.
Agenda Item:1 (WRC Agenda item 1.6)
Wideband Wireless Channel Characterization
of the
MLS Band (5000-5150 MHz)
Status Report
Larry Foore
NASAGlennResearchCenter
United States of America
David W. Matolak
OhioUniversity
United States of America
Summary
Under the Space Based Technologies (SBT) project (informally known as ACAST), the NASAGlennResearchCenter is sponsoring a channel sounding effort proposed by Dr. David Matolak of OhioUniversity. The effort plans to characterize the 5000-5150 MHz band (with emphasis on the MLS extension band 5091-5150 MHz) in the form of a set of statistical channel models, parameterized appropriately for the area(s) of interest (surface and potentially terminal areas). A proper statistical characterization of the channel is not only necessary for optimal waveform design, but for the evaluation of existing potential waveforms in these areas.
This report is being presented to the International Civil Aviation Organization (ICAO) WG-F to inform the aviation community of the progress made toward properly evaluating this channel for wireless wideband signaling. Included are photographs of the test equipment and current facilities, a brief description of the test plan, and example preliminary measurement results.
Background
The ACAST project at the NASAGlennResearchCenter has committed to supporting the FAA in providing research and technical support for the efficient use and protection of aviation spectrum. Through industry support functions such as the I-CNS 2004 Conference and the ACAST Workshop 2004, NASA has identified the protection of the 5000-5150 MHz band (with emphasis on the MLS extension band 5091-5150 MHz) for aviation use as one of the top priorities for the ACAST project.
In consideration of WRC-07 Agenda item 1.6, NASA (in consultation with the FAA) has decided to evaluate the use of the MLS extension band for short range, wideband signaling in the surface area environment. This channel will be thoroughly characterized and understood in order to provide a proper evaluation of any particular system or waveform. There have been some academic efforts concerning aeronautical channel modeling [1], which rely heavily on land-mobile channel models for surface area environments. However, much of the terrestrial research is concerned primarily with the statistical characterization of path-loss in relation to land and urbanization variations. The multipath associated with the airport surface environment has the potential to differ significantly from channel models developed exclusively for land mobile (cellular) communications, due to the wideband nature of the signals being considered.
In an effort to properly characterize the use of the MLS extension band for wideband signaling, the NASA Glenn Research center has sponsored a grant proposal created by Dr. David Matolak of OhioUniversity. Dr. Matolak has organized a research effort that will create a parametric, stochastic model of the surface area channel for the MLS extension band [2,3]. The model development process relies heavily on empirical data, which is obtained from channel sounding. Various transmit/receive positions, surface variations, and weather conditions will be used to develop a collection of channel impulse responses (CIRs). This data will be categorized, parameterized, and properly analyzed to develop the channel models.
Wideband Channel Sounding
Wideband channel characterization varies significantly from narrowband characterization, especially in potentially rich multipath environments such as the airport surface. Time-delayed reflections of the transmitted signal arrive at the receiver with varying amplitudes and potentially different Doppler shifts. For wideband signals, as signal bandwidth becomes larger there is a greater potential for the behavior of individual spectral components to become uncorrelated due to phase shifts along reflected paths. This results in frequency-selective fading. The bandwidth over which all frequency components are affected similarly is termed coherence or correlation bandwidth [4].
In the time domain, the time-delayed echoes of the transmitted signal overlap at the receiver causing intersymbol interference (ISI). Without corrective measures via channel equalization techniques, the delay spread sets the lower bound on error rate for a specified data rate [4]. This delay spread, coupled with Doppler shifts, must be properly characterized for a given channel for waveform design/analysis.
Much of the theoretical foundation for classifying randomly time-variant (fading) linear channels was developed in a classical paper by Bello [5]. Bello’s work describes dispersive channels in a statistical sense, bringing together the consideration of both random time delays and random Doppler shifts into system transfer functions. Bello also showed the relationship between various classifications of dispersive channels, namely the Wide Sense Stationary (WSS), Uncorrelated Scatterer (US), WSSUS, and Quasi-WSSUS channels.
Parsons’ text does an excellent job of explaining how this theoretical development relates to practical channel sounding and classification [4]. By assuming that a dispersive channel is stationary over short intervals of time, Parsons shows how the assumption of WSSUS can be applied to obtain the autocorrelation function of the channel output from the time distribution of received power, or the power-delay profile.
A relatively recent paper published by Bultitude [6] discusses some pitfalls and recommended practices when describing a radio channel with regards to empirical data collected via channel sounding techniques. More specifically, Bultitude demonstrates how to ensure the assumptions of stationarity and ergodicity from empirical data, as well as how to correctly estimate a channel’s frequency correlation function from this data. Ultimately, as Bultitude points out, the correlation bandwidth sets the lower bound for error performance.
From the theory set forth by Bello, Parsons has clearly explained how the theory has practical application in wideband channel characterization. Bultitude provides guidance beyond this by demonstrating actual data processing for the examples of Rayleigh and Rician fading channels.
Dr. Matolak has constructed a plan to collect the empirical data necessary to follow the suggested path to wideband channel characterization in the MLS extension band. This status report discusses the equipment being used and how it relates to the theory discussed. Additionally, sample power-delay profiles are discussed from preliminary indoor trials.
Modified Berkeley Varitronics Raptor Model Channel Sounding System
The sounding system utilized for data collection in this effort is a modified Raptor model developed by Berkley Varitronics Systems. The standard model of the Raptor was developed primarily for indoor multipath and propagation analysis in the range of 2.4-2.85 GHz. The modified model being used for the sounding effort has translated the output signal and corresponding tuned receiver to the MLS band.
In the simplest sense, the transmitter portion of this system outputs an oversampled PN sequence that is BPSK modulated. This signal is filtered via a 128 stage FIR to produce the desired baseband signal mask. This wideband signal is then converted up to the desired frequency range. The maximum output power of the Raptor transmitter is approximately 33 dBm, which is relative to the types of short range, wideband signaling that may be used in surface area networks.
The received signal, after proper down conversion and IF filtering, is I/Q demodulated and fed to a bank of digital correlators. These correlators allow the receiver to search for the multipath components and identify Doppler shifts to produce power-delay profiles.
The desired output for this effort is in the form of power-delay profiles that are saved offline and post-processed via Matlab routines. These routines will produce the important channel parameters (with consideration of Bultitude’s work [6]), such as the correlation bandwidth, coherence time, the RMS delay spread, etc.
Figure 1 displays the Raptor receive and transmit units side-by-side.
Figure 1 - Raptor Channel Sounding System (Left – Receiver, Right – Transmitter)
The following figure displays the Raptor transmit configuration as seen in the lab. Note that the transmitter is connected to an omni antenna. These antennas have an approximate 2 dBi gain.
Figure 2 - Raptor Transmitter and Omni Antenna
The following figure displays the Raptor receive configuration as seen in the lab.
Figure 3 - Raptor Receiver, Omni Antenna, and Data Logging Laptop
Sample Output
As previously mentioned, the desired output from the receiver comes in the form of power-delay profiles. The power-delay data will be post processed to determine important channel parameters such as the correlation bandwidth, coherence time, and RMS delay spread.
The following two plots are examples of power-delay profiles obtained with the Raptor system. These data sets were obtained from indoor testing, so the measured values of delay spread are expectedly small.
Figure 4 - Power-Delay Profile, Indoor, Minimal Multipath
Figure 5 - Power-Delay Profile, Indoor, Observable Multipath
The first plot (Fig. 4) displays a strong Line-of-Sight (LOS) component, with some correlated power received afterwards. There was no significant multipath component from this collection.
The second plot (Fig. 5) shows not only a primary component, but a secondary echo that is 5-6 dB down from the primary. This demonstrates the Raptor’s capability to discern multipath components and is preliminary to testing on OhioUniversity’s airport surface.
Going Forward
Over the course of the remainder of 2005, Dr. Matolak in conjunction with NASA and the FAA, plans to sound the surface areas of at least 3 major U.S. airports. A sounding campaign is planned to begin at the ClevelandHopkinsInternationalAirport at the end of February, 2005. OhioUniversity’s own local airport, as well as (potentially) the NASAGlennResearchCenter campus, will be used as a testing ground for procedures.
Ultimately, the development of the database and parametric models will aid in the evaluation of existing system performance in the MLS band or the design of an optimal waveform for the band. For example, existing waveforms may require additional channel equalization for optimal performance pending the degree of delay spread experienced in this channel. Correspondingly, an optimal bandwidth will be identified for the design of a waveform specific to this channel.
Recommendations
This channel characterization effort will contribute to efforts set forth to address Agenda Item 1.6 of the WRC-07. The data collected during these sounding campaigns will allow not only for the development of wideband channel models, but also narrowband channel models of the type that may be used in surface sensor collection networks (fixed, directional point-to-point links).
NASA respectively recommends that this work be considered as part of the investigation to develop requirements for aeronautical systems in the MLS extension band (note Resolution 414 of WRC-03). Particularly, this work may influence the re-characterization of this band for the use of AM(R)S systems to support next-generation aeronautical communication systems on the airport surface.
References
[1] M. Haas, “Aeronautical Channel Modeling,” IEEE Trans. on Vehicular Tech., vol. 51, no. 2, pp. 254-264, March 2002.
[2] D. W. Matolak, “Wireless Channel Characterization: Overview and Application to 5 GHzBandAirport Surface/Terminal Environments,” OhioUniversity Report for NASAGlennResearchCenter, May 2004.
[3] D. W. Matolak, “5 GHz Wireless Channel Characterization for Airport Surface Areas: Project Workplan and Approximate Schedule,” OhioUniversity Report for NASAGlennResearchCenter, August 2004.
[4] J. D. Parsons, The Mobile Radio Propagation Channel, 2nd ed., John Wiley & Sons, New York, NY, 2000.
[5] P. A. Bello, “Characterization of Randomly Time-Variant Linear Channels”, IEEE Trans. on Comm. Sys., CS11, pp. 360-393, December 1963.
[6] R. J. C. Bultitude, “Estimating Frequency Correlation Functions From Propagation Measurements on Fading Radio Channels: A Critical Review”, IEEE Journal on Selected Areas in Comm., vol. 20, no. 6, pp.1133-1143, August 2002