Relative Positioning Using Pseudolites in the Navigation Systems Testing Laboratory at NASA's Johnson Space Center

Background
This research investigates the use of GPS pseudolites as tools for relative positioning in the Navigation Systems & Technology Laboratory (NSTL) at NASA's Johnson Space Center in Houston, Texas. Pseudolites (false-GPS satellites) can be used to produce pseudorange and carrier phase measurements, which can be received and transformed into a position estimate in the software of a GPS receiver. An application such as this could be used to mitigate the effects of signal loss in the receiver of a spacecraft that is in the proximity of a larger vehicle, such as the International Space Station. The smaller spacecraft could then rely solely on pseudolite measurements from the larger vehicle for navigation. Other applications might include indoor positioning and machinery automation. In order to apply this technology in space, it must be fully understood. This investigation explores what has been learned about pseudolites and their receivers in the NSTL. The environment of the laboratory and its effect on the measurements is studied. Further, deterministic solutions that use pseudorange, carrier phase and carrier smoothed code measurements and Kalman filters that use carrier phase and carrier smoothed code are investigated. It is shown that in the NSTL the errors on the pseudorange measurements limit any practical system to a carrier phase only solution. However, if multipath is overcome and the noise on the pseudorange is within normal levels, a carrier smoothed code algorithm could be incorporated to produce acceptable measurements for navigation. Additionally, if the system is in motion, an extended Kalman filter can be used to estimate position and velocity based solely on the carrier phase measurements.

Space Shuttle Orbiter Approaching the ISS

Facility and Hardware
The NSTL is a high-bay converted into a testing facility at JSC. The pseudolites and their antennas are arranged in a constellation within the NSTL. In the center of the NSTL is a workbench where the receiver and its antenna sit. Experiments have shown pseudorange measurements in the NSTL to be too noisy for current use. Research is underway to mitigate the multipath which is believed to be the main source of this noise.

The NSTL "Ship Channel" Receiver

An IntegriNautics IN200C-XL signal generator. This is the pseudolite that sends a signal to the transmitting antennae.

Noise in the NSTL
For this collection, two antennae were separated approximately 0.5 meters in the x-direction and routed into a common receiver. The top four plots illustrate the single-difference of the pseudorange measurements. For the separation distance at which these data were taken, the means should be on the order of a meter. The standard deviations are within commonly accepted values of pseudorange noise. The four plots on the right illustrate the single-differences of the carrier phase measurements with mean biases removed.

Views of the NSTL

Pseudolites E, B, and F on the north wall	Pseudolites F, B, A, and C on the east wall and the ceiling	Workbench and pseudolites C and A Close up of pseudolite A on ceiling of NSTL
Pseudolites E and F on the north wall	Close up of pseudolite F with a helical transmitting antenna
The metallic high-bay door and air handling system (facing west)	Close up of pseudolite F with a Reduced Surface Wave antenna

Analysis and Results

A set of files was written to analyze the measurements made in the NSTL. Since the pseudorange measurements from the NSTL were unusable, a simulation was written to test the receiver navigation software. The flowchart to the right illustrates the logic in the code. Additionally, seven different estimation algorithms were written. Some were pedagogical in nature, such as deterministic solutions using pseudorange measurements, carrier phase measurements with initialization, or a combination of the two measurements. A batch algorithm that estimated the state at each epoch based on all data from prior to and including that epoch was written. The batch estimate used the same initial guess for each epoch. Another deterministic solution used carrier smoothed code (CSC) data. The CSC method uses an extended sequential filter to center the carrier phase measurements with the pseudorange measurements, thereby removing the need to solve for the ambiguity. Over time, the CSC solution relies on the rate of change of the carrier phase measurement. The figure below shows the CSC measurement estimate as a function of epoch. The next figure to the right is a deterministic position solution using CSC measurements. A RAIM was employed so only measurements that fit the predictions were used. Two extended Kalman filters (EKFs) were derived. The first uses CSC measurements for a position solution. The second uses only carrier phase measurements to estimate the position, velocity, and integer ambiguities. This method requires that the rover antenna is in motion.

Carrier Smoothed Code

Carrier Smoothed Code Position Solution

Extended Kalman Filter using Carrier Smoothed Code

Extended Kalman Filter using Carrier Phase

Conclusions

Future study should attempt to combine the algorithms studied. The simulation shows that if motion is measurable, only carrier phase is needed for an accurate position solution. If motion is not measurable, carrier smoothed code techniques that combine pseudorange and carrier phase measurements should be used to obtain the most accurate position possible.

Written by Geoffrey Wawrzyniak, May 2001