Irwin O. Kennedy

Steady state RF fingerprinting for identity verification: one class classifier versus customized ensemble

Mobile phone proliferation and increasing broadband penetration presents the possibility of placing small cellular base stations within homes to act as local access points. This can potentially lead to a very large increase in authentication requests hitting the centralized authentication infrastructure unless access is mediated at a lower protocol level. A study was carried out to examine the effectiveness of using Support Vector Machines to accurately identify if a mobile phone should be allowed access to a local cellular base station using differences imbued upon the signal as it passes through the analogue stages of its radio transmitter. Whilst allowing prohibited transmitters to gain access at the local level is undesirable and costly, denying service to a permitted transmitter is simply unacceptable. Two different learning approaches were employed, the first using One Class Classifiers (OCCs) and the second using customized ensemble classifiers. OCCs were found to perform poorly, with a true positive (TP) rate of only 50% (where TP refers to correctly identifying a permitted transmitter) and a true negative (TN) rate of 98% (where TN refers to correctly identifying a prohibited transmitter). The customized ensemble classifier approach was found to considerably outperform the OCCs with a 97% TP rate and an 80% TN rate.

Power Reduction in Network Equipment Through Adaptive Partial Reconfiguration

We introduce a new approach to reducing FPGA power consumption. By exploiting the time varying nature of a systems environment, we are able to extract power consumption savings. We do this by closely tracking environmental changes and adapting the implementation accordingly using partial reconfiguration. We chose network infrastructure equipment to provide the context for the work since it is a significant consumer of FPGAs and is deployed in diverse environments. The network industry is also very interested in reducing FPGA power consumption as part of a major system wide effort, since it faces regulatory pressure, environmental concerns and rising electricity bills. We present a new experimental framework for measuring the power consumption of FPGA cores. The framework is used in an illustrative case study of how the approach works with a Viterbi decoder. The experiments give encouraging results and show that significant savings in power consumption can be obtained.

Radio Transmitter Fingerprinting: A Steady State Frequency Domain Approach

We present a novel technique for radio transmitter identification based on frequency domain characteristics. Our technique detects the unique features imbued in a signal as it passes through a transmit chain. We are the first to propose the use of discriminatory classifiers based on steady state spectral features. In laboratory experiments, we achieve 97% accuracy at 30 dB SNR and 66% accuracy at OdB SNR based on eight identical universal software radio peripherals (USRP) transmitters. Our technique can be implemented using todays low cost high-volume receivers and requires no manual performance tuning.

Feature extraction approaches to RF fingerprinting for device identification in femtocells

Radio frequency (RF) fingerprinting is a technique which attempts to extract a unique identifier from wireless signal transmissions in order to perform automated device identification by exploiting variations in the transmitted signal caused by hardware and manufacturing inconsistencies. The problem of signaling storms caused by increased core network signaling load due to idle mode cell camping on femtocells is addressed and a novel approach to RF fingerprinting is proposed as a solution to tackle this problem by identifying device model types. This study describes a large data set containing 54 Universal Mobile Telecommunications System (UMTS) user equipment (UE) devices (41 model types), the largest of its kind reported in the literature. This study also presents a novel feature extraction technique, which greatly improves identification accuracy over standard spectral approaches while limiting the number of random access channel (RACH) preambles required for identification. Accuracy of 99.8 percent was achieved. Importantly, the proposed technique can be implemented using todays low cost high-volume receivers and requires no manual performance tuning.

Passive Steady State RF Fingerprinting: A Cognitive Technique for Scalable Deployment of Co-Channel Femto Cell Underlays

Recently, cellular operators have begun evaluating femto cells that aggressively reuse spectrum to cover a small spatial footprint (10 m radius). A large-scale femto cell underlay network will increase an operators total number of cells by two to three orders of magnitude and presents significant scaling problems in spectrum reuse. We argue that to achieve this scale and interoperability with the existing UMTS network and handsets, the femto cell must use novel yet simple cognitive techniques: sensing, smart handover and idle mode cell camping. We report in detail on the problem of signaling storms caused by increased core network-signaling load due to idle mode cell camping. We provide a novel passive RF fingerprinting technique as a solution to tackle the problem. The technique is based on frequency domain characteristics. Our technique detects the unique characteristics imbued in a signal as it passes through a transmit chain. We are the first to propose the use of discriminatory classifiers based on steady state spectral features. In laboratory experiments, we achieve 91% accuracy at 15 dB SNR based on seven different models of UMTS user equipment. In the largest known laboratory experiment of its kind, we report an accuracy of 85% using our technique on twenty UMTS user equipment. This large test set includes 10 identical devices. Our technique can be implemented using todays low cost high-volume receivers and requires no manual performance tuning.

Using Support Vector Machines for Passive Steady State RF Fingerprinting

Passive steady state RF Fingerprinting has recently been proposed as a promising new method for identifying a radio transmitter. In essence, the algorithm detects the differences imbued on a signal as it passes through the analogue stages of a transmit chain. In this paper we improve the algorithms performance and scalability by proposing a new more sophisticated classification engine. The classifier engine is based on a one-against-one multi class support vector machine. We measure the improved system’s performance in the largest, most representative case study of its kind - 73,000 measurements across 41 models of UMTS user equipment (UE). We achieve 94.2% classification accuracy. In addition we provide detailed misclassification analysis and outline how the analysis can be used to considerably further improve overall classification accuracy.

RF Fingerprint detection in a wireless multipath channel

An RF Fingerprint is a unique signature passively imbued on a signal as it passes through a radio transmitter. We present the first RF Fingerprint detection technique capable of excellent performance in the presence of multipath propagation and the absence of a constant frequency offset. Our algorithm differs from previous work by performing joint channel estimation and classification on a steady state signal. The technique may be applied to any radio system with a repeated symbol sequence — such as a preamble. Our laboratory demonstrator is capable of distinguishing between the preamble signal transmitted by Universal Mobile Telephone System (UMTS) user equipment. We report excellent results — in excess of 99 % for 20 different UMTS models in an indoor wireless environment.

How to sleep, control and transfer data in an energy constrained wireless sensor network

We propose a communication theoretic framework to model energy consumption in ultra low-power transceivers. We use it to estimate the optimal time split between sleep, control (page, resource allocation etc.) and data transfer modes. We specify the necessary conditions for the sleep duty cycle to minimize energy consumption. We consider practical issues in existing transceivers, replace the information theoretic bounds with practical signaling techniques and re-estimate sleep duty cycle. We propose complementary network architectures to decouple access control from data transfer. We explore the design space captured by the framework with initial simulations and include a comparison of 802.15.4 against a fundamental energy limit. Results show a significant reduction in energy consumption and control overhead.

A cross layer design and evaluation of IEEE 802.15.4 network with an enhanced sensor gateway: Injecting hierarchy into wireless sensor networks

We propose an enhanced IEEE 802.15.4 gateway for use in a hierarchical network architecture. We perform a system comparison between it and a gateway using a normal sensor mote. This comparison includes a PHY and MAC simulation benefiting from ray trace propagation modeling of a large office building. The proposed enhanced gateway offers a 21.2 dB improvement in receiver sensitivity and an ideal channel equaliser. We analyze these benefits before performing a full system simulation. We report improvements over a classical sensor gateway in all key areas of packet delivery, latency and mote energy efficiency. Just under a six-fold increase in indoors range is also reported. We observe a reduction in performance due to channel distortion and combat it by proposing a fractionally spaced equaliser. We show that for buildings with large RMS delay spreads - greater than 100ns (e.g. conference centres), significant benefit may be derived.