Stanford GNSS Monitor information on the signals in each of Station Antenna these frequencies. These signals, then, lie in the frequency band of GPS and gNSS Galileo signals. The Compass navigation signals are code division multiple access (CDMA) over China signals similar to the GPS and Galileo With the launch of its first signals. They use binary or quadrature phase shift keying (BPSK, QPSK, respec- the Compass MEO middle-earth-orbiting tively). Further, SU observations and (MEO) Compass satellite, analysis indicate that the codes from the satellite Codes China has put forth its current Compass M-1 are derived from Gold codes. GNSS entry. The key to Statements from Chinese sources using and understanding indicate that the system will provide the performance of the at least two services: an open civilian Compass M-1 navigation service and a higher precision military/ authorized user service. signals is revealed by its The Compass-M1 satellite repre- spread spectrum code. This sents the first of this next generation of article by a team of Stanford Chinese navigation satellites and differs significantly from China’s previous Bei- University researchers dou navigation satellites. Those earlier presents the spread satellites were considered experimen- spectrum codes being tal, and most were developed for two- broadcast by this satellite. dimensional positioning using the radio determination satellite service (RDSS) concept pioneered by Geostar. Compass M-1 is also China’s first graCE XiNgXiN gaO, alaN ChEN, MEO navigation satellite. Previous Bei- shErMaN Lo, DaviD DE LorENzO, PEr ENgE dou satellites were geostationary and Stanford UniverSity only provide coverage over China. The global implications of this satellite and the new GNSS it represents makes the M-1 satellite. Observations by CNES, Compass to an integrated GNSS receiver satellite of great interest to navigation SU, and other researchers indicate that without additional expensive hardware experts. the current satellite is only broadcast- or processing. Moreover, the rapid prog- n April 14, 2007 (local time), ous researchers around the world, rent design is have a system comprised of The rapid manner in which research- ing on three of the frequencies (E2, E6, ress of the Compass development (and China launched the Compass including Stanford University (SU), have 30 medium earth orbit (MEO) satellites ers have already trained their instru- E5b). the current state of the Galileo program) M-1 satellite. This satellite rep- been interested in examining the navi- and 5 geostationary orbit (GEO) satel- ments onto the satellite proves this point. To the best of the authors’ knowl- offers the intriguing possibility that the Oresents the first of a new global gation signal of this system. lites. The MEO satellites will operate in For example, Centre National d’Études edge, no observations of Compass E1 system may become operational before navigation satellite system (GNSS) that To understand its effects and to six orbital planes to provide global navi- Spatiales (CNES, the French space agen- broadcasts have been made. It also Galileo. is planned to have a total of 35 satel- develop receivers that can track the gation coverage. cy) published an informative overview of appears that the Compass satellite is not As such, great motivation exists lites. Unlike prior Chinese navigation signal, one must understand the sig- Compass will share many features their observations of the Compass-M1 continuously broadcasting navigation for understanding Compass and how satellites, Compass M-1 broadcasts in nal structure being used. In the case in common with GPS and Galileo, pro- signals a month after its launch in the messages on the other three frequencies; it may be properly and cost-effectively L-band, using signal structures similar of Compass, this means determining viding the potential for low cost integra- May/June issue of Inside GNSS. we have occasionally observed unmodu- integrated into a GNSS receiver. On the to other GNSS systems and sharing fre- and understanding its spread spectrum tion of these signals into a GPS/Galileo/ The interest has resulted in signifi- lated or continuous wave (CW) signals flip side, the signals may pose a source quencies near to or overlapping those of codes. This article will present the Com- Compass receiver. These commonali- cant basic information on the Compass in those bands. Apart from these basic of interference and degrade the perfor- GPS, Galileo, and GLONASS. pass codes and provide an overview of ties include multiple frequencies, signal observations of the Compass M-1 sig- mance of GPS or Galileo. Interference Frequency Modulation Type The addition of another GNSS, how our team at Stanford determined structure, and services. nal structure, little information has been with and degradation of GPS/Galileo particularly one that will broadcast in these. According to International Tele- 1589.74 (E1) QPSK(2) published on the actual codes. performance are possibilities if interop- the same frequency bands as GPS and communication Union (ITU) filings 1561.1 (E2) QPSK(2) The similarity in frequency, signal erability was not a driving concern in Galileo, both excites and intrigues the Compass Overview by China, Compass will broadcast on 1268.52 (E6) Q/BPSK(10) structure, and services with GPS and the signal design. GNSS community. Such a system has the The Beidou ( ) or Compass naviga- four frequencies centered at 1590 MHz, 1207.14 (E5b) BPSK(2), BPSK(10) Galileo makes Compass a tantalizing This latter possibility, of course, potential to introduce benefits — as well tion satellite system (CNSS) is China’s 1561 MHz, 1269 MHz, and 1207 MHz prospect for GNSS users. These simi- concerns military users as well because TABLE 1. Compass Frequencies and Modulation as concerns — for GNSS users. Numer- entry into the realm of GNSS. The cur- (rounded). table 1 provides general larities could allow for the addition of Compass overlays the GPS M-code and 36 InsideGNSS july/august 2007 www.insidegnss.com www.insidegnss.com july/august 2007 InsideGNSS 37 Compass E2 Signal Spectrum will help to develop prototype GPS/Gali- we collected additional data in June in order to obtain longer -160 leo/Compass receivers and help identify data sets with which to work. ways to best use the new signals together -180 with other planned or existing GNSS Frequency Domain Plots -200 signals. Rather than repeat the excellent spectrum plots from the CNES article mentioned earlier, this section will show the spectrum -220 Data Collection Equipment for each Compass signal without averaging. Figure 2 shows the -240 Use of a high gain antenna greatly aids unaveraged E2 signal spectrum from one of our data sets. The the effort to assess the Compass signal main lobe and the first side lobes of the 2 MHz chipped signal -260 and determine its navigation code. For are clearly visible even without averaging. -280 data collection, we used the Stanford An L1 signal from a nearby GPS satellite can also be seen ower Spectral Density (dB/Hz) GNSS Monitor Station (SGMS). The in this plot as well as narrowband signals on 1549 MHz. Figure P -300 SGMS has a 1.8-meter steerable para- 3 shows the unaveraged Compass E5b signal spectrum from -320 bolic dish antenna with an L-band feed. another data set. The main lobe of the BPSK(2) is clearly visible, -340 The system was developed to provide and the BPSK(10) main lobe can also be made out. As expected 1530 1540 1550 1560 1570 1580 1590 an on-demand capability for observing in this frequency band, we also see strong narrowband interfer- Frequency (MHz) GNSS signals. ence from distance measuring equipment (DME). FIGURE 2 Unaveraged spectrum of Compass M-1 E2 This antenna provided many of the Figure 4 shows the unaveraged E6 signal spectrum with the measurements seen in a previous article main feature being the main lobe of the QPSK(10) signal. Also Compass E5b Signal Spectrum in the May/June 2006 issue of Inside visible is an as yet unidentified 1 MHz–wide transmission cen- -160 Portable Ground Station Set Up GNSS to which the authors contributed, tered around 1257 MHz. -180 including some of the data used to deter- mine and validate the GIOVE-A codes. Deriving the Compass Codes -200 (See the citation for the article by S. Lo The main challenge to revealing the PRN code sequence is -220 et alia in the “Additional Resources” sec- the low signal-to-noise ratio (SNR). With an omnidirectional tion near the end of the article.) antenna, the received signal power is on the order of 10-16 watts. -240 We collected data for the analysis Even with the 1.8-meter dish antenna and high-quality low -260 described here using a vector signal ana- noise amplifiers (LNA), the received C/No is still roughly 65-70 lyzer (VSA). One change from our past dB-Hz (assuming a transmit power of 30 W). This still does not -280 set-up was to make the ground station provide enough gain to pull the code chips out of the noise, and ower Spectral Density (dB/Hz) P -300 (including antenna controllers) portable. the code is not directly visible in the time domain. -320 This was necessary because the original In order to decode the PRN code sequence, we need to ground station facility is being renovat- process the data to boost the signal above the noise floor. The -340 1180 1190 1200 1210 1220 1230 1240 ed. Accompanying photos show the SU main concept is to stack multiple periods of the PRN sequence Frequency (MHz) antenna and ground station. together so that the noise will be averaged. To achieve this, we Data sets on all three observed Com- need to determine the code period, wipe off Doppler offset, FIGURE 3 Unaveraged spectrum of Compass M-1 E5b pass frequencies were taken on multiple adjust the initial phase shift and demodulate the secondary days.