GPS World - January 2008 - (Page 27)

Receiver Design | SYSTEM DESIGN & TEST receiver’s front end. Otherwise, the receiver’s performance can be degraded in unpredictable ways. If an LNA is used, its gain and noise ﬁgure must be noted, and the generator’s power settings may need to be adjusted accordingly. The details of this adjustment are discussed in the “Subtleties” section. As for the attenuator, it must be included in the setup if the output signal power range of the GPS simulator does not go low enough to cover the signal range of interest. Many GPS simulators have a lower signal power limit of around –150 dBm, whereas the high-sensitivity regime is commonly assumed to refer to signal levels of –155 dBm and below. e eﬀect of the attenuator (and any other passive components between the output of the simulator and the RUT or LNA inputs, including cables, splitters, connectors, DC blocks, and so on) is treated as follows. Mathematically, the noise ﬁgure of a passive component is equal to its insertion loss. us the C/N0 at the output of a 12 dB attenuator is exactly 12 dB lower than the C/N0 at its input. Since, as explained previously, the convention is to refer all C/N0 values to a standard (ﬁxed) thermal noise ﬂoor density, any inline loss can simply be treated as an eﬀective reduction in signal power, without regard to the noise ﬂoor. us if we assume the simulator in Figure 1 is set to –130.0 dBm, the attenuator has a measured insertion loss of 12.1 dB, and the four-way splitter has a measured insertion loss of 6.3 dB (6.0 dB power split plus 0.3 dB splitter loss), the eﬀective signal level delivered to each RUT is –130.0 – 12.1 – 6.3 = –148.4 dBm. Cables, RF adaptors, DC blocks, and any other interposing passive components are treated in a similar fashion. A ﬁnal detail in the hardware setup that can signiﬁcantly aﬀect testing results is the reference oscillator in each RUT. e setup shown above assumes each receiver is running on its own standard clock. If an external reference is used with any of the receivers, or if any receiver has an unusually good frequency reference, the performance characteristics of the reference should be noted explicitly in the test results. is is important because using an exceptionally good reference clock (e.g. atomic frequency standard) will produce unrealistically good receiver performance that may be unattainable under real-world operating conditions. As best practice, the characteristics of each receiver’s frequency reference should always be stated explicitly. 15 10 Generator power setting (dB) 5 0 -5 -10 0 5 10 15 20 25 30 35 Scenario time (min) 40 45 50 55 p FIGURE 2 Power profile used to measure tracking sensitivity Test Cases Here we describe the power proﬁles used to evaluate receiver sensitivity in tracking and reacquisition, and the procedure to estimate acquisition sensitivity and related ﬁgures of merit. For data sheet validation tests, the RUT set should normally include best-of-breed hardware, that is, factory-made evaluation kits with latest available ﬁrmware. Since each manufacturer’s data sheet typically states the most aggressive sensitivity numbers that can be reasonably justiﬁed, this choice represents www.gpsworld.com the best way to validate the various data sheet numbers. On the other hand, it is also useful to include embedded devices when feasible. While the evaluation kits provide an idea of maximum achievable performance, embedded implementations (standalone personal navigation devices, handheld computers, etc.) should be used for more realistic measurements. Tracking Sensitivity. e power proﬁle used to measure tracking sensitivity is shown in FIGURE 2. Several features are worth noting. First, the RUT is driven with maximum signal power for at least 12.5 minutes to allow it to decode a full almanac, plus a small additional amount of time to accommodate the acquisition process. Fifteen minutes is a reasonable value here, but can be shortened considerably if there is a reliable way to initialize the RUT directly with the correct almanac. Second, the transition from full power to the highsensitivity region is rapid but not abrupt. e idea is to ease the receiver into high-sensitivity mode as quickly as possible without triggering any loss-of-lock routines in the navigation ﬁlter. Since the nature of any such routines cannot, in general, be known in advance, it is important to monitor the receiver during this transition to make sure no satellites are lost before the power level settles to the ﬁrst step in the high-sensitivity region. Finally, each power level must be maintained long enough to correctly ascertain receiver performance at that level. Experience shows that if the interval is too short, it may be impossible to tell whether the RUT loses lock on a given satellite because the power level changes, or because its tracking loops really lose the satellite signal. To that end, a reasonable duration for each step is 240–300 seconds. Once the RUT has been driven with a power proﬁle as shown above, its output (typically in the form of NMEA logs, though other formats may also be used) is examined to determine various aspects of tracking performance. ese include number of satellites versus power level (FIGURE 3) and position availability (FIGURE 4). Additionally, it is informative to plot position error versus power level, as shown in FIGURE 5. Since receivers can be optimized for diﬀerent aspects of performance, a plot of this type January 2008 | GPS World 27 http://www.gpsworld.com

Table of Contents Feed for the Digital Edition of GPS World - January 2008

GPS - January 2008 Contents Out in Front Expert Advice The Money-Go-Round u-Nav Latest Acquisition Apples to Apples Global SBAS 2008 GPS Receiver Survey Advertisers Index & Company Directory The Manufacturer's Road Year of the Who Working Indoor Up and Down Good, Better, Best Marketplace Classifieds Seen + Heard

GPS World - January 2008

For optimal viewing of this digital publication, please enable JavaScript and then refresh the page. If you would like to try to load the digital publication without using Flash Player detection, please click here.