UWB Tracking System Design with TDOA Algorithm for Space Applications

University of Houston • University of Houston-Clear Lake • ISSO Annual Report Y2006 • 11-16

Edward Dickerson, Dickey Arndt, David (Jianjun) Ni

ABSTRACT—This study describes a design effort for a prototype UWB (ultra-wideband) tracking system currently under development at the NASA-Johnson Space Center. The system is being studied for its practical use in procedures for tracking Lunar/Mars rovers during early exploration missions when satellite navigation systems (such as GPS) are not available. The Science Crew Operations Utility Testbed (SCOUT) vehicle under development at JSC provides a testbed for the utilization of the UWB tracking system in the space environment. Field tests were conducted jointly with the SCOUT vehicle at the Meteor Crater in Arizona to test the tracking capability for a moving target. These maneuvers showed the compatibility of the UWB tracking system under simulated conditions.

Research scientists are exploiting UWB technology to im plement a UWB tracking system with its fine-time resolution on the order of picoseconds, its low-power spectral density, which allows the system to coexist with other communication systems, and its resistance to multipath interference. In addition, the high data rate capability of UWB provides a multimedia communication channel that can support a passive tracking system with very little increase in system complexity, cost, or power requirements. A two-cluster Angle-of-Arrival (AOA) tracking method, using Time Difference of Arrival (TDOA) information, is utilized for implementation of the tracking system, not only to exploit the achievable fine time resolution of UWB signals, but also to eliminate the need for synchronization between the transmitter and the receiver. The UWB radio at each cluster is used to obtain the TDOA estimates from the UWB signal sent from the target. Assuming this is a long-range application, the TDOA data can be carefully converted to AOA data to find the angle of arrival. Since the distance between two clusters is known, the target position is computed by a simple triangulation.

To evaluate tracking performance of the AOA algorithm, several Matlab simulations have been conducted. These simulation results reveal that the tracking resolution is a function of the TDOA estimate noise level, the baseline receiver configuration, and the tracking range. Simulations show that the AOA algorithm achieves fine resolution with a low estimation noise level of TDOA data. This provides valuable guidance for design of the tracking system.

Field tests were conducted jointly with the SCOUT vehicle at the Meteor Crater in Arizona to test the tracking capability for a moving target. These tests demonstrated that the UWB tracking system can co-exist with other RF communication systems onboard SCOUT and that a tracking resolution less than one percent of the range (range up to 2000 feet) can be achieved.

Introduction
Researchers at the NASA-Johnson Space Center (JSC) are developing a manned Lunar/Mars rover prototype known as the Science Crew Operations and Utility Testbed (SCOUT). For space exploration, after a Lunar/Mars outpost has been established, it will be necessary to track the rover or astronauts' positions while they work off the base. In early exploration phases, there were no satellite navigation systems (such as GPS satellites) available around the moon or Mars. Hence, a RF tracking system is needed which can co-exist with other communication systems used by the landing vehicle and other spacecrafts.

A UWB R&D Group has been formed at NASA/JSC to develop and evaluate communication and tracking systems using UWB technology. In this report, we document a research effort that exploits UWB technology to meet the tracking system design challenges posed by the complex operational environments of NASA prototypes such as the SCOUT vehicle.

The remainder of this report is organized first with an overview of the UWB technology. Next, a tracking methodology Angle of Arrival (AOA) will be introduced and the simulation results discussed. The last section will discuss the field test results and provide some concluding remarks.

UWB Technology
Ultra-wideband (UWB), also known as impulse or carrier-free radio technology, is a promising new technology. It has been utilized for decades by the military and law enforcement agencies for fine-resolution ranging, covert communications, and ground-penetrating radar applications. In February 2002, the Federal Communications Commission (FCC) approved the deployment of this technology in the commercial sector under Part 15 of its regulations.¹ More researchers are recognizing that UWB technology holds great potential to provide significant benefits in many applications such as precise positioning, short-range multimedia services, and high-speed mobile wireless communications.

The DARPA (Defense Advanced Research Project Agency) study panel that coined the term ultra-wideband in the 1990s defines it as a system with a fractional bandwidth greater than 25 percent. Later, the FCC defined the term UWB to describe any signal with bandwidth equal to or in excess of 500 MHz or a fractional bandwidth greater than 0.2. The basic concept of current UWB technology is to develop, transmit and receive an extremely short duration burst of RF energy — typically a few tens of picoseconds to a few nanoseconds in duration. Whereas conventional continuous sine wave radio systems operate within a relatively narrow bandwidth, UWB operates across a wide range of frequencies (a few GHz) by transmitting a series of low-power impulsive signals.

For the emerging technology of UWB radar and UWB wireless communications, the transmitted signal can be regarded as a uniform train of pulses represented as

(1)

where Tr is the pulse repetition interval, and P(t) is the pulse-shaping waveform, which is often a Gaussian monocycle. In the time domain, the Gaussian monocycle is mathematically similar to the first derivative of a Gaussian function. It takes the form

(2)

where t is the duration of the monocycle. Figure 1 shows an ideal monocycle centered at 2 GHz in both the time and frequency domains.²

Conveying information over impulse-like radio waveform, UWB is characterized by several uniquely attractive features:

Low-power carrier-free ultra-wide bandwidth signal transmissions
Low impact on other RF systems due to its extremely low-power spectral densities
Immunity to interference from narrow band RF systems due to its ultra-wide bandwidth
Multipath immunity to fading due to ample multipath diversity (RAKE receiver)
Capability of precise positioning due to fine-time resolution
Capability for a high data rate, multi-channel performance
Low-complexity low-power baseband transceivers without intermediate frequency stage

At present, there are many UWB product manufactures. In 2003, there were only two companies, Time Domain Corporation³ and XtremeSpectrum, Inc.⁴ (acquired by Freescale Corporation in November 2003), offering commercially available products. In this research effort, we investigated their chipsets and exploited them to design the tracking systems.

Time Domain was the first to miniaturize UWB technology to a silicon solution, marketed as PulsON. PulsON systems use pulse position modulation. PulsON 100 is the first generation of UWB silicon chipsets that powered a variety of UWB product prototypes and demonstrations. In 2001, Time Domain launched the first commercial UWB chipset solution, PulsON 200, for many potential applications of UWB. The PulsON 200 chipset is comprised of three components:⁵

The PulsON Dual Timer, which precisely slices time into picoseconds
The PulsON Quad Correlator, the receiver cast in SiGe
The PulsON CMOS Logic Chip, a software-definable RF processor.

PulsON chipsets are available to customers as part of the PulsON Evaluation Kit (EVK), a comprehensive product development platform that supports the full range of UWB capabilities, including wireless communications, tracking, and radar.

XtremeSpectrum named its product TRINITY because it is a high-speed, low-power and low-cost chipset. The XtremeSpectrum XSI100 TRINITY chipset provides full wireless connectivity, implementing bi-phase modulated ultra-wideband and the IEEE 802.15.3 MAC protocol. In June 2002, the company demonstrated that TRINITY delivers up to a 100 Mbps data rate supporting applications such as streaming video, streaming audio, and high-rate data transfer at very low levels of power consumption. The four-chip set consists of an RF Front End, an RF transceiver, a digital baseband and a MAC.

The potential advantages of a UWB system for this application, including high-speed video transmission, multipath resistance, ease of interoperability with other RF systems, and precision tracking characteristics, together with the availability of commercial UWB products, makes this technology a good choice for implementation of the communication and tracking system. Rapid technological advances have enabled cost-effective implementation of UWB radar, communication, and tracking systems. Furthermore, array beam forming and space–time processing techniques promise further advancement in the operational capabilities of UWB technology to achieve long-range coverage, high capacity, and interference-free quality of reception.⁶ Hence, UWB technology is proposed with confidence to implement the communications and tracking system for space applications investigated in this research effort. The UWB radios (PulsON 200 EVK) from Time Domain Corporation have been chosen as the core hardware for the design of the tracking system since it was the only available evaluation and development platform in the UWB field when this project began. The latest available UWB products (such as PulsON 210 EVK from Time Domain Corporation) and next generation UWB development platforms will be utilized to enhance the system in the future work.

Figure 1. Gaussian Monocycle in Time and Frequency Domains

Tracking Methodology
Many different approaches can be applied to estimate the location of a radio source, including angle of arrival (AOA), time of arrival (TOA), time difference of arrival (TDOA), relative signal strength (RSS), and various hybrids of these parameters. For close-in applications, the TDOA approach has been chosen as the tracking method since it does not require synchronization between the transmitter and receiver but can still exploit the fine time resolution available with UWB signals.^7,8 For long-range applications, the AOA approach can be applied to estimate the location of a target since the approximation error under the far field assumption is relatively small. The AOA technique is discussed briefly below.

Figure 2. AOA Localization in 2-D Space

A two-dimensional AOA tracking case is illustrated in Fig. 2. Two receivers are used to locate the transmitter in this 2-D space by a simple triangulation. Two receivers' positions (Rx1(0,0), Rx2(d,0)) are assumed known. If the angle of arrival from the target to each receiver (i.e., q1 and q2) can be estimated, the transmitter's position can be computed using the Law of Sine as follows,

In order to find the AOA information (q1 and q2), two antennas spaced by distance a are connected with each receiver. Since the UWB signal has fine time resolution, the TDOA information (t12 and t43) has been measured and carefully converted to the AOA information as follows.

Since electromagnetic waves travel with constant velocity c in free space, the distance between the transmitter and the receiver's antenna is directly proportional to the propagation time of the signal. Under the long-range assumption (r1, r2 >> a),

ct12 ≈ a cosq1
q1 ≈ arcos (ct12 /a);

similarly,

ct43 ≈ a cosq2
q2 ≈ arcos (ct43 /a);

The advanced signal processing techniques are developed to estimate the TDOA information (t12 and t43) from the pulses transmitted from the target UWB radio.⁹ The TDOA estimates are then fed into the above AOA algorithm to locate the target.


Figure 3. Tracking Error with Perfect TDOA Estimates	Figure 4. Tracking Error with Noisy TDOA Estimates

Tracking Simulations
In order to analyze the tracking error behavior and gain some insight regarding achievable tracking resolution, several Matlab simulations were performed using the AOA tracking method described under “UWB Technology.” The results of two of these simulations are discussed below.

Tracking error with perfect TDOA estimates
In order to transform TDOA estimates to AOA data, a far field assumption (r1, r2 >> d so that the lines from the target to two antennas at each receiver are approximately parallel) is made. A 2-D tracking simulation is presented to illustrate the impact of this far field assumption on tracking performance. The simulation setup is as follows:

Baseline configuration: two receivers are 50 meters apart ( d =50m) and two antennas at each receiver are 15 meters apart ( a =15m).
Transmitter trajectory: the transmitter moves in a semi-circular orbit (from 0 degrees to 180 degrees) with a radius of 610 meters (=610m).
TDOA noise level: the TDOA estimates are assumed perfect.

The simulated tracking error due to the parallel approximation is plotted in Fig. 3. The simulation result shows that the tracking error due to approximation has a W-shape pattern. In general, this approximation error is relatively small with the average about 0.05 meters. The tracking error at certain trajectory range (angle of target from 30 degrees to 150 degrees) is below the average error.

Tracking error with noisy TDOA estimates
A similar simulation with noisy TDOA estimates (standard derivation s = 10 picoseconds) is conducted and the error analysis (from 30 degrees to 150 degrees) is illustrated in Fig. 4. The simulation shows that the tracking error is random and the average tracking error at range of 610 meters is 2.7595 meters, less than 0.5% of the tracking range.

Tracking error vs. affecting parameters
It can be shown that the tracking error MSE (Mean Squared Error) is a function of parameters a , d , r1 and s . Several simulations have been conducted to study how these parameters affect the tracking resolution. The default values of these parameters are as follows: a =15m, d =50m, r 1=610m, s = 10 picoseconds.


Figure 5. Tracking Error vs. Cluster Size	Figure 6. Tracking Error vs. Baseline Size

First, the relationship between the tracking error (MSE) and the cluster size (a) has been studied for both perfect TDOA estimates and noisy TDOA estimates. In Fig. 5, simulation results show that the tracking error increases as the cluster size increases for perfect TDOA estimates while the tracking error decreases as the cluster size increases for noisy TDOA estimates.

Second, the relationship between the tracking error (MSE) and the baseline size (d) has been studied for both perfect TDOA estimates and noisy TDOA estimates. In Fig. 6, simulation results show that the tracking error does not change as the baseline size changes for perfect TDOA estimates, while the tracking error decreases as the baseline size increases for noisy TDOA estimates.


Figure 7. Tracking Error vs. Tracking Range	Figure 8. Tracking Error vs. TDOA Noise Level

Third, the relationship between the tracking error (MSE) and the tracking range (r1) has been studied for both perfect TDOA estimates and noisy TDOA estimates. In Fig. 7, simulation results show that the tracking error decreases as the tracking range increases for perfect TDOA estimates while the tracking error increases as the tracking range increases for noisy TDOA estimates.

Finally, the relationship between the tracking error (MSE) and the TDOA noise level (s) has been studied for noisy TDOA estimates. In Fig. 8, simulation results show that the tracking error does not change when the TDOA noise level is low ( e.g. , less than 0.1 picoseconds) while the tracking error increases dramatically as the TDOA noise increases to picoseconds level.

The above simulation results in Fig. 8 show that the tracking resolution can be further improved by increasing the cluster size and the baseline size, if it is feasible, and by decreasing the TDOA noise level.

Figure 9. Test Site – Meteor Crater in Arizona

Outdoor Tests
Outdoor tests have been conducted to test the UWB tracking capability with extended range in an operational environment. A joint tracking test was conducted with the SCOUT vehicle at the Meteor Crater in Arizona. Figure 9 shows the test site. Two camps (SCOUT Camp and EC/Suit Camp) were set up near the Meteor Crater. Twelve way points (WP) with GPS coordinates were put between two camps and beyond to define the SCOUT running trajectory.

Figure 10. Configuration of the Two-Cluster UWB AOA Tracking System

The configuration of the two-cluster UWB AOA tracking system is shown in Fig. 10. Due to the conservative FCC limit on the UWB emission power (-41.3dBm/MHz), the transmitting range is limited. In order to increase the tracking range, four high-gain horn antennas are used as receiving antennas, and a low noise amplifier (LNA) is added at the receiving side after each receiving antenna. Two clusters are connected to a laptop computer through a hub.


Figure 11. Tracking System Baseline Setup	Figure 12. Tracking Target – SCOUT Vehicle

The two-cluster UWB AOA tracking system baseline has the following setup (Fig. 11): two receivers are placed 50 meters apart and two horn antennas at each receiver, 15 meters apart. One UWB radio was integrated with the SCOUT vehicle as the transmitter (Fig. 12).

The objectives of the test were to:

Test the real-time tracking of a moving target.
Test the interference with other communication systems on vehicle.

The SCOUT vehicle was running at normal speed (7 miles /hour) in the tracking area. Figure 13 shows the tracking accuracy compared to the differential GPS, with less than one percent error at ranges up to 610 meters (2000 feet). A SCOUT running trajectory recorded in Fig. 14 demonstrates the real-time tracking capability of the system. The tracking update rate for the trajectory was approximately 5 Hz. No RF interference was observed between the UWB tracking system and other on-board SCOUT systems (such as GPS at 1.6 GHz, video at 5.8 GHz, voice at 140 MHz and telemetry at 2.4 GHz).

Conclusion
A prototype UWB tracking system has been designed, implemented, tested, and proven feasible for space applications. UWB technology has been exploited to implement the tracking system because of its properties such as high data rate, fine time resolution, and low power spectral density. The AOA tracking method using TDOA information has been employed to avoid synchronization problems between the transmitter and the receiver. A two-cluster system with high-gain horn antennas has been implemented to increase the tracking range. Simulations demonstrate that the approximation error due to the far field assumption for the AOA algorithm is relatively small and the tracking scheme can achieve the desired fine tracking resolution. Outdoor tests have been conducted jointly with the SCOUT vehicle to test the tracking capability for a moving target. These tests demonstrate that the UWB tracking system can co-exist with other RF communication systems, and that a tracking resolution less than one percent of the range (range up to 2000 feet) can be achieved.

References
¹FCC First Notice and Order, “Revision of Part 15 of the Commission's Rules Regarding Ultra-wideband Transmission Systems,” ET-Docket , Feb. 2002.
²Time Domain Corporation, “PulsON Technology Overview,” <http://www.timedomain.com/Files/downloads/techpapers/PulsONOverview7_01.pdf>
³Time Domain Corporation, <http://www.timedomain.com/index.cfm>
⁴XtremeSpectrum, Inc., <http://www.xtremespectrum.com/index.html>
⁵K. Siwiak and M. Franklin, “Advances in Ultra-Wide Band Technology,” Radio Solutions, Commonwealth Conference & Events Centre, London, Nov. 6–7, 2001.
⁶M. G. M. Hussain, “Principles of Space-Time Array Processing for Ultrawide Band Impulse Radar and Radio Communication,” IEEE Trans. on Vehicular Tech. 51 (2002): 393-403.
⁷J. Ni, D. Arndt, P. Ngo, C. Phan, and J. Gross, “UWB Tracking System Design for Free-Flyers,” AIAA Space 2004 Conference and Exposition, Sept. 28–30, 2004.
⁸J. Ni and R. Barton, “Design and Performance Analysis of a UWB Tracking System for Space Applications,” IEEE/ACES Intl. Conf. on Wireless Communications and Applied Computational Electromagnetics, April 3–7, 2005.
⁹J. Ni, D. Arndt, P. Ngo, C. Phan, and J. Gross, “UWB Tracking System Design with TDOA Algorithm for Space Applications,” NASA/JSC Internal Report, August 2005.

PDF (528KB) | Contents