Antenna measurement errors due to multi-path effects could be significant in low frequency spherical near-field test ranges. The manifestation of such effects can include increased gain variations in the antenna frequency response, distortions in the shape of the main beam or the sidelobes, appearance of unwanted or displaced sidelobes, degradation of the cross-polarization performance.

In the VHF and low-UHF bands, the cost of reducing multi-path by lining the anechoic chamber walls with low frequency absorber material can well be prohibitively expensive. This paper discusses the impacts on the far-field gain and pattern data due to the presence of multi-path effects during the spherical near-field acquisition.

Several approaches to reduce multi-path and antenna/probe interactions have been investigated:

Introduction
The precise measurement of antenna gain for antennas in the lower half of the VHF range presents many challenges. Half-wave dipoles or shorter antennas are often used as gain standards for metrological purposes. However these possess a near uniform H-plane pattern which creates difficulties in achieving a reflection free measurement site, particularly up to 150 MHz. By measuring VHF dipole antennas over a large flat outdoor ground plane, the reflections can be precisely controlled [1, 2].
A calculable dipole antenna has been developed which is used to provide traceability for Antenna Factor, which can be directly related to realised gain, to uncertainties as low as ± 0.15 dB. The calculation relies on the moment method, as applied in the Numerical Electromagnetic Code (NEC). NEC includes the mutual coupling of the antenna with its image in the ground plane, enabling antenna factor to be calculated for a given antenna height and polarisation.

Method
The site insertion loss (SIL) between a pair of dipoles above a ground plane is measured, and compared with the calculated value [1]. SIL is obtained by combining a NEC model for the dipole elements with the measured complex S-parameters of the baluns. The antenna factor is derived from the SIL value.
A UHF version was developed to cover mobile phone frequencies and great care was needed in minimising the feed gap of the dipole and reducing the bulk of dielectric support. It is a miniature version of the VHF version, Fig. 1. The lowest uncertainty for antenna factor was achieved for resonant dipoles, however NEC enabled uncertainties of less than ± 0.3 dB at the band ends over a 200% bandwidth: four dipoles with lengths resonant at 60, 180, 400 and 700 MHz were used to cover the frequency range 20 MHz to 1020 MHz. The uncertainty for the UHF antenna factor was less than ± 0.3 dB .

Conclusion.
Antenna gain at VHF can be measured with uncertainties as low as ± 0.15 dB using the calculable dipole antenna. This has enabled uncertainties in the measurement of dipole like antennas to be identified, such as unwanted reflections from the test site and antenna mount. An advantage of measuring dipole-like UHF antennas with a dipole reference antenna is the reduction of site errors by substitution of similar antennas. The agreement over a broad bandwidth between measurement and model has quantified the accuracy of the method-of-moments code. The presentation of this work will describe how the combination of a high quality ground plane and antenna design, and the availability of NEC, led to the conclusion that the dipole gain was most accurately characterised by the NEC calculation.
[1] Alexander M J, Salter M J, Loader B G and Knight D A, Broadband calculable dipole reference antennas, IEEE Trans. EMC, Vol. 44, No.1, pp 45-58, February 2002.
[2] Garn, H. F., Buchmayr, M., Müllner W. and Rasinger, J.,"Primary standards for antenna factor calibration in the frequency range 30 - 1000 MHz", IEEE Trans. Instrum. Meas., Vol. 46, No.2, April 1997, pp544-548. Figure 1. 20-1020 MHz antenna with dipole element resonant at 180 MHz and 0.85-2.2 GHz antenna with 2 GHz dipole element, and a 30 cm rule.

Millimeter-wave applications have become increasingly important in the last years for both military and civil applications. To mention only a few of them, let us remind the radio astronomy, the remote sensing of clouds and precipitation, the collision avoidance radars, and the mm-wave indoor LANs.

The characterization and diagnosis of antennas can be fruitfully carried out, at these frequencies, by means of near-field indoor measurements [1] and a Near-Field to Far-Field (NFFF) transformation technique.

Propagating the near-field to the far-field requires both amplitude and phase measurements. However, as well known [1], accurate mm-wave phase surveys are difficult and require expensive set-ups, so that phaseless antenna characterization techniques [2] are highly desirable.

In fact, since more than one decade, phaseless antenna diagnosis approaches are available throughout the literature [2]. However, defining reliable and accurate techniques at mm-waves is an awkward task due to the usually large electrical dimensions of the antennas to be tested and to the environmental noise. Therefore, the development of an effective diagnosis technique requires
i) collecting the least possible number of data needed to the far-field reconstruction;

ii) involving the least possible number of unknown parameters to describe the radiating system and to be determined from the measured data.

Such strategy provides reliability and accuracy to the algorithm, strengthens the approach against the environmental clutter and significantly reduces the overall measurement duration.

In this communication, we present a phaseless antenna characterization algorithm satisfying the two above-mentioned requirements. In particular, the algorithm is based on the aperture field expansion through the Prolate Spheroidal Wave Functions (PSWFs) and a non-uniform spatial sampling of the measured near-field amplitude. Accordingly, both the unknown and the data representations effectively exploit the available a priori information on the size of the radiating system. At variance with other methods up to now presented in the literature, the PSWFs allow representing "at best", namely with the least possible number of parameters, the space of "essentially bandlimited" and "essentially spatially limited" aperture fields. Moreover, the adopted non-uniform spatial sampling allows to represent the measured field through a "minimum" number of samples [3].

An extensive experimental analysis has been carried out to assess the algorithm performances, also in comparison to other available approaches. In this communication, the results concerning data collected at 100GHz are presented.

[1] J. Ala-Laurinaho, P. Foster, G. Junkin, T. Hirvonen, A. Lehto, D. Martin, A. Olver, R. Padman, C. Parini, A. Räisänen, T. Sehm, J. Tuovinen, R. Wylde, "Comparison of antenna measurement techniques for 200 to 1500 GHz", Proceedings of the 20th ESTEC Antenna Workshop on Millimetre Wave Antenna Technology and Antenna Measurement, Noordwijk, 18-20 June 1997, pp. 345-351.

[2] O.M. Bucci, G. D’Elia, G. Leone, R. Pierri, "Far field pattern detemination from the near-field amplitude on two surfaces", IEEE Trans. Antennas Prop., vol. AP-38, n. 11, pp. 1771-1779, 1990.

[3] O.M. Bucci, G. D’Elia, "Advanced sampling techniques in electromagnetics", in Review of Radio Sci. 1993-1996. London, UK: Oxford University Press, pp. 177-204.

An indoor localization and communication project is described that proposes to use RFID tags, placed in the building beforehand, as navigation waypoints for an inertial navigation system carried by a first responder.

RFID (radio-frequency identification) devices commonly are attached to persons or to moveable objects so that the objects can be tracked using fixed readers (special-purpose radios) at different locations. In this project, we explore the "flip side" of this practice. Our concept is that detection of RFID devices in known, fixed locations by a moving reader provides a precise indication of location for tracking the person or moving object that is carrying the reader. This information can then be used to correct for any errors of an inertial tracking system.

The reception of GPS signals is not reliable inside most buildings. Inertial tracking systems inherently drift over time and produce errors in position, especially for inexpensive and lightweight systems. Corrections, in the form of the insertion of known locations that have been reached, can be developed automatically by the detection of an RFID device, either by correlating the identity of the device with a table of locations or by reading the device's location from data stored on it. As shown in Figure 1 below, there are regions where each tag can be detected. Based on which tags the reader detects, the reader can be located as being within a certain region. For example in the figure below, if the reader detects tags 4, 5, and 6, then it must be in the 4,5,6 region. If the reader only detects tag 4, it must be in the region where only tag 4 is detected.

Many portable radio transceivers require very small antennas, which because of their small electrical aperture exhibit a very small or negligible gain. In such circumstances, the main objective of the designer is obtaining high radiation efficiency. However, the measurement of this parameter for the manufactured prototype creates a significant challenge. The reason is that connecting a small antenna to conventional test equipment, as used for "normal-size" antennas, may considerably change its radiation properties.

This problem can be overcome by using a measuring system, which has been invented and developed at the Wroclaw University of Technology. This system performs the measurement of radiation pattern of a small radiator by employing a battery supplied built-in VCO and a dielectric positioner. The measurements are accomplished in the spherical system of coordinates. The setup uses circularly polarized probes which ensure all linearly polarized components are taken into account during the measurement process. During the spherical scan, the same angular step ranging from 2° to 10° is used in the azimuth and in the elevation planes. The recorded data are added together by computing a discrete sum all over sampled points on the sphere. Radius of the sphere varies from 1.2 to 4.5 meters in the setup, but during each data recording remains constant. The adopted method requires the use of a calibration step, similarly as in the reverberation chamber method proposed by Swedish researchers a few years ago. The simplest calibration standard is a dipole antenna, which is tuned to the test frequency. For this case, the dipole's efficiency is approximately 100%. An alterative calibration procedure relies on the check of power levels recorded by a receiver which are referred to absolute power values.

A straightforward extension of the proposed system concerns the task of determining of power absorbed by a lossy object, which is placed near the antenna under test. If only relative values are to be derived, we compare the discrete sum values for the isolated antenna and for the antenna located next to the lossy object. Two measurement cycles are required. This leads to about an 8 hour measurement procedure for the 2° step. Examples of lossy objects that are tested include a phantom of a human head filled with a jelly substance, bars of strongly absorbing materials and an electronic equipment.

Fig. 1. 'Screen shot' of a software developed in house for the needs of determination, among other parameters, antenna efficiency by summing up energy over a sphere.

3-D pattern measurement in anechoic rooms for evaluating the Total Radiated Power (TRP) of mobile terminals is now widely used. In particular, standard or pre-standard documents precisely describe the way to perform such measurements, with a good accuracy. These techniques are generally based on the acquisition of a large number of points (more than 200), which leads to a prohibitive far-field assessment duration. However, complete 3-D measurements are not always necessary, particularly in the mobile phones design stage. Following this remark, the most direct idea is to only evaluate the far-field in a few "well-chosen" cut-planes of the measurement sphere. The objective of this paper is to show that this simple idea works well, and allows to obtain fast and accurate estimations of the TRP of mobile handsets.

1. Relevant definition of the 2-D Partial Radiated Power

Firstly, the concept of Partial Radiated Power (PRP) in a given cut-plane is briefly examined, showing that a degree of freedom, related to the choice of the surface of integration, appears in the definition of this quantity. It is highlighted that the PRP has to be normalized, such that the TRP can be seen as an average of the PRP, in the various cut-planes. In that case, a few "well-chosen" PRP samples can give a good estimation of the TRP.

2. Experimental validation of multi-cut 2-D estimation of the TRP

3-D TRP of about 30 mobile phones are compared to averages of 2-D PRP, in various cut-planes. This comparison is performed in the case of the GSM 900, DCS 1800 and PCS 1900 frequency bands. As an example, Fig. 1 illustrates the relevance of 2-D multi-cut estimations, for the tested terminals operating in the DCS band. The "full TRP" is considered to be obtained by measuring the PRP in 6 cut-planes, successively separated by a 30° angle. Fig. 1a represents the ratio of averaged PRP to full TRP, as a function of the number of cut-planes used (the considered cut-planes are adjacent), showing among others that two "non optimal" cut-planes allow to estimate the TRP with less than 0.8dB of deviation.

Fig. 1b represents the maximal deviation of the PRP averaged over two various cut-planes, with respect to the 3-D TRP. It is observed that the average of the PRP in two perpendicular planes gives an estimation of the TRP, deviating with less than 0.2dB to the full TRP value. It is shown that it is also the case for every mobile phone and bandwidth tested, when the E and H-planes (1 and 4) are used. These results should help among others in accelerating the design process of mobile phone antennas.

We propose a novel design for a vector measurement system for the characterization of Gaussian beams and test results for mm-wave receiver optics alignment across the 211-275 GHz band. Previously published work, by C.-Y.E. Tong et al. (2003), W. Jellema et al. (2005), on vector beam measurements employs phase-locked loops (PLLs) with Gunn-oscillators and/or multiple frequency sources. We are developing a measurement set-up without any PLLs and employ the combination of a single frequency source, comb-generator and direct multiplication LO unit. The design takes advantage of different harmonics to generate the RF and LO signals and to create the desired IF. Importantly, at the same time it allows obtaining a perfect phase-coherence and initial phase-noise cancellation.

In order to minimize the phase error we use a vector network analyzer (VNA) as a signal source. In the suggested scheme the signal from port 1 of the VNA generates a signal, f_source, which is fed into a comb-generator that generates number of frequencies ∆f_source apart. When f_source is low, this results in a large number of closely spaced, phase-coherent frequencies. Since any selection among the generated frequencies is phase-coherent, any two of them can be used to produce phase-coherent RF and LO signals by filtering and multiplication. As a test source we use a harmonic mixer mounted on a xyz-scanner. Its absolute position with respect to the receiver optics is predefined by means of triangulation system comprising lasers and position sensitive detectors. The IF from the SIS mixer is down-converted using suitable reference from the comb generator to produce an IF equal to f_source. This signal is fed into Port 2 of the VNA and amplitude and phase are measured.

The advantage of using comb generator is to have closely spaced frequencies to choose from and yielding the desired IF-frequency and obtaining initial-phase-noise cancellation. Most of the phase-noise is cancelled in the down-conversion in the SIS-mixer and the remains should be cancelled in the second down-conversion before the measurements in the VNA. The cancellation of the remaining phase-noise present at the IF is obtained by selecting also the LO, for the second down-conversion, from the harmonics generated by the comb-generator.

We also propose to use the same measurement set-up for the frequency bands 275-370 GHz and 385-500 GHz. We specifically designed our system such that it has a potential to cover all three bands by only replacing two filters and the LO multiplication unit. Scalar beam measurements up to 320 GHz have already demonstrated a dynamic range of about 20 dB with the harmonic mixer as the RF transmitting source [private communication by M. Panteleev]. We expect that vector measurements provide even greater dynamic range, indicating that we can push the frequency even higher with this harmonic mixer as the RF comb source.

Now, we are in the verification phase of the system design and intend to present results from measurements of the Band 1 of the Facility receiver for Atacama Pathfinder EXperiment (APEX) for 211-275 GHz.

One of the most important stimuli to the development of new antenna measurement techniques in recent years has been the progress in small antennas for wireless transceivers, such as cellular phones. In these applications, antennas are miniaturized to the highest possible degree in order to match the growing demand for very compact wireless and mobile portable terminals. This demand has activated the wide range of research activities in the field of designing small antennas. These designs are often carried out using very sophisticated and accurate antenna analysis and design computer aided packages. However, once these antennas are designed and developed the final step is to test them experimentally. This creates a considerable challenge, as the well-established antenna test procedures that are valid for "normal size" antennas are difficult to extend to their miniaturized counter parts. For example, the testing of normal size antennas can easily neglect the presence of connecting cables as they can be hidden in the antenna structure. This assumption does not apply to small antennas, as their dimensions can be comparable with the width of the connecting cable. The solution to this problem can be provided by the high fidelity radiation pattern measurement system, which has been developed at the Wroclaw University of Technology. The solution features two main attributes. One, the system uses a large dielectric positioner so that the small antenna under test is placed away from any disturbing objects. Two, to generate the radiating pattern, the antenna is fed with a minute VCO mounted back-to-back to a small antenna ground. The dc supply to the VCO is provided with a rechargeable battery, which is mounted in a cluster with the VCO and the antenna under test. This arrangement emulates the situation as it would be in a compact wireless transceiver such as cellular phone. The positioner, which is solely developed in a dielectric material, is shown in Fig. 1. Differences between the radiation patterns for a small antenna obtained with the use of the conventional and the newly developed radiation pattern measurement equipment are large. The differences between the two patterns are apparent, indicating the advantages of the newly developed small antenna test equipment. The differences are due to the absence or presence of the connecting cable. For completeness, an example of the spherical radiation pattern, as measured with the developed system, is shown in Fig. 2.

Fig. 1. A photograph of the antenna positioner of a measuring system, which was developed at Wroclaw University of Technology for the purpose of measuring radiation pattern of small antennas.

Fig. 2. A spherical radiation pattern of the TCR small antenna, as recorded at 2020 MHz by the system with a battery-powered built-in antenna source.

Ultra-large antennas with a diameter of several 100 m or even 1 km are required for microwave transmission aboard a solar power satellite, or satellite mobile communication systems. The array antenna of aperture elements was proposed for these purposes. But there tend to exist mechanical gaps or electromagnetic field dips between element apertures, which cause the growth of grating lobes and the reduction of antenna gain.

Accordingly, we propose to load parasitic elements of half wavelength dipoles to enhance the coupling and to bury the gaps between the apertures. We have studied the validity through experiment and analysis. The experimental system is composed of two parabolic reflectors for the reception of satellite broadcasting with specially designed primary radiators. The parasitic elements are installed between the apertures as shown in Fig. 1, and are illuminated by the radiated field from the apertures. It is important to adjust the phase relation between the directly radiated wave from the apertures and the scattered illumination wave by the dipoles.

In this paper, we propose two novel methods of phase adjustment as follows: (1) to change the lengths of scattering dipoles.
(2) To imbed the scattering dipoles in dielectric material.

The experimental results indicate that the method (1) is effective and preferable in practical applications. The analysis requires modeling and several contrivances in simulation, and eventually offers a valid tool for realization of the proposed antenna.

The presence of electrical and mechanical errors in an antenna is usually observed in its measured far-field pattern, but normally the source of such errors can only be identified by analyzing the antenna extreme near-field. Several non-invasive diagnostics techniques have been proposed over the years, being all methods limited either in terms of the type of antennas for which they can be used, or in terms of the accuracy that they can provide. There is thus a need for an antenna diagnostics technique that applies to general types of antennas, and that is intrinsically accurate.

We propose a new diagnostics technique to be applied at the DTU-ESA Spherical Near-Field Antenna Test Facility located at the Technical University of Denmark. The measurement technique employed at the DTU-ESA Facility is based on the Spherical Wave Expansion (SWE) of the field radiated by the antenna. For mathematical reasons the computation of the aperture field in the extreme near-field of the antenna can not be accomplished through the SWE. One way to overcome this is to transform the SWE of the radiated field into a Plane Wave Expansion (PWE). In particular, the plane wave spectrum can be computed by the knowledge of the coefficients of the SWE, on any aperture plane in the antenna source-free region. This gives two main advantages. The first is that the plane wave spectrum can be evaluated also in part of the spectrally invisible region, the second is that the aperture field can be computed as an Inverse Fourier Transform of this spectrum. Hence, the spatial resolution achieved in the aperture field can in principle exceed the traditional value of half a wavelength, provided by the traditional techniques.

While the fundamental properties of the SWE-PWE transformation have been described in previous articles, we will here concentrate on how real and non-ideal aspects, such as noise and finite dynamic range, affect the proposed diagnostics technique. To do that, we will consider antenna models characterized by different directivities and made by a combination of electric and magnetic Hertzian dipoles. The dynamic range of the measurement system will be varied and noise will be added to the radiated near-field. The effects of such quantities on the obtained Q coefficients and on the extreme near-field provided by the diagnostics will be studied. Finally errors will be introduced in the antenna model and the ability of the diagnostics technique in identifying them will be tested. Simulated results will be provided for the test cases.