Gerard B. Rowe

A Quantitative Analysis of Spectral Occupancy Measurements for Cognitive Radio

Cognitive radio (CR) is viewed as a possible solution to the problem of radio spectrum congestion. The CR technique utilizes a temporarily unoccupied licensed band. Before investigating the technical and political implications of CR, it is necessary to know to what extent the licensed bands are temporally unoccupied. This paper describes a spectral occupancy measurement campaign conducted in the frequency range between 806 MHz and 2750 MHz in urban Auckland, New Zealand. The purpose of the measurement is to identify potential spectral opportunities for CR. Statistical analysis of the measurement results are presented in the form of noise distributions, signal amplitude probability distributions, and spectral occupancy rates as percentages of time. These analyses indicate that, on average, the actual spectral usage in this band is only about 6.2%. Pint-to-point links and some mobile uplink channels are identified as the most probable candidates for future CR operations. These results suggest that CR could bring significant gains in spectral usage. These results also imply that, in order to provide a reliable detection of the primary signal, cooperative sensing techniques may be necessary to mitigate various wireless channel effects.

Modeling the Effects of Nearby Buildings on Inter-Floor Radio-Wave Propagation

Two buildings (A and B) have been modeled and analyzed with a 2D TEz implementation of the finite-difference time-domain (FDTD) algorithm in order to identify and characterize the mechanisms allowing signals to propagate between floors, specifically reflection and scattering from nearby buildings. Results have been extended to 2.5D by assuming isotropic spreading in the third dimension. In both scenarios considered, reflections from surrounding buildings are found to increase the average received power on adjacent floors-up to 9.7 dB and 32 dB for buildings A and B respectively. Measurements of the impulse response in Building A, made with a sliding correlator channel sounder, show a number of long-delay pulses, which can be attributed to specific reflection paths. Based on these findings, a simple two-component propagation model to predict the sector-average signal strengths is proposed and validated against measurements of the received power. The direct component is modeled as free space with a 22 dB/floor attenuation factor, and the reflected component is modeled as free space with reflection/transmission coefficients of 0.5. The RMS prediction error for this model is 3.2 dB.

Modeling Propagation in Multifloor Buildings Using the FDTD Method

A three-dimensional parallel implementation of the finite-difference time-domain (FDTD) method has been used to identify and isolate the dominant propagation mechanisms in a multistorey building at 1.0 GHz. A novel method to visualize energy flow by computing streamlines of the Poynting vector has been developed and used to determine the dominant propagation mechanisms within the building. It is found that the propagation mechanisms depend on the level of internal clutter modeled. Including metallic and lossy dielectric clutter in the environment increases attenuation on some propagation paths, thereby altering the dominant mechanisms observed. This causes increases in the sector-averaged path loss and changes the distance-dependency exponents across a floor from 2.2 to 2.7. The clutter also reduces Rician K-factors across the floor. Directly comparing sector-averaged path loss from the FDTD simulations with experimental measurements shows an RMS error of 14.4 dB when clutter is ignored. However, this is reduced to 10.5 dB when the clutter is included, suggesting that the effects of clutter should not be neglected when modeling propagation indoors.

An Investigation Into the Understanding and Skills of First-Year Electrical Engineering Students

In response to demands from industry and the profession for more graduates, first-year engineering numbers have grown considerably over the last decade, matched by an increasing diversity of academic backgrounds. In order to support first-year students effectively, and ensure the courses they take remain appropriately pitched, the academic preparedness of these students must be determined. Since 2007, the lecturers in the compulsory first-year Electrical and Digital Systems course at the University of Auckland (UoA), Auckland, New Zealand, have administered a short diagnostic test to determine the level of conceptual understanding of electricity and electromagnetics possessed by the incoming students. This paper presents and discusses student understanding of dc circuit theory as revealed by the diagnostic test and subsequent investigations. The evidence is indicative of both flawed conceptual models and context-triggered misapplication of fundamental rules. Parallels are drawn with the results of research conducted elsewhere, indicating the misconceptions are robust and pervasive, crossing institutional and national boundaries. Not only are concepts such as current and voltage poorly understood, but even more basic concepts such as series and parallel connections are confusing for a significant number of students. Understanding the incorrect models that underlie these basic misconceptions is the first step to correcting them. Only then can students proceed to the more advanced concepts that engineering graduates are required to master.

Outage Probability Estimation for Licensed Systems in the Presence of Cognitive Radio Interference

Cognitive radios (CRs) improve spectral efficiency by transmitting unlicensed signals over licensed frequency bands. A key requirement for CRs is to limit the interference to the licensed system. However, the variability in the radio channel caused by fading and shadowing makes this task more difficult. In this paper the performance of the licensed system has been analysed by calculating the probability of outage of the licensed receiver (LR). A Poisson field of CRs is considered such that the CRs are distributed uniformly around the LR. A circular exclusion zone of radius R has been defined with the LR at the center so that CRs are permitted to transmit only if they are outside this zone. For the first time, expressions for outage probabilities have been developed for different values of R, taking into account the variability in the radio channel. Rayleigh, lognormal and Suzuki channels have been analysed. The effects of channel characteristics on the outage probability have been investigated. Results show the extent to which increasing R improves the performance of the licensed system significantly.

Physical layer engineering for indoor wireless systems in the twenty-first century

Deploying wireless systems so that they offer the best performance for all stakeholders (e.g. users, service providers and shareholders) is identified as a critical issue in advancing the wireless industry into the second decade of the twenty-first century. An historical perspective of the key issues is initially presented, from which a set of societal expectations of future wireless performance are derived. The engineering challenge is to meet these expectations, and this has motivated a programme of research being undertaken at the universities of Auckland and Kent. Suggestions are made as to how this research might be encapsulated in the form of a wireless Code of Practice to assist practitioners charged to provide high performance wireless functionality to users in built environments.

Spectrum sensing using principal component analysis

In the recent past considerable research has been performed on blind signal detection techniques that exploit the covariance matrix of the signals received at a cognitive radio (CR). These techniques overcome the noise uncertainty problem of the energy detection (ED) method and can even perform better than ED for correlated signals. Contrary to the previous work where the main evaluation technique has been theoretical analysis and simulations, in this paper we use Software defined radios (SDRs) with correlated signal reception capability to evaluate the sensing performance of the existing covariance based detection (CBD) techniques. The existing techniques considered in this work are; Covariance absolute value (CAV), Maximum-minimum eigenvalue (MME), Energy with minimum eigenvalue (EME) and Maximum eigenvalue detection (MED). Most importantly this paper presents a novel technique for blind signal detection that uses Principal Component (PC) Analysis. The PC based signal detection algorithm and the CBD algorithms are tested in a real scenario with SDRs and their sensing performance is compared. The PC algorithm outperforms the MED and EME algorithms under all conditions and it performs better than the MME and CAV algorithms under certain conditions.

Blind Signal Detection Using a Linear Antenna Array: An Experimental Approach

An experimental investigation is reported on the use of a linear antenna array to perform blind signal detection. A software-defined-radio (SDR)-based receiver system is used to capture signals received by an antenna array while ensuring time and frequency synchronization across the radio-frequency (RF) front ends. The antenna array is moved in circular and linear paths to cause variation in received signal strength and to cause variation in the incident angles of the incoming waves. Covariance-based detection (CBD) algorithms are applied to the received signals to perform blind signal detection. Multiple antennas serve to enhance the received signal correlation, thereby causing an improvement in the detection performance of the CBD algorithms. In the presence of noise calibration error (which is inevitable in a practical system), the maximum eigenvalue of the correlation matrix (MEC), Hadamard ratio test (HRT), and covariance absolute value (CAV) algorithms exhibit a significant improvement in detection performance as the number of antennas is increased, whereas the eigenvalue-based detection (EBD) algorithms show no improvement. This paper also investigates the effect of varying the antenna spacing on the received signal correlation and its subsequent effect on the detection performance. It is observed that the detection performance of the linear array is directly related to the mean signal cross-correlation achieved by the array. A significant improvement in detection performance was observed for an antenna spacing of 0.1λ; however, to avoid degradation of antenna efficiency at such close spacing, proper impedance matching must be performed.

Estimating Protection Distances in Spectrum Sharing Systems

In this paper, we analyze spectrum sharing on broadcast primary frequency bands. We employ protection regions (with size defined by protection distances) around the broadcast transmitter service contours for meeting the primary system outage constraints, hence implementing the Federal Communications Commissions regulations. For the unlicensed nodes distributed in a Poisson Point Process, it is shown that exact protection distances cannot be analytically determined. In this paper, tools for estimating the protection distances are presented. Previously we have shown that combinations of lognormal and shifted-lognormal distributions always overestimate protection distances, hence guaranteeing primary system protection. In this paper, we show that combinations of gamma and shifted-lognormal distributions as well as the Edgeworth expansion also guarantee primary system protection. A comparative study of the three techniques and their comparison with the protection distances obtained from the Markov and Chebychev inequality is performed. It is shown that combinations of lognormal and shifted-lognormal distributions provide the most efficient spectrum utilization by the unlicensed system.

Adjacent Channel Operation of Portable Unlicensed Devices inside DTV Service Contours

The operation of unlicensed devices (UDs) on underutilized licensed frequency bands can significantly enhance spectrum utilization. The FCC regulations on UD operation on vacant TV channels focus on controlling the interference to the primary system, which is a critical operational requirement for UDs. Personal/portable UDs are allowed adjacent channel operation inside the TV service contour. However, the FCC analysis of this operation is based on interference effects of a single UD, while the accumulative interference from multiple UDs is not considered. In this paper we extend the FCC analysis and develop a realistic model for interference from multiple UDs operating in Wireless Local Area Network (WLAN) clusters. We develop analytical expressions for the probability of outage of a primary user (PU) interfered by single and multiple clusters. The effect of channel variability due to shadowing is included. Results present the increase in PU outage probability with increased PU-TV station distance. For a single cluster operating in close vicinity of a PU, we analyse the effects of UD transmit power control and radiation pattern on PU performance.