The 60 GHz band has recently attracted a great deal of attention due to the available wide bandwidth and the limited propagation range, which enables frequency reuse. This makes it particularly attractive for home and office environment for high data rates on the order of 1 Gbit/s such as those investigated in the WIGWAM project (Wireless Gigabit with Advanced Multimedia). The achievement of such data rates depends on the time dispersion properties of the radio channel, which give rise to intersymbol interference and frequency selective fading.

To characterise the radio channel in the 60 GHz band, the frequency and bandwidth of the wideband channel sounder previously developed for indoor measurements at 2 GHz have been upgraded to operate at 60 GHz with 1024 MHz bandwidth. The sounder uses a chirp waveform with programmable bandwidths up to 300 MHz at an update rate of 250 Hz. The technique provides optimum use of the available transmit power, high processing gain at the receiver, and bandwidth compression. The up-conversion to 60 GHz is performed using three stages. The first stage converts the 2 GHz to 5 GHz. This is followed by a 15 GHz stage where the bandwidth and the operating frequency are quadrupled to 60 GHz. To provoke multi-path behaviour for short paths moderate directivity antennae have been used. The antennae used here were conical horns with a nominal gain of 20 dBi. The sounder uses rubidium standards at both ends, which permits full dynamic operation in contrast to sounders, which use cables between the transmitter and receiver to provide timing synchronisation or local oscillator reference.

The sounder was used in line of sight and non line of sight indoor measurements. The spurious free dynamic range of the power delay profiles obtained over 1 s were on the order of 20 dB. Therefore, rms delay spreads for a 15 dB threshold level were estimated. These range from 0.77 ns for a single detected component to 13.7 ns for the worst case.

In this paper the architectures of the 60 GHz up-converter and down-converter are discussed with performance tests. Distribution of rms delay spread in measurements taken in the School of Engineering at Durham University is presented.

Zdenek NEMEC, Vladimir SCHEJBAL, Dusan CERMAK, Pavel BEZOUSEK University of Pardubice, Studentska 95, 53210 Pardubice,
Czech Republic P21 - UWB channel modelling
zdenek.nemec@upce.cz, vladimir.schejbal@upce.cz, dusan.cermak@upce.cz, pavel.bezousek@upce.cz
Prime Author: Zdenek NEMEC

The propagation of ultra wide band (UWB) signals is analyzed. The UWB radar output signals are formed both with transmitter and antenna. Moreover, they can be substantially affected due to electromagnetic wave propagation through walls. Experimental results and numerical simulations are analyzed and compared.

The UWB concept is very useful for radars and communications [1], [2], [3]. The effect of various antenna receiving and transmitting responses as well as UWB signals (pulses) are analyzed in [4] - [6]. Various combinations of signals, transmitting or receiving antennas (small and aperture antennas) and wall structures as well as multipath propagation have been calculated and compared. The papers [4] - [6] studies spectra and UWB signal propagation through walls. This paper is analyzing experimental results.

Of course, the real antennas cannot work from DC to infinity and therefore, they form band-pass filter. Therefore, the real antennas do not exactly perform differentiation or integration and their responses are causal. The considered wall parameters are given [4] both for brick and concrete walls for various wall thickness t.

Several experiments of UWB through-wall propagation for various walls both for frequency and time domains are currently carried out. Several examples of measurement will be given both for time and frequency domain for various walls (such as brick or concrete walls between various university laboratories and offices) and signals. That is compared with numerical simulations, which consider used antenna and signal types.

References
[1] HEYMAN, E., MANDELBAUM, B., SHILOH, J. Ultra-Wideband Short-Pulse Electromagnetics 4. New York: Plenum Press, 1999.
[2] TAYLOR, J. D. Ultra-Wideband Radar Technology. N. York: CRC, 2001.
[3] BEZOUSEK, P., SCHEJBAL, V. Radar technology in the Czech Republic. IEEE Aerospace & Electronic Systems Magazine. 2004, vol. 19, no. 8, p. 27 - 34.
[4] SCHEJBAL, V. et al. UWB Propagation through Walls. Radioengieering. 2006, vol. 15, no. 1.
[5] SCHEJBAL, V. et al. Electromagnetic wave propagation through obstacles. In International Workshop on Microwaves, Radar and Remote Sensing MRRS 2005. Kiev (Ukraine), Sep. 19 - 21, 2005, p. 255 - 260.
[6] BEZOUSEK, P. et al. UWB signal propagation through walls. In International Workshop on Microwaves, Radar and Remote Sensing MRRS 2005. Kiev (Ukraine), Sep. 19 - 21, 2005, p. 249 - 254.

The need for mobile applications that are spectrum-hungry due to the high bandwidth requirements (e.g. video transmissions, TV on the move) has fuelled research into wireless mobile communication systems capable of high data-rate throughputs at short ranges. The development of such systems requires an understanding of dynamic variations in the propagation channel characteristics. This work describes measurements for the characterisation of the dynamic wideband propagation channel at 60GHz in a cluttered environment, with anticipated mobile platform speeds of up to around 80mph (130km/hr).

The vast potential of wireless communications at 60GHz, where the radio spectrum is currently less congested, can be exploited to provide a solution to the ever-increasing demand for higher bandwidth transmissions on the move. At these frequencies, transmissions are typically of short ranges of 1 - 2 km due to the high oxygen attenuation. As a result, there is greater potential for frequency re-use. Users of such wireless systems might demand the reception and transmission of real-time and location-specific information instantaneously. Typical access speeds may be around 40 - 400Mbps. The successful development of such a communication system in a cost-effective manner would require detailed understanding of the propagation channel, and its dynamic variations.

A measurement campaign is conducted to provide dynamic wideband channel measurements at 60GHz for a range of platform speeds. The demanding practical issues involve the mechanical arrangements for the deployment of the channel sounding equipment, and the choice of waveform repetition rate, which must be sufficiently high to measure the large Doppler spreads while still providing a sufficient range of delay spreads.

The QinetiQ wideband channel sounder, Jasmine™ [http://www.cpar.qinetiq.com/jasmine.html], is deployed for the dynamic channel measurements. The sounder offers an extremely wide Doppler range of ±20kHz, whilst still providing ample bandwidth of up to 250MHz and delay spread measurement capability of up to 10µs, with a resolution of up to 10ns. The sounder also allows sequential MIMO (multiple-input multiple-output) operation, and features a 2GB/s fibre-channel disk array.

Measurements are conducted to study the effect of vehicular traffic on fixed point-to-point communication links at a height of approximately 3 - 5m. The effect of traffic movement on the propagation channel is investigated. In addition, measurements are collected for fixed-to-mobile communication links in various locations at a variety of different mobile platform speeds. The results provide a preliminary characterisation of the dynamic propagation channel at 60GHz, which would facilitate the development of channel models required for the development of wireless mobile communication systems at 60GHz.

The 5 GHz band is interesting for future broadband wireless access systems, which might be deployed even in outdoor scenarios. Due to the high path loss, a larger area can only be covered with multiple overlapping mircocells. A promising way out to reduce the resulting interference is a coherent cooperation between these microcells. However, little is known about the channel parameters in the 5 GHz band for distances beyond the typical coverage area [1]. Knowledge about the macrocellular propagation conditions may, hence, be beneficial for the system design even if microcellular deployment is desired.

In this paper we analyze data from 5.2 GHz channel measurements in an urban macrocell scenario in order to figure out the statistics of important channel parameters. The most common parameters for a realistic simulation of the inter-cell interference are the path gain, Ricean factor [2], delay spread and excess delays.

The measurements were done with the RUSK HyEff channel sounder [3] at 5.2GHz with 120MHz bandwidth and up to 10 W transmit power. Data have been collected in a typical metropolitan area in Berlin, Germany. While the transmitting (Tx) antenna is vertically polarized and omni-directional, the 8 receive antennas have relatively wide spacing of about 1 m. They are made of dual polarized patches, with horizontal or vertical polarization. The receiver (Rx) was located on rooftop of a 52 m high building while the Tx was mounted on a trailer driving up to 21 km/h.

Owing to a careful noise reduction in the delay domain, the measurement coverage reached distances up to 700 m between Tx and Rx. The map including all measurement routes is given in Fig. 1. Path gains down to -140 dB have been detected reliably which are plotted in Fig. 1. Additionally, we have considered the delay parameters (delay spread, excess delay). In Fig. 2, the cumulative distribution function (cdf) of the excess delay is shown when 95% of the signal power is received in selected distance regions. Within the microcell, the widely accepted COST207 value of 1 μs for the microcell environment is verified by our measurements in the first distance region below 200 m. If this is the radius of the microcell, we need values up to at least 600 m distance, for prediction of the interference in the adjacent cell having a diameter of 400 m. The maximum observed excess delay in this region is 2.3 μs, while the mean path loss increases from 107 dB (<200 m) to 130 dB (400...600m). For inter-cell interference mitigation, the stronger temporal dispersion of the interference signal must be taken into account, resulting in somewhat larger guard intervals.

[1] Uwe Trautwein, Markus Landmann, Gerd Sommerkorn, and Reiner Thoma. System-oriented measurement and analysis of mimo channels. In COST 273 TD(05) 063, Bologna, Italy, January 2005. [2] Lars Thiele, Michael Peter, and Volker Jungnickel. Statistics of the Ricean K-Factor at 5.2 GHz in an Urban Macro-Cell Scenario. 2006, yet unpublished. [3] http://www.channelsounder.de

The objective of the measurements presented in this contribution is the study of the wideband radio channel in the 40GHz band. There are not many references related to wideband characterization in the 40GHz band. The problem grows if we look for references including outdoor and indoor environments, as well as different antenna patterns and polarizations to increase the variability of channel characterization. So, we can conclude that the measurement campaign performed and presented in this contribution constitutes a complete research work developed for wideband characterization in this band.

Measurements were taken within three groups of environments, in order to get results for a variety of multipath situations: two indoor environments, three outdoor environments and one outdoor-indoor environment. Locations selected for measurement campaign reflect complex variety of rooms, buildings and places. The objective of this selection of different scenarios is to obtain consistent results relative to the behavior of the channel at 40 GHz, in wideband transmissions.

Two pattern antenna configurations have been selected for performing measurements: directive and omnidirectional. Due to the different magnitude and phase of reflection coefficients for vertical and horizontal polarization, millimetric wave channel characteristics also depend on the polarization of the antennas used. So, these antenna configurations have been combined with two orthogonal polarizations in the outdoor scenarios, and also in the special outdoor-indoor case. In indoor scenarios, only vertical polarization is used. Polarization is identical for both transmitting and receiving antennas.

A Swept-Time Delay Cross-Correlation (STDCC) sounder was built to carry out the measurement campaigns. This STDCC sounder transmits a PRBS sequence with a chip rate of 12.5ns. We have introduced changes to the classical configuration of this kind of sounders to improve the performance. The main of these novelties is the suppression of the sliding factor, by clocking transmitter and receiver PN sequences to the same rate. Another important change is the implementation of the receiver based in two down conversions to intermediate frequencies. We have demonstrated that although this technique with off-line crosscorrelation implies a non real time sounding, it provides an increased capacity for resolving multipath components.

The ranges of estimated time and frequency channel parameters achieved have been grouped in the three types of environments and summarized in Table I.

The estimated coherence bandwidths values are important to study the applications which can be implemented in the 40GHz frequency band. Results obtained from the measurement campaigns carried out show good expectations for the use of the 40GHz band to offer broadband services with high bit rates. To support a data rate of 155Mbps, the bandwidth should be of 155/4=38.75MHz for QPSK modulation; 155/16=9.7MHz for 16QAM modulation and 2.42MHz for 64QAM. Those values are compatible with the coherence bandwidths obtained in our measurements.

Abstract is included in the pdf attachment.

This paper presents results on mobile channel characterization for above the sea propagation paths at 1.9 GHz. The study is based on measurement campaigns conducted in various locations, carefully chosen to cover all kind of environments that can be met in the Aegean Sea. The purpose was to model the over-the-sea channel. The measurements were separated in five sets, representing five different propagation environments (four with line of sight-LOS and one with non line of sight-NLOS conditions). Thus, the conclusions derived for each group can be applied to any similar environment under investigation. The groups include narrow, average and wide sea passages with or without onshore vegetation, "open sea" environments, harbors and ports. The NLOS measurements formed the fifth group regardless of the surrounding environment. Generally NLOS propagation for these channels causes rapid worsening of the parameter values that allows us to study these cases as a unity.

This study presents conclusions and results on large scale as well as on small scale characterization (time dispersive and time variant nature of the channel).

Large scale modeling was accomplished with the use of the log-distance model. Attenuation factor n and shadowing factor s were estimated for each group of measurements. For LOS groups, n was estimated close to 3 while s varied from 1.5 to 4 dB depending on the complexity of the environment. NLOS measurements resulted higher values as expected (n = 3.6 and s=4.5 dB). Analysis showed that large scale behavior depends strongly on the environment and also that the results can correlate with some theoretical models, e.g. path losses at small Tx-Rx distance present similarities with the "plain earth" model.

As far as the time dispersive nature of the channel is concerned, the study focuses on the Power Delay Profile (PDP). Typical results on the average PDP for successive regions where the channel can be regarded wide sense stationary are presented in Fig 1.
PDP has a 'spiky' shape that cannot be described by simple mathematical expressions but can be explained by the nature of the channel. A simple methodology for predicting the delays of possible incoming signal echoes is also included in this study. Moreover, mean excess delay and rms delay spread were calculated and empirical cumulative density functions of these quantities for each group were estimated (Fig 2). Generally delay parameters for LOS remained at low values; however NLOS conditions cause a remarkable increase.

Finally the envelope distribution for each identified (direct or non-direct) path was estimated. It was concluded that the LOS direct path follows the Ricean distribution. Moreover, the first incoming echo for NLOS cases (possibly a diffracted version of the direct path) also follows Ricean distribution with a significant lower K-factor. Non-direct echoes at the following delay bins obey the Rayleigh distribution. Results on the average estimated K-factor for each group of measurements are also included in the study.

A number of communication systems operating in millimetre-wave bands have been proposed in the past few years in order to provide broadband services to residential users and companies. In these bands, the troposphere causes several propagation impairments. In particular, rain attenuation can compromise the link availability and, consequently, the quality of the service provided. More traditional systems used to rely on the provision of link margins, calculated on the basis of first-order statistics of rain attenuation. Recently-defined systems, such as the standards IEEE 802.16-2004 or ETSI DVB-S2, include adaptation techniques that compensate in real-time the variations in the propagation channel by adjusting the signal modulation/coding formats or the transmission power. The design and operation of these systems requires information on the dynamic characteristics of the fades caused by rain and other propagation impairments.

A propagation experiment is being carried out at Universidad Politecnica de Madrid, Spain, with the double objective of gathering experimental propagation data in millimeter-wave bands in the first place and, subsequently, gaining insight on the channel characteristics through the analysis of these data. The first objective is in itself important, as propagation data in these bands is scarce and not available for different climates. Because the propagation impairments are originated mainly in the atmosphere, data must be obtained at least for one full year, ideally throughout several years, to take into account the year-to-year variability of meteorological phenomena.

In this paper, the experiment objectives and set-up will be described. The set-up includes a terrestrial radio link at 38 GHz, a Ka-band satellite receiver, and a meteorological station that continuously records surface meteorological data, including rain rate with a tipping-bucket rain-gauge. The terrestrial link is deployed between two buildings in the campus. The path distance is approximately 1 km. The transmitter sends an unmodulated carrier, whose level is measured and registered at the receiver. Rain attenuation can be derived by comparison of the registered levels in the presence of rain, and in clear-sky. The system has no means to measure more subtle propagation effects, as for example gas attenuation. The link is operative since the end of 2004.

The satellite receiver measures the Eutelsat HB-6 Ka-band beacon at 19.7 GHz. Calculations have shown that the received level is sufficient to measure link characteristics for more than 99.99% of the year. The receiver is working since the beginning of 2006. The beacon receiver is able to measure rain attenuation and tropospheric scintillation.

In the months ahead, new meteorological equipment will be added, to better characterise the rain structure. The satellite receiver is also being completed by a radiometer in the same band, that should help to measure other propagation effects that need a more precise reference level, as gas or cloud attenuation. The radiometer is due to start operation in the Spring of 2006.

The paper is mainly focused on the experiment itself, describing the propagation effects to be measured, the function of each instrument, their interactions, the installation and calibration processes, and the final objectives of the project. Some preliminary results will also be presented, both from the terrestrial link and from the satellite receiver.

The BFWA (Broadband Fixed Wireless Access Network) systems are terrestrial cellular point-to-multipoint networks, wherein both the TSs (Terminal Station) and the BSs (Base Station) have fixed locations. The system is applicable to provide high-speed duplex communication for broadband services, i.e. Internet browsing, multiplayer video games, etc. Because of the applied wide frequency band the system might be also applicable to as feeder network of GSM, UMTS, B3G and/or other networks. In a BFWA network the evolved intra system SINR (Signal to Interference plus Noise Ratio) has the most effect to the quality of service. The evolving interference situations depend on more factors such as sector layout, applied duplex method, etc. At high frequency bands, i.e. 38GHz the SINR is relevant influenced by precipitation therefore the network planning procedures must consider it because a time-dependent rain event can cause SINR fluctuation, and that can result in termination of the radio links. Our previous results [2] in figures the time fluctuation of service quality at certain locations of the coverage are of the BFWA network caused by time dependent SINR related to the simulated moving rain cell can be seen. However, to predict the outage performance of the system, statistical analysis of the SINR fluctuation jointly at each certain TS location over the whole coverage is needed. In this paper we focus our attention on the investigation of effects of rain fading to the evolving SINR in a BFWA system. With computer simulation we will estimate the probability of outage of the communication at any terminal location. For this the distribution of rain intensity is available from measured point rain rate time data. Applying this distribution the two dimensional distribution of rain intensity over the BFWA service area could be estimated so the rain attenuations on the radio links (point-to-multipoint in a cell of BFWA network). Take into account these attenuations the statistical probability of the SINR values at each location in the service area are determinable. As a result an outage probability map could be generated, which could be useful in the network planning if the distribution of rain intensity is known.

In order for systems engineers to be able to determine optimum methods of mitigating the impairments caused by multipath propagation, it is essential for the transmission channel to be satisfactorily characterised. It shows also the extent to which propagation effects in a radio environment influence the capacity of wireless systems, and outlines possible countermeasures for mitigating multipath effects in such systems. By describing the various measurement techniques to sound the channel and the interrelationships between them, applicable for the determination of important channel design characteristics this paper will serve as a supplementary document for those available in the open literature with respect to radio channel characterisation with particular interest at millimetre waves (55-65GHz). The challenges of designing and constructing a channel sounding system at as high frequencies as 65GHz are discussed. The actual hardware as well as propagation measurements taken in a University building to characterise the channels transfer function, the delay spread and coherence bandwidth are presented.