To utilize signals transmitted through or reflected off walls, indoor GPS receivers require sensitivities 20dB below that of older, conventional receivers. At these levels, jamming may come from low-power sources in the room, in the equipment, or in the chipset itself. The authors address these "inside" sources--arising from co-location of antenna, baseband, and host electronics--producing emissions that readily exceed interference levels, and they offer some design considerations for indoor locatability.
Until recently, jamming investigations and mitigation efforts focused on outdoor signal levels and the effects of relatively high-power but distant emissions. With the advent of indoor location, emissions much closer to the receiver have become a problem. These include clock- and bus-related harmonics, frequency plans; serial communications and peripherals; proximity issues (obtaining enough radio frequency isolation); power supplies; and cellphone transmission bursts, blanking the receiver.
With information technology (IT) equipment pervading the modern world, even compliance with electromagnetic compatibility (EMC) regulations does not prevent interference with indoor GPS. These present an additional form of "inside" enemy--inside the building. Even with the cleanest design and layout, hardware designers likely cannot entirely eliminate these interferences.
We have discovered that industry outside GPS itself is not familiar with dealing with signals in the 1[0.sup.-18] watts region, and has little desire to anticipate the finer integration details. Thus it is left to intellectual property (IP), chip, and module providers to find and fix problems. Test equipment such as spectrum analyzers and oscilloscopes have difficulty providing useful help unless the jammers are so large that they obliterate the GPS signal altogether.
Indoor applications of GPS receivers are relatively new. Using network-assist techniques, the viability of GPS has been extended down to -180dBW and below, approximately 20dB lower than earlier GPS receivers. At the same time the antenna, traditionally mounted remotely (on masts in marine applications, on bodywork in automotive and airborne applications) has migrated to the equipment case--or even inside--for handhelds. These factors significantly change the environment and issues relative to GPS design.
Here we explore some of the new problems these receivers experience: self-jamming mechanisms inside the receiver electronics, the lower levels of interference that become significant when working indoors, and some commonplace sources that exist inside buildings. We discuss possible approaches for mitigating the effects through signal processing, such as interference detection, filters, and pre- or post-correlation compensation. We conclude with some speculation on areas for future developments.
Inside The Room
Some interference problems, particularly the cross-correlation issue, have been with GPS engineers from the earliest days. However, indoor receivers now encounter significantly larger problems.
One or more signals may enter the room substantially unimpeded, whilst others are attenuated either by passage through or reflection off walls. Typically, an indoor receiver must work down to -180dBw or lower in most large buildings with steel frames, ducting, and wiring. Immediately this makes the cross-correlation problem worse. Additionally, the antenna location inside the building puts it much closer to potential sources of interference. The modern office, shopping mall, airport, and vehicle carry a great deal of IT equipment. Many homes also have digital equipment such as personal computers (PCs), MP3 players, and other digital devices; analog TV may have a digital set-top box, and digital audio broadcasting has joined analog radio.
Radiated Interference. EMC regulations ensure that reputable equipment conforms to some standards, to avoid destructive interference with other electronic devices, or interference with communications, radio, and TV. Typical EMC regulations limit radiated interference to the order of 40 microvolts per meter at a range of 10 meters. In the case of indoor GPS, the receiver and its antenna may be closer to the equipment in question than the EMC test distance. For a substantially omnidirectional GPS antenna, the effective aperture will be 0.01square meters or less, and the potential interference power at the antenna is of the order of -130dBW. Figure 1, taken in an electronics laboratory with many PCs and other test equipment, shows interference spikes at approximately -120dBW at the antenna.
Figure 2 shows the harmonic of a PC clock utilizing clock jitter as a means of EMC reduction--the 10MHz flat topped area near the center of the screen, and crossing 1575 MHz. The emission is not reduced in power, but is spread, in this case by about 6 percent of its center frequency. EMC measurements are usually made with ~10kHz bandwidth, so that spreading the energy much wider allows more radiated power to pass the regulations. This is not good news for the indoor GPS community!
Whilst PCs, and especially their clock-related harmonics, are the most obvious sources, in practice most electronic equipment (printers, faxes, MP3 players, and so on) now contains microprocessors and so generates similar EMC. Those devices that use screens may also generate interference related specifically to the screen-drive circuits and scan rates, or to backlighting charge-pump circuits.
By many standards, these interference sources are small powers, but for GPS, with a nominal received power at the antenna of -160dBW, they represent large figures! Of course, if the interference is out-of-band, then filters will reject it, whilst for in-band interference the correlation loss (typically 30dB for C/A code) may be sufficient to keep the outdoor type of receiver working. For an indoor receiver looking for -180dBW, correlation loss alone does not suffice, and some form of jammer rejection is required.
Inside The Host
The receiver must accommodate interference from the outside world, even if it is inside a building. Usually, the GPS will have the benefit of at least a small distance between it and most of the devices described in the previous section. But there are other sources to consider, even nearer to home: those inside the host equipment itself. For a GPS integrated into a modern, very compact handset, these will be very close indeed. The host equipment designer may not be aware of the requirements of an indoor GPS, nor have freedom to choose a frequency plan to suit. This can produce problems when GPS and host are finally integrated.
Entry Points. The most likely entry points are at the L-band antenna or low-noise amplifier (LNA) circuits, these being the most sensitive points (that is, with the largest gain before the digital signal processor, or DSP). The first downconversion will be sensitive to both the GPS band and the image frequency. If the mixer used for this downconversion is an image-reject type, there will likely be 20-30dB less gain for the image frequencies, and a jammer getting in at the RF input would need to be this much bigger to cause problems. From the antenna, there should also be an RF filter that will attenuate out-of-band signals, perhaps by another 40 or 50dB. For an indoor receiver where jamming is a serious problem, the frequency plan should ensure that the image lies outside the 20-50 MHz bandwidth of the RF filter, to benefit from this additional rejection.
Signals can also enter at the intermediate frequency (IF) stage, but must be substantially higher in power (by perhaps 30dB). However, the IF filter elements themselves (especially inductors) may make antennas sensitive to low clock harmonics. In most cases, such a jammer must be close to the GPS IF band (probably between [+ or -]2 MHz) unless it has sufficient power to overcome the filter rejection, which may be up to 60dB, depending on which IF component is sensitive to the interference.
Other entry points within a host environment include power supplies, that may directly introduce noise into amplifier stages, or modulate a local oscillator (or even another out-of-band jammer) to provide an in-band spurious signal. Usually these coupling mechanisms predominantly create noise and lower sensitivity.
