable local oscillator frequencies for the complex mixers.

This paper will describe our developmental process, exploring compromises using modelling techniques, decisions and some of the novel architectures and solutions that were needed. Key results from the characterisation phase will be presented.
[Back to Program]

Session B1: RECEIVER DESIGN 1
Paper #2

AN EFFICIENT WEAK SIGNAL ACQUISITION ALGORITHM FOR A SOFTWARE GPS RECEIVER:
D.M. Lin, J.B.Y. Tsui, Sensors Directorate, Air Force Research Laboratory

Compared with hardware GPS receivers, software GPS receiver has lots of advantages for the weak signal acquisition. The frequency domain circular correlation can be used(1). More intelligent schemes can be built-in. Longer data can be used for coherent integration. Several acquisition methods were reported previously for a typical software GPS receiver(2). Also several innovative approaches have been reported to integrate long data by partition(3). In order to acquire the weak signal, there is another difficulty to be overcome. The difficulty is how to handle the navigation code phase transition which potentially can occur every 20 ms. The paper submitted will discuss a new acquisition algorithm to process 40 ms of data regardless of whether or not there are the phase transitions and where they occur. The algorithm applies coherent integration by partition. The correlation peak of this approach is in the frequency domain. If there is no phase transition, the result will be the same as regular 40 ms coherent integration. If there are phase transitions occur in the 40ms data or the carrier is off the bin or both, the spectrum of the signal spreads. However, there are only three dominant spectral lines surrounding the bin of the carrier frequency. In this new approach, if the peak of the result of the coherent integration by partition is above the threshold the acquisition is completed. Otherwise, the adjacent frequency bin is added and peak of the result is compared against the threshold that is developed based on the noise in this level. If the peak is above the threshold the acquisition is completed. This will take care of the case if the carrier frequency is off the bin. If the detection fails then every other bin is added and the same process repeats. This will take care of the case of two phase transitions in the data set. If the detection fail, three adjacent bins are added the same process repeats. The threshold used here is developed based on the noise in this level. This will take care of the case of one phase transition, and the case of the combination of both mentioned above. Depending on the carrier frequency and the location of the phase transitions in the 40 ms of data, the preliminary result shows the new algorithm can improve C0/N by 12 to 16 dB, compared with the typical C0/N which can be obtained by 1 ms of coherent integration. The paper will also discuss how the threshold for each noise level is developed. In order to measure the efficiency of the algorithm, the method of continuously non-coherent summation of 10-ms coherent integration(4) will be analyzed to compare the performance.

(1) D.J.R. VAN NEE and A.J.R.M. COENEN "New Fast GPS Acquisition Technique Using FFT", Electronic Letters 17th January 1991 Vol. 27 No. 2, PP 158-160
(2) DAVID M LIN and JAMES B.Y. TSUI "Acquisition Schemes for Software GPS Receiver", Proceedings of ION GPS 98" September 15-18, 1998, Part 1, pp. 317-326
(3) DAVID M LIN and JAMES B.Y. TSUI "Direct P(Y)-Code Acquisition Algorithm for Software GPS Receivers", Proceedings of ION GPS 99" September 14-17, 1999, Part 1, pp. 363-368
(4) DAVID M LIN and JAMES B. Y. TSUI " A Software Receiver for weak signals", IEEE International Microwave Symposium, May 20-25, 2001, http://estd-www.nrl.navy.mil/ims2001/fileuploads/THIF-37-digest-28175629.pdf
[Back to Program]

Session B1: RECEIVER DESIGN 1
Paper #3

A NEW NAVIGATION BIT SYNCHRONIZATION METHOD FOR A GPS RECEIVER:
M. Kokkonen, Nokia Research Center, Finland; S. Pietila, Nokia Mobile Phones, Finland

Since each of the transmitted data bits in a GPS signal are repeated 20 times i.e. transmitted on 20 successive code periods, there is initially uncertainty about the bit boundaries in a GPS receiver. This uncertainty must be removed in order to be able to detect navigation data bits, use coherent integration in the tracking loops of the receiver and compute pseudo-ranges in a receiver tracking only C/A-code. If receiver position and time is roughly known, these can be used for bit synchronization, otherwise bit boundaries have to be detected from the received signal using statistical signal processing. We concentrate on this latter alternative in this paper.

The traditionally used statistical bit synchronization procedure (the histogram approach) is based on looking at the sign changes of the prompt correlator output, which is obtained by integrating the received signal over one code period. If the number of sign changes at some candidate position is clearly larger than in other positions, the bit boundary is declared to have been detected. The use of signs of correlator outputs corresponds to making hard-decisions of those outputs.

A new method to find bit boundaries in a GPS receiver is proposed. The method basically computes time averages of bit energies for all twenty possible bit boundary positions and selects the position which maximizes the bit energy. The proposed method uses unquantized values from the correlator all the way until the final decision is done. This efficient use of available information means that the algorithm finds the bit boundaries quickly in moderate conditions and can reliably operate with a very weak signal, which is very important in e.g. positioning indoors. With this method bit boundaries can be detected e.g. from a signal with C/N0 equal to 20 dB Hz (simulation results are provided for verification), in contrast to the traditional method, which fails to find bit boundaries from this signal because the bit error rate of data from single code periods is excessive.

We also show how phase offset and small frequency errors can be tolerated, so that phase lock is not necessary for bit synchronization.
[Back to Program]

Session B1: RECEIVER DESIGN 1
Paper #4

MITIGATION OF THE NEAR-FAR PROBLEM BY SUCCESSIVE INTERFERENCE CANCELLATION:
P.H. Madhani, P. Axelrad, University of Colorado at Boulder; K. Krumvieda, J. Thomas, Data Fusion Corporation

Since the inception of GPS, plans to augment the system to improve its accuracy, availability, and integrity have been addressed. One such augmentation is the pseudolite (PL), which transmits GPS like signals. The PL can be deployed on ground, air, or water (aboard a ship). It has been used in critical environments such as aircraft approach and landing. The major drawback of the use of PLs is the associated "Near-Far Problem". Different approaches have been implemented to overcome the problem, e.g. pulsed PL transmission, PL signals transmitted with a frequency offset from L1 and receiver antenna design.

This paper discusses a successive interference cancellation (SIC) technique applied at the signal processing stage of the GPS receiver to mitigate the near-far problem. In SIC, strong signals are identified using standard channel tracking architecture. The strong signal replica is then removed from the composite signal and acquisition of the weaker signals is then attempted. Each signal identified may be successively removed from the composite, allowing for a significant reduction of the acquisition threshold for the weaker signals.

The paper will illustrate the performance of SIC applied to both simulated and experimental GPS data corrupted by near-far. It is shown that the SIC technique makes a receiver far more robust to continuous transmissions from the PLs, and eliminates the requirement to modify the PL transmitted signals and receiver antenna design. This technique can make use of the existing receiver technology and has the potential to provide significant gains over the conventional GPS receiver in the near-far scenario.
[Back to Program]

Session B1: RECEIVER DESIGN 1
Paper #5

ENHANCED SENSITIVITY FOR ACQUISITION IN WEAK SIGNAL ENVIRONMENTS THROUGH THE USE OF EXTENDED DWELL TIMES:
J.L. Shewfelt, R. Nishikawa, Sr., C. Norman, SiRF TEchnology Inc.

The assimilation of GPS technology into a variety of low power, small form factor consumer devices and the need to operate in obstructed, attenuated, and fluctuating signal environments has resulted in unique challenges for commercial GPS receivers. Due to antenna and other implementation losses typical of handheld receivers, the ergonomic considerations of handheld devices, and the anticipated use model scenarios for many consumer applications, the signal levels at which the GPS receiver is expected to perform are significantly lower that typical open sky conditions. As a result, the GPS receiver must acquire and track signals at increasingly lower signal levels.

The SiRFstarIIe chipset provides a flexible hardware and software architecture for implementing signal acquisition and tracking of low level signals using a combination of coherent and non-coherent hardware and software integration techniques. In the standard Satellite State Tracking Engine (SSTE), a hardware tracker is implemented to perform signal search, acquisition, detection, code lock, and carrier lock. Internal detection thresholds and hardware integration registers are programmed to provide a theoretical 95% Probability of Detection (POD) with a False Alarm Rate (FAR) or false detection rate of 3.0 x 10-2 per second.

The SiRFstarIIe reference receiver utilizes a two step threshold detection algorithm whereby satellites are initially searched at a threshold equivalent to approximately 40 dB-Hz, followed by a secondary search threshold equivalent to approximately 35 dB-Hz. The primary goal of the development project was to develop a system that could acquire signal at an initial threshold of 36 dB-Hz and a secondary threshold of 33 dB-Hz. These values were chosen because characterization of typical handheld receivers in real signal environment under a variety of conditions indicated that peak satellite C/No varied between 35 dB-Hz - 40 dB-Hz. In addition, the reference receiver utilizes a reacquisition threshold of approximately 36 dB-Hz which governs the performance of satellites dropping in and out of visibility. The goal of the development project is to lower the reacquisition threshold to 34 dB-Hz.

Software modifications were performed which varied the non-coherent integration dwell times in accordance with theoretical values to provide the improved acquisition sensitivity while maintaining POD and FAR. Laboratory tests were performed in order to demonstrate an improvement in acquisition sensitivity and to characterize acquisition performance in terms of minimum signal level, Time to First Fix (TTFF), and percentage of successful starts under various start-up conditions. Road tests were performed in order to demonstrate improvements in reacquisition performance and fix density in challenging environments. The tests were performed in both normal and attenuated signal environments and compared to the reference receiver baseline

The preliminary results of the testing has shown that extending the signal integration times and lowering the detection threshold can result in improved acquisition performance at lower signal levels without severely degrading acquisition performance at normal signal levels, and without an increase in FAR.

The results of this study also show that the system sensitivity performance closely tracks the predicted acquisition performance under nominal open sky signal conditions.
[Back to Program]

Session B1: RECEIVER DESIGN 1
Paper #6

MITIGATING SHORT DELAY MULTIPATH: A PROMISING NEW TECHNIQUE:
J.M. Sleewaegen, F. Boon, Septentrio Satellite Navigation, Belgium

In the recent years, several signal-processing techniques have been devised to mitigate errors induced by multipath signals. They have proved very efficient against multipath having a medium or large delay with respect to the direct signal. Typically, the errors are largely removed for multipath delays higher than around 20 m. From a theoretical point of view, this constitutes a dramatic improvement with respect to simpler techniques as the narrow correlator.

However, when analyzing field data, the resulting performance is often disappointing: the improvement over the narrow-correlator is marginal. The reason is that most of the multipath signals enter the receiver with a short delay after the direct signal, rendering the mitigation scheme ineffective.

This paper presents a new signal-processing algorithm for multipath mitigation, based on the a-posteriori estimation of the tracking errors. This method is shown to achieve at least 50% better immunity to short delay multipath compared to other state-of-the-art techniques. The performance is evaluated both from a theoretical analysis and from real data. A comparison with other methods demonstrates an improved mitigation capability in most practical situations.
[Back to Program]

Session B1: RECEIVER DESIGN 1
Paper #7

HIGH PRECISION ALGORITHM FOR PHASE ESTIMATES:
A. Steingass, German Aerospace Center (DLR), Institute for Communications and Navigation, Germany

State of the art
A high precision receiver for GPS uses the delay estimate as well as the phase estimate. The GPS signal consists of the carrier signal, the spreading code and the data signal. A DLL gives an estimate of the group delay of the satellite-signal. Another output is the despread navigation signal. This despread signal is then passed into a PLL for the phase estimate. Since the data signal is included in the signal, only an incoherent PLL (e.g. a Costas PLL) can be used for the phase estimate.

Turbo PLL - the new Approach
The algorithm presented here uses an incoherent Costas PLL for the coarse phase estimate. When the phase is estimated one can use a standard bit detector to receive the data-signal. GPS has an extreme high bit energy to noise ratio Eb/N0. The lowest C/N0 at which the satellite signal is trackable by a DLL is 25 dBHZ. But even at this lower end of this dynamic area, GPS obtains an Eb/N0 of 8dB. For this Eb/N0 an uncoded BPSK data transmission reaches a bit error rate of BER=10-4. After decoding of the GPS forward error correcting code ((32,26) Hamming code) the remaining BER goes down to about 10-10. This shows that there are almost no errors in the navigation data stream, a result which allows the following "turbo"-like approach:

1. Receive the GPS Signal.
2. Use an incoherent Costas PLL for a coarse phase estimate.
3. Correlate and decide the navigation data stream.
4. Use the forward error correction code to reduce the BER.
5. Remove the now known data stream from the despread signal.
6. Use a coherent PLL for a accurate estimate of the carrier phase.

Performance of the new algorithm
Figure 1 shows the performance of the new algorithm. For high SNR the Turbo PLL reaches the same performance as the Costas PLL. For the low SNR range the Costas PLL looses performance due to the squaring loss. In contrast to that behavior the Turbo PLL performs like a coherent PLL and reaches a gain of up to 3 dB at the lower end of the dynamics range.

FIGURE 1: Performance of the new algorithm (Turbo PLL in comparison to a standard PLL and the linearised analysis.
[Back to Program]

Session B1: RECEIVER DESIGN 1
Paper #8

GPS FOR THE E911 LOCATION REQUIREMENT - THE PRACTICAL IP APPROACH:
P. Anderson, J. Bickerstaff, Parthus (UK) Ltd

The paper discusses a practical approach to providing GPS as Intellectual property (IP). It details an implementation of GPS in a handset for indoor use. This includes the advantages of providing an IP solution compared to the methods normally used in implementing a stand-alone GPS.

GPS is the leading contender for incorporation into handsets to satisfy the E911 location requirement. GPS is also a mature technology, with proven location ability. It would appear, therefore, that the problem is simply to miniaturise GPS even further, and, at the same time, also reduce its power consumption. But the problem is more complex than this.

This paper takes a deeper look into the problem. GPS is usually designed for outdoor use. How will it fare indoors, where the signal (previously well below the noise) is attenuated another 20 dBs or more, and how does one achieve the time-to-fix of just a few seconds. More sensitivity, and faster time-to-fix can be addressed with large, parallel processing, architectures - more hardware (and more power). But handset and PDA manufacturers want less hardware, less chips, and more IP (intellectual property) solutions. However, existing GPS designs do not transfer well into IP, having been designed for single instantiations in custom chips.

This presents the IP designer with many challenges. They must build an expanded architecture in a small die space, and achieve lots of parallel processing with the use of little power. Simultaneously they must turn a complex, customised, design, into portable, technology independent, IP. In addition, the interaction with other system functions, is a key design issue. These considerations are often contradictory, and not easy to resolve. The paper addresses the design interactions between the various issues, illustrated with the design of a practical, and commercially available product, the Parthus NavstreamT3000 IP, designed specifically for portable appliance and mobile phone applications. The practical problems of interference from co-located systems (such as the phone transmitter) are also covered, and solutions proposed.

In conclusion the paper reviews the issues which separate a good, portable, E911 capable, IP solution from a conventional 'GPS chipset'.
[Back to Program]

Session B1: RECEIVER DESIGN 1
Alternate #1

SYNERGIES ON HANDSET ARCHITECTURE LEVEL FOR COMBINED GNSS AND WIRELESS LOCATION:
G. Heinrichs, MAN Technologie AG, Germany; B. Eissfeller, University Federal Armed Forces, Germany

Satellite-based navigation and positioning will have in future a tremendous influence on our daily life. In addition, due to the great success and still increasing dissemination of mobile phones interest in cellular network and/or handset based positioning techniques is growing rapidly in the mobile telephone community.

The U.S. Federal Communications Commission's E-911 mandate will soon require network carriers to provide location or geo-coding of emergency callers. Similar activities related to the E-112 are presently starting within Europe. Moreover, the increasing demand for commercial location-based services has driven cellular phone and network manufacturers to focus on accu-rate positioning solutions.

The location requirements for emergency callers outside urban areas can hardly be fulfilled without GNSS. Usually, with GSM the position is determined with cell identifica-tion/intersection and/or time difference techniques applied in the base station network. The additional drawback is that the user has no direct access to his position. However, this will change in future due to the new standard for cellular network-based wireless communication called UMTS/IMT-2000. Currently the DS-CDMA spread-spectrum access technique is the strongest candidate for the air interface of the third generation wireless communication sys-tems. Therefore, similar position determination techniques as known from GNSS can in princi-ple be applied which will result in a self-location capability of the mobile phone. Moreover, this opens the road for the combination of GNSS with cellular network-based wireless location and positioning particularly on receiver/handset architecture level. Thus, the combination of GNSS with cellular network-based wireless location and positioning will become an growing research and industrial area of interest in future. This will effect both, the GNSS/cellular network side and the GNSS receiver/mobile handset side.

The purpose of this paper is to give first answers on the question if handset architecture syner-gies exist for the combination of GNSS with wireless location. In order to identify synergies it is important to outline similarities and differences between wireless communication and satel-lite navigation. After this, guidelines for a combined use of GNSS and cellular network-based wireless location in terms of accuracy and availability will be determined and the basic signal characteristics of the existing and future GNSS as well as the new standard for cellular net-work-based standard UMTS/IMT-2000 will be presented.

Preliminary conclusions on the synergy issue will be drawn. In particular, it is tried to find first answers on questions related to the advantages, complexity, and component savings of new hybrid handset architectures. Finally, future trends are outlined.
[Back to Program]

Session B1: RECEIVER DESIGN 1
Alternate #2

FFT ACQUISITION OF PERIODIC, APERIODIC, PUNCTUAL, AND OVERLAID CODE SEQUENCES IN GPS:
C. Yang, Sigtem Technology, Inc.

This paper presents a comprehensive study of FFT-implemented circular correlation and its application to fast direct acquisition of GPS codes. This includes the periodic C/A-codes, practically non-periodic P(Y)-codes, never-repeating cryptographic M-code, and punctual codes which have been proposed to aid direct M acquisition as well as overlaid codes which are created by surface-reflected GPS signals extended beyond one code chip.

FFT operates on blocks of incoming and replica code samples, thus providing simultaneous search over the entire block of code phases. It is straightforward to work with periodic codes for circular correlation. However, it is not obvious for punctual and particularly long code sequences. One major concern is how to ensure that the incoming and replica code samples contained within the working block could be correlated. In addition, it is quite possible that the data bit sign may reverse in the middle of an integration interval. Jamming or co-located RF interference is another serious problem for acquisition. For weak signals or under wideband jamming, an extended coherent or incoherent integration may be employed to boost the signal to noise ratio but it requires an extra care to handle both the carrier and code Doppler effects and receiver clock errors during the prolonged integration. Furthermore, how to efficiently make use of complex FFT when the data length is not a power of two or highly composite is vital for practical implementation.

These design and computation issues are properly formulated in this paper and pertinent acquisition schemes are suggested. The solutions are illustrated with computer simulations and some of them are demonstrated with recorded GPS signals.
[Back to Program]

Session B1: RECEIVER DESIGN 1
Alternate #3

THE POST PROCESSING ALGORITHMS FOR TRANSLATOR-TYPE GPS RECEIVERS:
A. Pratt, D. Molyneux, Parthus (UK) Ltd

One important category of GPS receivers operates entirely autonomously and under high dynamic conditions. Such receivers would have a high risk of losing the tracking process on the satellite signals received if a conventional GPS receiver architecture were to be employed. For this reason several approaches have been developed which are distinguished by have all or part of the tracking processor installed in ground based equipment. This is connected to the GPS receiver by a data-link with (usually) a variable transmission delay.

The recovery of GPS data and the tracking of both carrier and code phase is more complicated for this receiver architecture. There are new problems too since the GPS receiver in the dynamic vehicle does not form any location solution nor GPS time.

This paper introduces the processing structures which have been implemented in ground based equipment to form tracking loops, derive the GPS satellite data message and transfer GPS time to the host vehicle without a data uplink.

Many performance benefits can be derived from post-processing the received GPS measurement including the removal of aiding data latency, processing the tracking loops with both forward and reverse time order, processing the tracking algorithms with the GPS data modulation removed. One new structure which is reported in the paper is a revised form of the Costas tracking loop using analytic signals working to zero frequency in base-band translated GPS. This has some special properties in that the filtering requirements are greatly simplified in this form.

Results showing the reconstruction of translated GPS carrier wave signals with dynamic host vehicle motions of up to 80g are shown together with the Costas tracked equivalent waveforms. This structure is viewed as a narrow band tracking filter.
[Back to Program]