NavX -NCS A Multi-Constellation RF Simulator: System Overview and Test Applications

Transcription

1 NavX -NCS A Multi-Constellation RF Simulator: System Overview and Test Applications Markus Irsigler, Bernhard Riedl, Thomas Pany, Robert Wolf and Günter Heinrichs, IFEN GmbH BIOGRAPHY INTRODUCTION Markus Irsigler received his diploma in Geodesy and Geomatics from the University of Stuttgart, Germany, and a doctoral degree (Dr.-Ing.) in Aeronautics and Astronautics from the Bundeswehr University Munich. He joined IFEN GmbH in 2007, where he currently works as product manager and systems engineer. He is involved in development, product management, marketing and support activities for IFEN s multi-frequency and multiconstellation GNSS signal generator NavX -NCS. The NavX -NCS is the first product in a new generation of combined Galileo/GPS products from IFEN GmbH. It has been designed to serve single-frequency mass market applications like A-GPS testing as well as multifrequency applications such as GNSS receiver development or RTK surveying. The NavX -NCS is a complete and flexible Galileo/SBAS/GPS laboratory test equipment which provides a reliable, stable and repeatable platform for a variety of test applications. Bernhard Riedl received his diploma in Electrical Engineering and Information Technology from the Technical University of Munich. Since 1994 he has been concerned with research in the field of real-time GNSS applications at the University FAF Munich. In 2006 he joined IFEN GmbH, where he is currently working as product development manager. The NavX -NCS has been developed in co-operation with WORK Microwave GmbH and is available in the form of two different product variants. The NCS Standard is a single-frequency signal generator whereas the NCS Professional supports all relevant GPS and Galileo frequency bands. Product shipping to the first customers has started in January By that time, the NCS was the first commercial combined Galileo/GPS signal generator on the market. Dr. Thomas Pany has a PhD in geodesy from the Graz University of Technology and a MS in Physics from the Karl-Franzens University of Graz. From , he was assistant professor (C1) at the University FAF Munich, and he is currently senior research engineer at IFEN GmbH. His major areas of interest include GNSS receivers, software radio technology, navigation signal processing and GPS science. Dr. Robert Wolf holds a degree in Aerospace Engineering from the Technical University of Munich and a Dr.-Ing. from University FAF Munich. He joined IFEN GmbH in From he was head of integrity systems department at IFEN GmbH. Since 2008 he is the head of Navigation Products Department, responsible for NCS development. Dr. Günter Heinrichs received a Dipl.-Ing. degree in Communications Engineering from the University of Applied Science Aachen in 1988, a Dipl.-Ing. degree in Data Processing Engineering and a Dr.-Ing. degree in Electrical Engineering from the University Paderborn in 1991 and In 1996, he joined MAN Technologie AG in Augsburg, Germany, where he was responsible for GNSS receiver development. As from 1999, he was head and R&D manager of MAN Technologie s satellite navigation department. In 2002, he joined IFEN GmbH, Poing, Germany. Since 2008, he is head of the Customer Applications department, responsible for all customer, marketing and sales related activities. SYSTEM OVERVIEW Key features. The NavX -NCS is a multi-channel, multifrequency GNSS RF constellation simulator capable of simulating combined GPS/Galileo/QZSS/SBAS signals. It provides reliable and repeatable generation of all relevant GNSS signals (GPS L1/L2/L5, Galileo E1/E6/E5ab, SBAS L1/L5 and QZSS L1) out of one single chassis. The NavX -NCS chassis features a 19 rack mounting option allowing easy integration into standard laboratory equipment and a very compact design (maximum of two height units only!). Figure 1: NavX -NCS signal generating unit/control PC.

2 The signal generating unit consists of one or several RF modules, each of them featuring 12 channels. One NavX -NCS chassis can be equipped with up to 9 RF modules, so that it can be upgraded to a maximum number of 108 channels per chassis. Each RF module is fully configurable in terms of frequency and signal modulation (frequencies can be individually assigned to any RF module). Together with the large amount of available channels, this architecture ensures full flexibility for a variety of test applications. Interfaces and Hardware Upgrade Options. NavX - NCS devices feature various input and output interfaces like Ethernet connection, an input for external reference oscillators, 1 PPS, a hardware trigger input or an output for the internal clock signal. In addition to the standard RF output at the front of the NavX -NCS, an additional monitoring port is available at the rear panel. These input and output interfaces offer the potential to integrate the NavX -NCS to other hardware or to integrate it into existing test environments. NavX -NCS systems are delivered with a control PC, which is used for simulation configuration and interactive control (see section NavX -NCS Software ). The control PC is basically an external one. However, depending on the desired NavX -NCS product variant, an embedded PC can be selected instead of an external one. Further upgrade and configuration options comprise the implementation of additional RF modules or RF outputs, or the implementation of an internal noise generator. Product variants. The customers can select among two different NavX -NCS product variants, namely NavX -NCS Standard and NavX -NCS Professional. Figure 2: NavX -NCS Standard (multi-constellation, single-frequency, 1HU). The NavX -NCS Standard device is a single-frequency constellation simulator covering GPS/Galileo/GLONASS signals on L1/E1/G1. It supports a maximum of 36 channels. Since the NavX -NCS Standard is fully compatible with the testing requirements specified in the current 2G and 3G mobile phone standards, it is suited to set up an integrated single-frequency, multi-constellation test environment for GNSS receiver production testing. It comes in a single height unit chassis. Figure 3: NavX -NCS Professional (multi-constellation, multi-frequency, 2HUs). The NavX -NCS Professional is a multi-frequency multi-constellation simulator covering GPS, Galileo, QZSS, SBAS and GLONASS signals on L1/E1, L2, E6, and L5/E5ab. It supports a maximum of 108 channels and a multitude of hardware and software upgrade options. Due to the large amount of supported hardware channels and GNSS signals, it is ideally suited to set up an integrated multi-frequency, multi-constellation test environment for GNSS receiver development and testing. It comes in a two height unit chassis. Latest developments. During the past few months, the simulation capabilities of the NCS have been upgraded significantly. Major enhancements include: CBOC for Galileo E1. So far, the NCS supported only the BOC modulation for the Galileo E1 signal. It is now possible to choose between BOC and CBOC modulation. Internal noise generator. A noise source has been integrated into the NCS chassis allowing the simulation of a dedicated (and configurable) noise floor. The noise generator can be configured by the control software. External noise sources are not required any more. SBAS functionality. SBAS functionality include configuration of SBAS satellites in terms of satellite locations, PRNs, and service areas. The SBAS signals provide range corrections for GPS satellites. EGNOS, WAAS and MSAS are available as preconfigured options. SBAS signals can be generated using both the L1 and the L5 carrier frequency. QZSS capability. Configuration options include the definition of QZSS orbits, configuration of the satellite clock behaviour and configuration of the signal power levels. NCS CONTROL CENTER SOFTWARE The NavX -NCS is configured and controlled by the NCS Control Center, a flexible and powerful software which allows intuitive simulation configuration and interactive control. It provides access to all important simulation parameters and features all logging, monitoring and visualization options necessary to evaluate and analyze each simulation run. Figure 4 illustrates the general appearance of the graphical user interface of the NCS Control Center software.

3 Software Features. In conjunction with the NavX - NCS Professional device, the NCS Control Center software supports the generation of GPS, Galileo or GLON- ASS simulations (at all relevant frequency bands), the incorporation of satellite-based augmentation systems like EGNOS, WAAS, MSAS or QZSS and the generation of assisted GNSS data. The software supports definition and simulation of user trajectories, so that either static or dynamic simulations can be performed. In addition to select from a pool of pre-configured simulations, the user can configure its own simulation by adjusting the relevant simulation parameters. Some important simulation configuration options are described in the following paragraphs. Space Segment. Basic features for the definition of satellite orbits are the capability of importing precise ephemeris or almanac data (e.g. SP3, RINEX or YUMA) or the definition of orbit parameters per satellite. Entire satellite constellations can be generated by using the Single-Step Constellation Generation function. One important feature is the capability to generate combined GPS/Galileo/SBAS/QZSS constellations. Figure 4 illustrates such a combined constellation using three orbit planes for Galileo and six for GPS. In addition, two geostationary and 2 QZSS satellites were incorporated into the constellation. Other simulation configuration options related to the space segment are the definition of satellite clock errors and the definition of hardware group delays. Signal Propagation and User Environment. Modelling signal propagation characteristics basically comprises the definition of tropospheric and ionospheric properties as well as the definition of a multipath environment. Besides the capability of defining simple multipath conditions like the presence of ground or fixed-offset multipath, the NavX -NCS Control Center software can be upgraded with an Advanced Multipath Simulation extension. Other simulation configuration options related to the user environment include the definition of user antenna patterns or the use of elevation masks. All NavX -NCS systems support the simulation of user trajectories. User trajectories can be imported via a file interface. The file format is proprietary but documented, so that in addition to a set of pre-installed trajectories, the user can generate its own trajectories or convert existing trajectory data into the desired format and import it into the Control Center software. It is also possible to import externally recorded GPS tracks (like NMEA data) into the Control Center and replay them. The software also provides different types of closed user trajectories, for which the dimensions (e.g. horizontal extent) and the user dynamics (e.g. along-track velocity) are configurable. These trajectories do not have a starting nor an end point, so that a moving receiver is simulated throughout the entire simulation run. The following motion types are supported: Circular and elliptic motion Helical motion (circular motion with height changes) Motion along a lemniscate (8-shape) Motion along a Hypotrochoid Motion along a Epicycloid Motion along a Lissajous curve Motion along loops of different shapes Logging options. Simulation data can be logged in the form of RINEX observation (ranging data) or RINEX navigation files (orbit data). In addition, the content of the generated navigation messages can be logged as well. Simulation modes. The NavX -NCS system supports four different simulation modes. The standard mode of operation covers the case where the signal generator is connected to the control PC and the start of a simulation is initiated by the user. In this case, the simulation starts immediately after calling the related GUI-function. Figure 4: Simulation of a combined GPS/Galileo/ SBAS/QZSS constellation with 2 geostationary and 2 QZSS satellites. Three different mechanisms are available to configure/adjust signal power levels. A detailed description on these configuration options is given in the section Signal Power Levels. Since every signal generator is equipped with a trigger input, the start of a simulation can also be triggered by an external signal. In this case - which is the second mode of operation - the simulation does not start immediately after the user is calling the Start function, but waits for the trigger signal instead. The other two simulation modes do not require the signal generator to be connected to the control PC. They are software-based simulations that run either in real-time or as fast as possible and can be used to record simulation data. The performance of the latter simulation mode (i.e. the time the PC needs to process a simulation of a certain duration) only depends on the PC environment.

4 The following sections provide examples for specific test applications and test cases that the NavX -NCS is able to cover. The main focus is on the different simulation options for signal power levels. SIGNAL POWER LEVELS The NCS Control Center software provides three different approaches to configure and simulate the power levels of the satellite signals. The first method is to specify the satellite s transmit power, the second is to specify the received power level on the ground. Finally, power levels can be assigned by using a file interface, which is a very flexible and powerful tool to meet the needs of special test cases. The following diagrams illustrate the received power level for a dedicated SV as a function of time. Such power level profiles can be easily implemented by using power level files. Figure 5 visualizes the content of a power level file which configures a stepwise increase of the received power level for PRN1 from -160 dbm up to -140 dbm over a time period of 2000s. The power level is increased by 1 db at each step. After each step, the new power level is maintained for 100s. Transmit Power. The transmit power can be configured individually per frequency band. It determines the signal power radiated from the satellites (EIRP, Equivalent Isotropic Radiated Power). By using this mode, the NavX -NCS accounts for the free-space loss, so that the simulated power levels (measured at the signal generator s RF output) are different for each satellite due to the different elevation angles of the space vehicles. This mode provides the most realistic simulation of signal power levels. Received Power Levels. In addition to the configuration of the satellite s transmit power, signal power levels can also be defined in terms of received power, that is the power a user would measure at the output of his antenna. Received power levels can also be configured individually per frequency band. In this case, the simulated power levels (measured at the generator s RF output) are the same for all satellites transmitting in this frequency band. Moreover, a connected receiver will show more or less the same C/N 0 s for all satellites. Figure 5: Stepwise change of signal power levels as a function of time. Power level files can also be used to create power level profiles. Such a profile is illustrated in Figure 6. During the simulation, the signal power smoothly decreases by 15 db, remains constant for 100s, is increased again by 10 db and remains constant at a value of -135 dbm for the rest of the simulation. Power Levels from File. In contrast to the definition of the satellite transmit power or the received power levels, the third approach of configuring simulated power levels is much more flexible. It is based on a proprietary file format containing power level information individually for each satellite. The file format is open and documented, so that power level files can be easily created by the user. Using power level files provides the following advantages: Power levels can be configured individually per satellite Power level information can be combined with a time stamp, i.e. it is possible to configure power level changes over time (individually per satellite) It is possible to define periods of time where the signal is switched off temporarily (simulation of signal obstructions) As a side-effect, it is also possible to restrict the simulated constellation to a subset of satellites (satellites for which no power level information is given are not simulated at all) Figure 6: Power level profile defined by using ramps and times of constant signal power. Figure 7 illustrates the case where the satellite signal is temporarily switched off to simulate a signal loss due to signal obstruction. The signal is switched off for 150s. This period of time is visualized by the red portion of the power level profile.

5 The capability of configuring customized power level profiles forms a very flexible tool to serve the needs for dedicated test applications. In the field of receiver testing, for example, customized power level profiles can be used to test the receiver s capability to acquire and track weak signals or to deal with amplitude scintillation. In the next section, such an application is discussed in more detail. TEST APPLICATION Figure 7: Power level profile which contains information about temporary signal losses (caused by signal obstruction). In addition to the simulation of power level steps, ramps and signal-off times, power level files can also be used to simulate more complex effects. One example is amplitude scintillation, a phenomenon which is typically caused by ionospheric irregularities and which may cause rapidly changing power level variations. Especially during times of increased solar activity, ionospheric scintillation can be observed frequently. Figure 8 illustrates the power level profile for a dedicated SV showing the effects of amplitude scintillation. The signal is configured to have a constant power level of -130 dbm at the beginning of the simulation. After 200 seconds, the scintillation effects are switched on and are present for the next 200s. Figure 8: Power level profile showing the influence of amplitude scintillation. Power level data extracted from a power level file can be biased by applying a constant offset to all file contents. Thus, the content of a power level file does not need to have a dedicated unit (like dbm or dbw), because any values listed in the file can be transformed to reasonable power levels by just applying a suitable offset. By using this mechanism, it is also possible to replay recorded C/N 0 data. Based on the power level files presented in the previous section, the ability to acquire and track weak signals has been evaluated for four commercial GNSS receivers, namely a NovAtel OEMV GPS receiver, a ublox GPS receiver using the ANTARIS 4 chipset, a Javad Delta receiver and IFEN s NavX -NSR, a software-based GPS/Galileo receiver. The NavX -NSR is the first interactive GPS/Galileo L1/E1 software receiver which consists of a single frequency RF front-end (L1) and desktop software for baseband and navigation processing. The frontend provides 10 MHz RF bandwidth and is connected via a USB 2.0 Hi-Speed interface. The receiver is able to work in real-time or in postprocessing mode on previously recorded IF sample files. The NSR software is composed of the following key elements: Processing core Real-time USB IF sample interface Graphical User interface/configuration manager Use of a XSense IMU for GNSS/INS integration Application programming interface (API) The latter is one of the main advantages of a softwaredefined GNSS receiver in comparison to traditional hardware based receivers. Our API gives the user the possibility to gain access to all stages of a typical GNSS receiver. These are namely the IF Sample API, which provides neat access to the raw sample data packets before any signal processing has taken place. Such an interface is useful for applying a digital filter to the signal or to add noise or other signal types for jamming and interference simulation. The Baseband API enables users to integrate their own acquisition and tracking routines. In this mode the NSR does not run its built-in acquisition. Furthermore, the Baseband API provides full control and access to the tracking procedures and parameters of each receiver channel in the NSR. This is probably the most interesting API giving the advanced user a wide range of possibilities from just monitoring the tracking loop behavior up to creating its own implemented algorithms for the delay lock loop (DLL), phase locked loop (PLL) and frequency locked loop (FLL). Even vector-delay-lockloop implementations are possible to test various GPS/INS integration schemes (from loosely to ultratightly coupling). The Navigation extension API offers the raw observations, i.e. pseudorange, carrier phase, code phase and navigation data are available. It enables to

6 implement capabilities for navigation processing or any other kind of post-processing and positioning algorithms. Figure 9 provides an overview of the receiver s main dataflow and the API extensions. Figure 9: NavX -NSR dataflow and programming interfaces. Test Setup. The setup (cf. Figure 10) comprises the NavX -NCS as signal source for generating a repeatable satellite constellation and provides an RF output which is attenuated by 20 db and then fed to an LNA (low noise amplifier) to simulate the amplification of a typical antenna. After the LNA, the signal is identical to a received GNSS signal as measured at the antenna output. This LNA boosts the signal by 15 db before splitting it into several paths and feeding to the receivers. Figure 10: Test setup with NavX -NCS as signal source. The LNA mimics an integrated antenna LNA and is required to allow a fair comparison between the different receiver frontends. According to Friis formula the total noise figure of a navigation system (antenna plus receiver) is mostly determined by the integrated antenna LNA. Those antenna LNAs have noise figures of around 1-4 db. The noise figures of the receiver frontends are often much worse and might achieve values up 5-15 db. Doing the experiment without an LNA would overweight the influence of the frontend noise figure. The performance tests have been carried out under specific test conditions. The simulation has been restricted to generate GPS signals for a static user. A dynamic scenario has not been considered. Moreover, the receivers under test were configured to operate in specific receiver/processing modes: The ublox receiver can be operated in different modes ( High Sensitivity, Fast Acquisition, Normal ). Since the tests aimed at determining the capability of the receivers to work under weak signal conditions, the High Sensitivity mode has been selected for the ublox receiver. The NovAtel receiver does not provide modes of operation similar to those provided by the ublox receiver. However, the signal tracking process can be influenced by adjusting the tracking loop parameters which change the loop s response to signal dynamics. Available presets are foot, land and air. As it turned out that the test results do not differ significantly when using these presets, the land setting has been used for the tests. Since no configuration options similar to those provided by the ublox or the NovAtel receiver were available, the standard receiver settings have been used for operating the Javad receiver. The NavX -NSR was operated in two different modes. For a first test run, the standard tracking mode was used. For a second run, a vector tracking implementation was used. The latter approach is expected to maintain lock even in case the signals are very weak. The main difference between both tracking modes is that the standard tracking mode estimates code phase and Doppler for each channel individually, whereas vector tracking merges all channels by directly estimating velocity and and position. Vector tracking requires a lot of processing power, which is a scarce resource on many receiver platforms. It is important to note that the test results presented below can only be seen as a snapshot, because they are only valid under these test conditions (i.e. specific receiver settings, simulation of static user position). Changing the receiver settings or performing a simulation with a moving receiver may lead to different results. However, the main objective of this paper is to present an application example for the NCS rather than performing a competition between different GPS receivers. Therefore, the test results are based only on the assumptions presented above. In order to evaluate the performance of the different receivers, the power level profile illustrated in Figure 11 has been simulated.

7 This diagram illustrates the C/N 0 s of 7 GPS signals as obtained from the NovAtel OEMV receiver. The underlying power level profile is indicated as the grey dashed line. The weak signal tracking performance (test case 1) is defined in the falling part of the power level profile and determined by the C/N 0 at which the receiver cannot track the signals any more (loss of lock). Reacquisition and cold start performance (test cases 2 and 3) are defined in the rising part of the power level profile and determined by the C/N 0 at which the signals can be reacquired after having lost lock (warm start condition, test case 2) or after having performed a cold start (cold start condition, test case 3). Figure 11: Power level profile used to evaluate the performance of different GNSS receivers under weak signal conditions. This profile is applied to all simulated satellites, so that all signals undergo the same power level changes. At the beginning of the simulation, the power levels remain constant for 5 minutes allowing the receivers to acquire the signals. Then, during a period of 5 minutes, the power levels are decreased by 40 db, remain constant for 3 minutes and then increase again. The profile illustrated in Figure 11 has been used to cover 3 different test cases: High-Sensitivity Performance Results. Figure 13 illustrates the results obtained from test case 1 (Weak Signal Tracking Performance). All C/N 0 profiles obtained from all receivers under test are collected in this diagram. The Javad receiver is the first one to lose lock at a C/N 0 of about 33 db-hz. The NovAtel receiver maintains lock for C/N 0 s larger than ~27 db-hz. The ublox receiver shows a very good performance keeping lock down to C/N 0 s of about 10 db-hz. Due to its vector tracking implementation, the NavX -NSR maintains lock even if the C/N 0 drops below 10 db-hz. 1. Weak Signal Tracking Performance. Evaluation of the receiver s ability to maintain lock when the signal power levels are decreasing. 2. Reacquisition Performance. Evaluation of the receiver s ability to reacquire the signal when the power levels are increasing again (warm start condition). 3. Cold Start Performance. Evaluation of the receiver s ability to acquire the signal under weak signal conditions (cold start condition). The latter test case is put into effect by resetting the receivers (initiation of a cold start) when the simulated power levels reach their minimum. The criteria for evaluating test cases 1 and 3 are illustrated in Figure 12. Figure 13: Summary of results for test case 1 (Weak Signal Tracking). Test case 2 (Reacquisition Performance) covers the case that a signal has to be reacquired after it had been lost. The test results are illustrated in Figure 14. In this case, the ublox receiver reacquires the signals as soon as their C/N 0 s exceed ~25 db-hz. Also the Javad receiver shows a fairly good reacquisition performance (reacquisition limit: ~33 db-hz). The NovAtel receiver needs rather strong signals to reacquire them successfully (reacquisition limit: ~39 db-hz). The reacquisition performance of the NavX -NSR cannot be evaluated, because due to the vector tracking implementation, the receiver did not lose lock when the signal power levels were decreased. Figure 12: Simulated power level profile vs. recorded C/N 0 as a function of time.

8 Figure 14: Summary of results for test case 2 (Reacquisition Performance). The results of the cold start performance (test case 3) are illustrated in Figure 15. The NSR is the first receiver that is able to acquire the signals after a cold start at C/N 0 s of about 31 db-hz, followed by the ublox receiver which starts tracking at ~37 db-hz. The Javad receiver acquires the signals at C/N 0 s of about 40 db-hz. The NovAtel receiver shows a rather poor acquisition performance requiring signal power levels that translate to C/N 0 s of at least 47 db-hz. Figure 16: Reacquisition performance (test case 2) for the NavX-NSR using both the standard tracking mode and the Vector Tracking mode. It has already been mentioned that the test results strongly depend on the simulation conditions and the receiver settings or processing modes. One example for such a dependency is illustrated in Figure 17 and Figure 18 showing the reacquisition performance for the ublox receiver for two different simulation runs. During the first run (illustrated in Figure 17), the signal power level was simulated such that the C/N 0 observed by the receiver was at about 49 db-hz. During the second simulation run (see Figure 18), the initial signal power level was decreased resulting in an initial C/N 0 of about 42 db-hz. This leads to an earlier loss of lock during the second simulation run relative to the GPS time scale (x-axis in Figure 18) and the time difference between the loss-of-lock event and a point in time where the signal has a dedicated power level increases as well. This in turn increases the time the receiver needs to reacquire the signal, and consequently, the reacquisition limit (i.e. the C/N 0 at which the signal can be reacquired). Figure 15: Summary of results for test case 3 (Cold Start Performance). So far, the test results obtained from the NSR are only based on the Vector Tracking mode. Due to the use of this processing mode, the NavX -NSR is able to maintain lock on very weak signals even when the C/N 0 drops below 10 db-hz. If the Vector Tracking mode is switched off and the standard tracking mode is used instead, the ability to track weak signals is still good. Figure 16 illustrates the results for test case 2 (Reacquisition Performance) for the NavX -NSR using both the standard tracking mode and the Vector Tracking mode. While vector tracking ensures continuous tracking during the entire simulation run (red curve), the receiver lost lock when the C/N 0 dropped below approximately 7-8 db-hz but was able to reacquire the signals when the C/N 0 s reached ~22 db-hz (blue curve). Figure 17: Reacquisition performance of the ublox receiver in case of high initial C/N 0 s.

9 Figure 18: Reacquisition performance of the ublox receiver in case of decreased initial signal power levels. SUMMARY The NavX -NCS is a reliable GNSS signal source providing a simulation environment which ensures repeatable signal generation under controlled simulation conditions. It can be used as a stand-alone test device or integrated into more complex test environments and is a suitable and valuable tool to realize and implement a variety of test applications. One such application has been presented in this paper, namely the evaluation of the tracking performance of different commercial GPS receivers under weak signal conditions.