Sensitive Receivers Enable Effective GPS Design

When developing receiver designs, engineers can turn to chip set solutions that speed and ease the design process.

Contents
Four Equations; Four Unknowns
Error Processing
Signal attenuation
Building GPS systems
The heart of the chip set
DSP interfaces
Differentiating features
Power consumption

By: Greg Turetzky, Product Line Manager, SiRF Technology

The US Department of Defense (DoD; Washington, DC) has built an elaborate Global Positioning System (GPS) for pinpointing locations virtually anywhere on the planet. A constellation of 24 NAVSTAR satellites circles the Earth every 12 hours emitting radio signals that contain information about their positions. Specialized receivers on or near the planet surface receive these signals and calculate their positions relative to these satellites.

Four equations; four unknowns
Essentially, a GPS system works by having a receiver pick up signals from four satellites and measuring the time it took for those signals to arrive. From this timing information, one can calculate the distance between the receiver and each satellite. The four satellites' ephemeris data provide the satellite's X, Y, and Z positions. The range, R, is the receiver measurement made by calculating the time it took for the signal to reach the receiver. The user's position, (Ux, Uy, Uz), and the clock bias, Cb, is then calculated (Figure 1).

Each of the 24 satellites transmits a set of signals using spread spectrum technology. Spread spectrum technology enables low-powered satellites to produce signals that can be detected at very low received-signal levels.

Essentially, the carrier signal is modulated by a unique coding sequence, which has the effect of spreading the signal's frequency spectrum. Using a replicated code sequence, a GPS receiver searches that spectrum looking for a match. The signal can then be "unspread" and decoded. By transmitting several signals over the same spectrum, but using distinctly different coding sequences, each signal can share the spectrum without interfering with any of the others.

Error processing
The GPS system assumes that signals will be traveling between satellite and receiver in a straight line. The signal will actually be delayed upon going through the ionosphere, and the receiver timing references will have some small error. Both of these errors are predictable and correctable.

Multipath error, on the other hand, can create very large deviations. Multipath is caused by satellite signals that arrive at the receiver after having bounced off some nearby structure (e.g., a tall building), or the ground. Since the path is not straight, the time delay will be longer, and the distance from the satellite will also seem to be longer (Figure 2). This can produce location errors that are unacceptable, particularly in urban automobile navigation applications.

Signal attenuation
Non-restricted GPS signals are transmitted at 1.575 GHz. Such signals are blocked by steel and concrete structures (e.g., buildings and tunnels), and attenuated by passing through trees and leaves.

The GPS specification for minimum detectable signals renders reception marginal when the signal is attenuated by foliage. The denser the foliage, the more marginal the signal. As such, receivers that just meet this specification are not reliable for use in forests or even tree-lined streets.

To ensure that a signal can be received in a forest, the receiver must provide sensitivity that exceeds the current standard. For example, the receiver must be able to detect signals whose power has been attenuated to a level of about 5% of the initial level.

Building GPS systems
To build a GPS-based navigation product, an engineer must design a radio that can receive the spread-spectrum signals. The detected signals are then converted from RF signals into appropriate digital input formats. These digital inputs are processed and converted into position information, and the information is then processed to produce the required application output (e.g. a blinking cursor on a map overlay, a readout of latitude and longitude, etc.)

To develop a GPS receiver, engineers may want to turn to chip set technology, such as Santa Clara, CA-based SiRF Technology's SiRFstarI/LX product.

The heart of the chip set
The RF front end provides the heart of the GPS chip set solutions. For example, GRF1/LX chip serves as the front end in the SiRFstar/LX solutions. This chip converts GPS signals from their 1.575 GHz frequency into baseband signals. To accomplish this, this IC integrates an LNA, mixers, amplifiers, a synthesizer, and an analog-to-digital (A/D) converter.

RF front ends, such as the GRF1/LX, can also incorporate an on-chip voltage-controlled oscillator (VCO). By including the VCO on-chip, GPS receiver designers can replace the costly external oscillator required by traditional GPS designs with a cheaper external crystal (Figure 3).

In general, the RF front end IC interface with both the standard active antenna and a digital signal processing (DSP) companion chip. In the SiRFstar solution, the DSP chip, dubbed the GSP1/LX, provides a parallel processing architecture and a SnapLock feature.

The SnapLock feature allows the receiver to reacquire satellite signals in 100 ms. By doing this, the receiver can more quickly provide location information if a satellite signal is temporarily lost.

DSP interfaces
The baseband processor in a GPS chip set solution provides interfaces to other aspects of the system. For example, the GSP1/LX is designed to interface with any standard 8-, 16-, or 32-b microprocessor as well as 8- or 16-b static random-access memory (SRAM) chips. It also features a 2-b interface to the GRF1/LX and can produce 10 positions-per-second output (Figure 4).

Differentiating features
Overall, many of today's GPS chip sets offer the same number of chips as competing solutions. The key for engineer, therefore, is to seek out the features that differentiate one GPS solution from another. To illustrate, let's look at the SiRFstar architecture.

SnapLock Acquisition
As stated above, the SiRFstar chip set provides a SnapLock feature acquisition feature provides re-acquisition of satellite signals in only 100 ms. SnapLock acquisition results from a parallel spectrum search to find code correlation, involving 20 code samples.

SnapLock acquisition is an important feature in navigation applications. Cars lose satellite visibility in cities because they are blocked by tall buildings and tunnels, but they get a clear view in intersections, or when exiting a tunnel. The average time in an intersection is 1 to 2 s, but a re-acquisition time of 2 or 3s leaves no time for collecting signal data.

SnapLock acquisition re-acquires the signal and collects a measurement for a position update in one-tenth of a second. Thus, an intersection offers enough time for both re-acquisition and positioning when a system employs SnapLock technology.

This high-speed re-acquisition is also a key part of the power management scheme. Since the signal can be re-acquired in 100 ms the chipset can be power cycled at a rate faster than the standard 1 Hz update rate, causing no apparent loss of data but at greatly reduced power consumption.

SingleSat Positioning
When driving in an urban area, intervening buildings often block a car's satellite visibility. For other GPS systems, when less than three satellites are visible, positioning calculations cannot be made.

However, SiRF's SingleSat positioning mode allows positioning calculations, for short periods, when only a single satellite is visible. SingleSat positioning works by using a single satellite's data to determine how far along a current path the car has traveled. Any errors in position can be corrected as soon as SnapLock reacquires three or more satellite signals (e.g., when the car passes through an intersection). Car navigation systems employing SiRFstar technology will thus provide more position fixes than other systems when navigating in an urban setting.

Dual Rejection
Multipath errors occur when signals reach a receiver along an indirect path. Low level reflected signals bouncing off far-away objects are simply eliminated. Errors caused by nearby reflected signals are filtered.

In order to stop multipath problems, the SiRFstar solution employs a dual rejection scheme. Without such a rejection scheme, multipath-induced errors often cause random, large-scale errors in positioning for car navigation systems being used in urban areas.

Power consumption
GPS technology is finding its way into smaller and smaller portable devices. In these space-constrained environments, power consumption is a key concern. So engineers must turn to low-power chip set solutions for their receiver designs.

Chip set manufacturers are employing many techniques to reduce consumption in their designs. For example, new foundry technology and peripheral integration is being used to reduce consumption in both the RF front end IC and DSP. In addition, high-precision real time clocks are being employed to keep very accurate time (to a few microseconds) during power down to enable very fast restarts.

To reduce power consumption even further, manufacturers are also turning toward locking features, such as the SiRFstar SnapLock approach. These features allow the receiver to turn off in inactive periods and quickly re-lock on a signal when needed again. This enables the receiver to provide a continuous 1 Hz update while consuming less power.

About the Author
Greg Teretsky, Product Line Manager, SiRF Technology, Inc., 3970 Freedom Circle, Santa Clara, CA 95054. Tel: 408-980-4700; Fax: 408-980-4705.