Session 1A SDR Architectures and Implementations 1
Software Defined Visible Light Communication
Michael Rahaim (Boston University & NSF Smart Lighting ERC, USA); Ali Mirvakili (Tufts University, USA); Sagar Ray (RPI, USA); Valencia J Koomson (Tufts University, USA); Mona Hella (RPI, USA); Thomas DC Little (Boston University & NSF Smart Lighting ERC, USA)
Software Defined Radio (SDR) has been shown to be an effective and practical tool in the area of RF communications, essentially allowing flexible and rapid exploration of RF signal processing techniques while accelerating the advancement of configurable RF antennas and front-end hardware. The software defined concept can also be adapted to other physical communication media; we consider an optical channel using the visible spectrum and describe the issues and challenges in adopting SDR for this medium. Optical Wireless (OW) communications, more specifically Visible Light Communications (VLC), have gained recent interest as dual-purpose lighting and communication systems due to the ubiquity of emerging solid-state lighting. OW systems typically implement Intensity Modulation with Direct Detection (IM/DD); therefore limitations in LED bandwidth constrain modulation to low frequency or baseband techniques. In addition, optical devices typically have nonlinear electro-optic conversion characteristics and require a positive drive signal; therefore device-dependent signal predistortion and proper biasing is needed. Finally, OW directionality and line-of-sight constraints support the use of multi-channel optical diversity receivers. We investigate and propose a Software Defined VLC (SDVLC) solution implementing an optical front-end to adapt SDR platforms to the constraints of an OW channel. Utilization of an SDVLC system allows for parallel development of signal processing techniques and front-end hardware within an integrated testbed. This modularity, along with the ability to quickly bring up an OW system and implement new test scenarios, makes SDVLC a powerful concept for facilitation of future VLC research. The presented implementation is an instance of this novel concept.
Software Defined Radio based Global Sensor Network Architecture
David Haley (University of South Australia, Australia); Linda M. Davis (University of South Australia, Australia); André Pollok (University of South Australia, Australia); Ying Chen (University of South Australia, Australia); Gottfried Lechner (University of South Australia, Australia); Marc Lavenant (University of South Australia, Australia); Adrian Barbulescu (University of South Australia, Australia); John Buetefuer (University of South Australia, Australia); William G Cowley (University of South Australia, Australia); Alex Grant (University of South Australia, Australia); Terry Kemp (University of South Australia, Australia); Ingmar Land (University of South Australia, Australia); Rick Luppino (University of South Australia, Australia); Robby G. McKilliam (University of South Australia, Australia); Hidayat Soetiyono (University of South Australia, Australia)
The Institute for Telecommunications Research has led a consortium to develop a Global Sensor Network architecture for remote sensor data gathering and communication. The research has focused on a Low Earth Orbit (LEO) micro-satellite system, with support from the Australian federal government's Australian Space Research Program. We have explored the use of software defined radio (SDR) technology for the space segment, ground station and terminals. We have developed novel techniques for highly efficient one and two-way data communication with large numbers of remotely located sensors and devices. The system includes new architectures and waveform designs and makes innovative use of SDR. The end result is a cost effective, scalable and flexible system that is able to support very large numbers of users while requiring only a small amount of radio bandwidth. Development has been driven using a living reference model as the basis for an agile SDR methodology. This approach has enabled rapid transfer of fundamental research into a working system. A successful bench demonstration with space hardware was achieved within 12 months of commencing waveform design. Less than 12 months later the system was field trial proven using an aircraft as a satellite surrogate. Recently the system has been validated using a LEO satellite. This paper describes the SDR system architecture, development and test methodology, and presents results from the field proven system.
FPGA-based Implementation of Multiple PHY Layers of IEEE 802.15.4 Targeting SDR Platform
While SDR platforms become more and more accessible to a large community of researchers with affordable prices, open source FPGA-based implementation of wireless communication systems are still not available. This paper presents an open source FPGA-based design of IEEE 802.15.4 PHY Layers developed in the context of a new SDR testbed named FIT/CorteXlab. We propose a VHDL implementation of the three available options of the IEEE 802.15.4 physical layer parametrized and easily reconfigurable. We have validated our design on Nutaq platform which combines Xilinx Virtex-6 FPGA and tunable Radio420x RF transceiver. A tutorial style approach is adopted to describe the multiple PHY Layers of both the baseband TX and RX IPs of IEEE 802.15.4 standard. More focus is given to the symbol timing/carrier recovery and considerations for FPGA implementation are outlined. The IPs presented here will participate to the building of an open source hardware SDR library similar to GNU radio but targeted to FPGA-based platforms.
Strategies for Pushing Software Radio Closer to the Antenna
Rodger Hosking (Pentek, USA)
New technology offers software defined radio system engineers many diverse opportunities to apply digital signal processing much closer to the antenna than ever before. Various strategies include the latest wideband data converters, monolithic receiver chips, compact RF tuners, new FPGA families, and remote data acquisition modules using gigabit serial interfaces. Each approach presents benefits and tradeoffs that must be considered in choosing the optimal solution for a given application.
Session 2A CR and DSA Architectures and Systems I
The FP7 CogEU TV White Space Radio Transceiver
Paul D Sutton (Trinity College Dublin, Ireland); Timothy K. Forde (University of Dublin, Trinity College, Ireland); Justin Tallon (University of Dublin , Trinity College & CTVR, Ireland); Jose Ribeiro (Instituto de Telecomunicacoes Aveiro Portugal, Portugal); Paulo Marques (Instituto de Telecomunicações, Portugal); Pawel Kryszkiewicz (Poznan University of Technology, Poland); Linda Doyle (Trinity College Dublin, Ireland)
COGnitive radio systems for efficient sharing of TV white spaces in EUropean context (CogEU) is a three-year Specific Target Research Project (STREP) supported by the EU 7th Framework Programme which concluded in 2013. This paper presents the CogEU TV white space radio transceiver, a key demonstrator and deliverable of the project. The demonstrator consists of an SDR platform using the Iris software radio architecture for dynamic spectrum access using spectrally shaped, non-contiguous waveforms with intentionally embedded cyclostationary signatures for network rendezvous and coordination. A TVWS database is used for channel selection and a dedicated wireless microphone detector is integrated to ensure coexistence with PMSE devices. An overview of the system architecture is provided and each of the algorithms used for microphone detection, transmit waveform shaping, network rendezvous and synchronization is described. Finally, the paper discusses the key challenges faced in the system implementation and details the solutions which were applied.
On the Requirements of Dynamic Radio Frequency Mapping
Garrett Vanhoy (University of Arizona, USA); Haris Volos (University of Arizona, USA); Mohammed Hirzallah (University of Arizona, USA); Carlos E. Caicedo Bastidas (Syracuse University, USA); Tamal Bose (University of Arizona, USA)
Intelligent cognitive radio systems require information about their environment to make operational decisions. Dynamic Radio Frequency (RF) mapping provides estimates of the RF power levels over an area where spectrum activity or changes in the environment may be transient. These power levels can be used for a variety of applications such as interference management, spectrum policing, facilitating spectrum auctions, etc. The RF mapping can be accomplished by a network of sensors that are distributed in the geographical area and using them to sample the signal level spatially. Previous works have proposed a number of methods to accomplish this; however, the estimations were performed using band-limited RF maps as models. Many actual RF maps are not band-limited due to small-scale fading and there is currently no practical way to band-limit the RF map. As a result, the estimated maps will suffer from aliasing estimation errors. In this paper, we quantify the effect of aliasing on the estimation of an RF map as a function of the sampling density and the number of antennas (local samples) used at the sensing node. Our results show that adding 1 uncorrelated local sample per sampling location reduces the estimation error by approximately 25% and by adding 11 local samples by 50%.
An Orthogonal Spectrum Sharing Scheme for Cognitive LTE Networks
Aditya V Padaki (Virginia Tech, USA); Avik Sengupta (Virginia Tech, USA); Mahi Abdelbar (Virginia Tech & Wireless@VT, USA); Jeffrey Reed (Virginia Tech, USA); William H. Tranter (Virginia Tech, USA)
Spectrum sharing has been proposed as a solution to the problem of under-utilization of licensed spectrum. It has the potential of not only increasing efficient spectrum utilization, but also of increasing revenue for cellular operators who can lease out spectrum at times of high demand to other operators. In this paper, we develop a novel architecture for spectrum sharing at the base station level for LTE-A cellular operators. The operators, acting as a primary user in their licensed spectrum, have a choice of dynamically sharing parts, or all, of their spectrum with co-located/adjacent secondary co-operators. Such an approach, as opposed to the conventional static band access, has two fold advantages. Firstly, it ensures maximum spectral utilization, thus increasing spectral efficiency. Secondly, it increases the flexibility at eNBs providing higher opportunity to use alternate channels to improve system throughput. We propose an orthogonal spectrum sharing and resource allocation scheme, focusing on rate maximization with a minimum per-user rate constraint. A linear optimization framework at the network level is developed for maximization of throughout and spectrum access cost. We study a use case with two operators and two cells, for proof of concept. We obtain analytical solutions to the rate maximization problem and show that a mutually beneficial secondary spectrum access cost exists for both operators as a trade-off with the combined system sum throughput.
Design of a Self-Organising Wireless Network Improving Application QoE in Shared Frequency Band
Masayuki Ariyoshi (Advanced Telecommunications Research Institute International (ATR) & NEC Corporation, Japan); Kazuto Yano (ATR, Japan); Mariko Sekiguchi (Advanced Telecommunications Research International, Japan); Tomohiro Miyasaka (ATR, Japan); Kiyoshi Kobayashi (Advanced Telecommunications Research Institute International, Japan)
We have proposed a new self-organising wireless network on which applications on many terminals can run in high quality even in congested frequency bands such as 2.4 GHz. The system is designed by taking users' Quality of Experience (QoE) as a key metric. In this paper, we present the concept and design of the proposed wireless network with QoE-based control, which consists of three major functional elements: smart multi-layer sensing; prediction of achievable application quality; and radio channel access management. The smart multi-layer sensing measures available radio resources and estimates radio systems (WiFi, Bluetooth, ZigBee, etc) as well as their traffic types. Using those sensed and estimated results, amount of achievable traffic considering required application quality is calculated by the prediction function. The radio channel access management function decides and allows channel access for applications depending on those priorities. We will show the detailed design of our prototype of the system and evaluation results on improvement in application quality with the proposed system.
Session 3A SDR, CR and DSA Algorithms and Implementations
An Efficient Digital Pre-Distortion Filter For Ultra-Wideband QAM Modems
frederic j harris (San Diego State Univ, USA); Xiaofei Chen (San Diego State University, USA); Elettra Venosa (San Diego State University, USA); Bhaskar Rao (University of California, San Diego, USA)
The digital shaping filter in a DSP based transmitter has to perform pre-distortion tasks to account for the distortion introduced by the digital-to-analog converter (DAC) and analog smoothing filter. The DAC introduces SINC spectral response and the analog smoothing filter introduces non-uniform gain and phase distortion. This pre-distortion problem was well solved by cascading the pre-distortion filters after the square-root raised cosine (SRRC) shaping filter. However, this traditional solution fails in the wideband scenario due to the heavy computational burden. For example, any digital filtering on a 500 MHz sample rate signal is problematic due to the limited hardware processing speed. In order to build realizable and cost-effective modems for future wideband systems, the authors propose a new solution which takes advantage of the spectral shaping property of the perfect reconstruction (PR) non-maximally decimated filter bank (NMDFB) which allows the hardware to operate at significantly lowered speed and avoids the needs for a highly oversampled DAC.
Decoding of DVB-T2 LDPC codes on a Tilera processor: Optimizations and performance comparisons
Sudeep Kanur (Åbo Akademi University, Finland); Stefan Grönroos (Åbo Akademi University, Finland); Kristian Nybom (Åbo Akademi University, Finland); Jerker Björkqvist (Åbo Akademi University, Finland); Johan Lilius (TUCS and Abo Akademi, Finland)
The second generation Digital Video Broadcasting (DVB) standard for terrestrial (DVB-T2), satellite (DVB-S2) and cable (DVB-C2) broadcasting media specify the use of capacity approaching Low-Density Parity-Check (LDPC) codes for forward error correction. The standard specifies two different codeword lengths of size 16200 bits and 64800 bits with multiple code rates. Software defined radios can benefit from real-time decoding of these codes on general purpose computing hardware. However, iterative decoding algorithm for LDPC is a NP-complete problem. Multiple code rates and codeword lengths coupled by the inherent problem with decoding LDPC codes that they generate irregular memory access has posed significant implementation challenges. Recent advancements in multicore processors have enabled massively parallel computation that can, in some cases, out perform general purpose Central Processing Units. Irregular memory access and communication network bottleneck is handled by achieving a balance between computation and communication load between processors. Presented in this paper are the algorithms and the data structures used to implement log-domain decoding of the long LDPC codes specified by the DVB standards on a Tilera TILEPro64 processor, featuring 64 RISC cores interconnected by an on-chip network. In addition, the paper also compares the throughput and performance of decoding against a similar implementation on a modern multicore central processing and a graphics processing unit.
Viterbi decoding employing a heterogeneous cognitive radio for public mobile communication systems
Masafumi Moriyama (The University of Electro-Communications, Japan); Takeo Fujii (The University of Electro-Communications, Japan)
To improve reliability and service coverage of public mobile communication systems (PMCS), which is used for communication in public safety systems, we consider combining the PMCS with commercial systems such as cellular systems based on heterogeneous cognitive radio techniques. More specifically, when bit error rate (BER) performance of the PMCS degrades in the cognitive radio terminal constituted by the PMCS and the commercial systems, a base station of the commercial systems will transmit the information to improve PMCS' communication quality. This information is a part of code-words of PMCS' information and is called subsidiary information. The cognitive radio terminal receives the subsidiary information, and then makes synthesized signal by combining this subsidiary information and received signal transmitted from a base station of the PMCS. In this paper, we propose a decoding method for the synthesized signal, employing the maximum likelihood sequence estimation (MLSE) for the convolutional code in the cognitive radio terminal. Then, we customize the Viterbi algorism and derive distance spectrum, which represents capability of the convolutional code. From the distance spectrum, we can derive theoretical BER performance of the proposed decoding method. The theoretical BER closely approximates BER determined by computer simulation. Finally, the BER performance of the proposed system is analyzed and evaluated.
Hardware Implementation of Gold's Algorithm for Rendezvous in Adaptable FH Cognitive Radio Networks
Curtis Medve (Air Force Institute of Technology, USA); Michael Seery (Air Force Institute of Technology, USA); Mark D. Silvius (Air Force Institute of Technology, USA); Robert J McTasney (University of Colorado at Boulder, USA); Kenneth Mark Hopkinson (Air Force Institute of Technology, USA)
Adaptive frequency hopping cognitive radios provide robust performance in rapidly changing and contested spectrum environments. The typical assumption that frequency hopping networks require a common control channel may be eliminated by using Gold's algorithm for rendezvous and synchronization. This paper presents a novel hardware implementation of Gold's algorithm, and demonstrates its feasibility and significance for reducing synchronization time in adaptive cognitive radio systems. First, this paper reviews the mathematical operations for Gold's algorithm. Second, it discusses how these operations are represented in VHDL for implementation as a custom intellectual property core on an FPGA. This core is optimized for speed using an iterative programming approach. Third, the core is targeted to the Xilinx Virtex-5 ML507 FPGA development board and its embedded PowerPC440 processor. Fourth, a proof-of-concept testbed is constructed using a transmitter/receiver pair implemented on the ML507 hardware. RF carrier frequencies are modeled in the audible spectrum to simplify prototyping and testing. A master transmits a sequence of audio tones based on a linear feedback register pseudo-noise (PN) generator. A slave is used as an acquisition device which listens to the transmitted tones, estimates the PN sequence, and synchronizes the receiver to the transmitter. In future work, the Gold's algorithm core will be targeted toward Rice University's Wireless Open-Access Research Platform (WARP) with an adaptive RF frequency hopping modem, and the rendezvous algorithm will be tested in the RF spectrum in a laboratory setting.
Session 4A SDR Architectures and Implementations II
Interference Mitigation using Adaptive Polarization
Robert Conley (Eigen Wireless, USA); Steven Schennum (Gonzaga University, USA); Tim Hillstrom (Eigen Wireless, USA)
Many mission critical communications standards such as Project 25, TETRA and IEEE 802.15.4g (Smart Utility Network) were defined as Single Input Single Output (SISO) systems and thus lack provisions for smart multiport antennas. These and other standards are often narrow band, lack coding techniques and may implement non-coherent modulation. Such systems are increasingly overwhelmed by interference from network congestion, secondary users and out-of-band induced receiver intermodulation products. Furthermore they operate in highly scattered environments that include narrowband multipath fading and polarization dependent losses. In this paper, a dual receiver architecture is presented with baseband algorithms implemented in an SDR platform along with unobtrusive antenna arrays that address all of these issues resulting in an average 23 dB Signal to Interference plus Noise Ratio (SINR) link budget improvements for mission critical communication systems. Also addressed is the requirement for new receiver performance criteria and associated measurement techniques needed to validate cognitive receivers as they enter commercial production.
A PR NMDFB Based Carrier Recovery Architecture for Next Generation Wideband Modems
frederic j harris (San Diego State Univ, USA); Elettra Venosa (San Diego State University, USA); Xiaofei Chen (San Diego State University, USA)
Currently, the hardware capabilities set the limit on the bandwidth that can be processed digitally. Data must be sampled and processed at least at two samples per symbol. ADCs offer sample rates which are of the order of a few Gsps. However, at such high speeds even the simplest filtering task can saturate the hardware processing limits and this prevents wideband processing. Perfect reconstruction (PR) non-maximally decimated filter bank (NMDFB) receivers overcome this problem removing the constraint on the bandwidth of the signal that can be processed. In this paper the authors propose a wideband receiver based on PR NMDFB. The main focus is the carrier recovery task, which traditionally requires a pair of band-edge (BE) filters and a square-root raised cosine (SRRC) filter, plus a frequency locked loop to drive the carrier offset error to zero. The conventional approach, thus requires three FIR filters operating at the input sampling rate. When the input sample rate becomes higher, which is the case in wideband receivers, the filtering tasks become a heavy burden for any hardware processor. In this paper we propose a novel carrier recovery architecture which is embedded in a wideband PR NMDFB receiver. The carrier recovery is implemented by performing BE and SRRC filtering in the channelized domain. The PR NMDFB provides the capability of simultaneously implementing the three required filters while allowing the signal processing to be performed on deeply decimated sampling rate, which enables removing the limits on the input bandwidth in the next generation modems.
Information Assurance of LTE-Advanced Self-Organizing Networks
Munawwar Sohul (Virginia Tech, USA); Raghuprasad Bettadapura (Virginia Tech, USA); Aman Singhal (Virginia Tech, USA); Jeffrey Reed (Virginia Tech, USA)
One of the most anticipated features of LTE-Advanced is Self-Organizing Network (SON) - the capability to configure and optimize the network automatically by adapting to network characteristics and failures. This capability results in reduced human intervention in network operations and is essential to achieve the capacity promise of 4G, especially for small cell deployments in a Het-Net scenario. However, SON functionalities introduce new network vulnerabilities. Network intelligence and behavior characterization could reduce throughput. Node emulation attack could decreases cell coverage, affect scheduling algorithms and serving node preferences. Automation requires measuring the environment through sensing, but an intruder can modify these measurements and control reports. Although the security implication of SON is a very important aspect, there has not been any significant research done so far. This paper investigates the security vulnerabilities of SON and identifies the point of vulnerabilities for intrusion, and possible security threats and their expected impact on system performance. A focused analysis on the impact of node emulation attack on the Inter-Cell Interference Coordination (ICIC) is presented here. Anomaly detection is used as the intrusion detection technique to identify the malicious intrusion. We devised two anomaly detectors: correlation band detection with low probabilities of false alarm and missed detection and threshold detection. Based on the decisions from the intrusion detectors necessary modification was applied to the system parameters to recover from the attack. Simulation results show reasonable improvement in system performance in terms of cell-edge users SIR and throughput due to the proposed detection and recovery scheme.
DO-178 and Common Criteria - How to Combine both Design Assurance Standards
The standards DO-178B/C, DO-254, and Common Criteria (CC) both focus on design assurance albeit in different domains. While DO-178B/C and DO-254 focus on safety the Common Criteria standard is concerned with security. It is not uncommon, though, that both safety and security of an SDR need to be evaluated by certification authorities, so in order to keep the development effort at bay it is advisable to efficiently combine the evidences that DO- and CC-standards demand. This paper explains the similarities as well as the differences of DO and CC and suggests an easy-to-follow developmental approach which both makes use of the similarity-overlap and mitigates the effects resulting from the differences.
Session 5A CR and DSA Architectures and Systems II
Learning Characterization Framework and Analysis for a Meta-Cognitive Radio Engine
Hamed Asadi (University of Arizona, USA); Haris Volos (University of Arizona, USA); Michael Marefat (University of Arizona, USA); Tamal Bose (University of Arizona, USA)
Cognitive radio engine (CE) designers are continuously working to understand and develop better learning techniques to be employed within their designs. Each of these techniques has its strengths and deficiencies depending on the application scenario. Distinguishing when each learning technique should be used is a complex and time-consuming task and currently CE's do not have this ability. The main objective of this paper is to demonstrate that a CE with meta-abilities is able to choose which of the learning techniques are better suitable for the various scenarios. The concept of a meta-CE is not new; however, no significant progress was made that provides the analytical tools necessary for a strong meta-CE. In this paper, first we characterize different wireless scenarios (e.g. high/low SNR, available bandwidth, and primary user activity.) By using the scenario characterization, we are able to evaluate the effectiveness of each learning technique in terms of the characterized scenario. This analysis allows the meta-CE to differentiate the learning techniques. By enabling the meta-CE to evaluate each learning technique, it also paves the way for equipping the meta-CE with the techniques needed to develop and evaluate new learning techniques that can be used during the meta-CE's operation. Our results show, by using meta-CE techniques the CR achieves the highest performance rate 80% of the time. Finally, we test three different learning techniques in 100 different channel situations from low SNR to high SNR; we identify and explain the meta-CE work flow, and evaluate the performance of our meta-CE techniques.
Development of Measurement based Spectrum Database for Efficient Spectrum Sharing
Masayuki Kitamura (The University of Electro-Communications, Japan); Takeo Fujii (The University of Electro-Communications, Japan); Kei Inage (The University of Electro-Communications, Japan); Yuya Ohue (The University of Electro-Communications, Japan)
In order to realize a frequency sharing by cognitive radio, the method of Primary User (PU) protection is the most important function. As one of protecting methods, Secondary User (SU) cooperation with external information of spectrum is focused in these days because high reliable PU protection can be realized. Currently, geolocation database which stores the spectrum availability information calculated according to the propagation model for protecting PU has been considered. This type of database is well known as FCC database of TV white space. However, since propagation model cannot perfectly model an actual wireless environment, extra margin for protecting PU is required and SU communication opportunity is reduced. Conversely, if the measured data can be used for creating the database, more accurate propagation environment and more SU communication opportunity can be obtained with keeping PU protection. Therefore, in this research, we develop a radio environment database that is stored a statistical information of radio environment measured and reported by SU. We implement the measurement device for SU based on USRP and actual radio environment is measured to test the developed database. Here, the stored value of database is the averaged received signal power with measured location. The server analyzes the measured data for deriving their stochastic information. We set measurement device on the vehicles and measure the actual radio environment for testing the developed radio environment database. From experimental results, we can clarify the difference between the current geolocation database and the developed radio environment database based on measurement value.
Enhanced and Embedded FPGA GNU Radio Flow
Ryan Marlow (Virginia Tech, USA); Peter Athanas (Virginia Tech, USA)
This paper presents a Zynq capable version of GNU Radio with an enhanced flow that utilizes the processing capability of FPGAs. This work features TFlow -- an FPGA back-end compilation accelerator for instant FPGA assembly. The Xilinx Zynq FPGA architecture integrates the FPGA fabric and CPU onto a single chip, which eliminates the need of a controlling host computer; thus, providing a single portable, low-power, embedded platform. By exploiting the computational advantages of FPGAs in the GNU Radio flow, a larger class of software defined radios can be implemented. Once the FPGA is programmed with a design, modules can be altered with specific sets of parameters to realize an even larger class of applications and further solidify the idea of rapid assembly of software defined radios.
Resource Allocation Control Scheme for Secondary Users in Cognitive Radio Networks
Salman A AlQahtani (King Saud University, Saudi Arabia)
Recently, cognitive radio (CR) has been proposed to promote the efficient utilization of the spectrum by exploiting the spectrum holes. The fundamental issue in cognitive radio (CR) networks is the efficient utilization of radio resources. Call admission control (CAC), which controls the number of calls based on the available resources and bandwidth, is an important functionality to ensure the quality of service (QoS). Thus, in this paper, a new call admission with eviction control of secondary users (CAEC) is proposed to improve the spectrum utilization efficiency. This proposed scheme provides better balance between SU QoS and primary users (PUs) protection. Performance evaluation is conducted in terms of the quality of service (QoS) requirements, such as useful utilization and loss probability. Numerical results show that the proposed CAEC achieves higher admitted traffic for SUs and ensure higher system utilization.
Session 6A Spectrum Sensing Architectures
Large Scale Spectrum Sensing Implementation for Cognitive Radio Systems
Raied Caromi (University of Arkansas, USA); Dragoslav Stojadinovic (WINLAB Rutgers University, USA); Ivan Seskar (WINLAB, Rutgers University, USA); Seshadri Mohan (University of Arkansas at Little Rock, USA)
While the problem of spectrum sensing has been extensively studied in recent years, most of the accomplished research lacks practical implementation and consideration for large bandwidth sensing. In this paper, we develop a large scale spectrum sensing scheme for cognitive radio. We first design a framework that is capable of collecting and combining sensing results from multiple nodes. We then implement and compare the performance of three sensing algorithms in standalone nodes equipped with USRP software defined radio. In particular, the three algorithms are: 1) a multiband time-based energy detector, 2) a multiband frequency-domain-based energy detector, and 3) a low complexity sequential multiband spectrum sensing algorithm with delay constraint. Furthermore, we develop a simple, yet efficient procedure for choosing threshold values and number of samples required based on observational statistical averages for the three detectors. Finally, we use the developed framework for sensing large bandwidth using multiple USRP nodes and ORBIT testbed network infrastructure. Spectrum sensing results are provided to demonstrate the efficiency of the developed algorithms and framework for sensing large spectrum bandwidth in a real-time environment.
Efficient Spectrum Sensing/Monitoring Methods and Testbed Development for Cognitive Radio based WSN
Sumit Kumar (IIIT Hyderabad & Center for Communication Research, India); Rama Garimella (IIIT Hyderabad, India)
In this work, we propose efficient spectrum sensing and spectrum monitoring techniques based on Doubly Cognitive Architecture (DCA) and Partial Sensing respectively. We have developed a testbed for Cognitive Radio-based WSN(CRWSN), where the proposed techniques are successfully implemented and tested. The testbed enables real-time implementation of CRWSN in the range of 2.4-2.5GHz with GNU Radio and USRP. The testbed is equipped with significant features such as user access to realize re-configurability, cooperation between CRWSN nodes and ability to emulate several wireless propagation scenarios, especially multipath and hidden nodes. It facilitates real time hardware implementations and in-system software based simulations. The testbed can perform several critical cognitive radio tasks like Dynamic-Spectrum-Access(DSA), Cooperative-Spectrum-Sensing and Multi-frequency-Multi-hop relaying. Our experiments on the testbed show that DCA is more efficient with average Spectrum Discovery Time(3 sec.) and variance(2.66) reduced to 0.42 and 0.19 times the average Normal Spectrum Sensing(7 sec.) and variance(13.83), respectively. Partial sensing-based spectrum monitoring also outperforms periodic spectrum monitoring with average PU detection time(3 sec.) and variance(15.42) significantly reduced to 0.27 and 0.12 times the average of 11 seconds and variance 15.42 of periodic spectrum monitoring. Time-bandwidth-product of partial spectrum monitoring is found to be 50% more than periodic spectrum monitoring leading to higher bandwidth efficiency. We are also developing a toolbox based on this testbed where user can use different prediction methods such as ANN, ARMA, SVM. Our testbed can serve as a development platform providing insights into CRWSN and will be made available to remote users with web-UI.
Weighted Cooperative Sensing based on Spectrum Database for Cognitive Vehicular Networks
Yuya Ohue (The University of Electro-Communications, Japan); Masayuki Kitamura (The University of Electro-Communications, Japan); Kei Inage (The University of Electro-Communications, Japan); Koji Ishibashi (The University of Electro-Communications, Japan); Takeo Fujii (The University of Electro-Communications, Japan)
Cognitive Radio (CR) can adaptively modify own radio parameters based on recognized surrounding radio environment. In the spectrum sharing scenarios, CR users reuse unoccupied spectrum of the Primary User (PU) without giving harmful interference toward PU. It has been studied that CR is applied to Wireless Vehicular Networks (WVNs). In both general CR and CR in WVNs, it is still important to detect PU communication by using spectrum sensing. However, CR in WVNs, has different characteristics compared to general CR. CR nodes in WVNs are located in front of and behind the Master Node (MN) on a roadway. It is difficult to achieve desired gain by using conventional cooperative sensing because the difference of propagation loss and shadowing degrades the sensing performance. To improve sensing reliability in WVNs, we propose a weighted cooperative sensing method based on radio environment database information. Considered database stores long-term average signal power calculated from measured data of CR users. This information is included in each spatial block as geographical data. MN requests the data of average powers at the block locating Cooperative Nodes (CNs) and own position from Database. After that, MN calculates weights of sensing value of all cooperative nodes, and derives the test statistics by weighted average. It becomes possible to give the weights by considering reliability of sensing data from each node. In this paper, we show simulation results and also emulated results from the data of measurement of real TV broadcasting signals to confirm the effectiveness of the proposed technique.
Dynamic Spectral Sensing and Resource Allocation in Optical-Wireless Networks
Egidio Raimundo Neto (Inatel- Instituto Nacional de Telecomunicações, Brazil); Marco Casaroli (Instituto Nacional de Telecomunicações, Brazil); Jeferson Rosa (National Institute of Telecommnunications, Brazil); Igor da Costa (INATEL, Brazil); Antonio M Alberti (National Institute of Telecommunications, Brazil); Arismar Cerqueira Sodré Junior (INATEL, Brazil)
The current Wi-Fi networks are based on access points, which can operate at two ISM bands. As soon as the device is switched on, the frequency band is chosen and kept constant until it is rebooted. Therefore, its system characteristics are fixed and, consequently, possible changes in the random wireless scenario might not be taken into account. This work proposes the concept and reports the implementation of an innovative, adaptive and cognitive optical-wireless metropolitan network. It is able to perform Spectrum Sensing (SS) and Dynamic Resource Allocation (DRA) in order to dynamically adapt the frequency band and/or antenna radiation pattern as a function of the mobile radio environment. The proposed network is based on a Central Office (CO) and Remote Antenna Units (RAUs). The CO concentrates not only SS and DRA functions, but also antenna control. It is based on an efficient integration of a spectrum analyzer, an adaptive and dual-band access point, and a notebook. At RAU side, we have developed reconfigurable dual-band antennas, which are responsible to simultaneously sense the environment over 2.4 and 5.8 GHz bands and to enable full-duplex communication. Its operation can be described as follows. First spectrum occupation is measured and then transmitted to CO by using Radio over Fiber (RoF) technology. The SS data is processed by using an algorithm based on energy detection technique. Finally, DRA is ensured by dynamically choosing the best channel and/or frequency bandwidth in order to optimize the communication link between CO and the network users.
Session 7A CR and DSA Test Beds and Demonstrators
Micro Cognitive Radio Network Testbed (MICRONET) for education, experimentation, and demonstration
Eric Sollenberger (Virginia Tech, USA); Vuk Marojevic (Virginia Tech, USA); Carl B. Dietrich (Virginia Tech & Wireless @ Virginia Tech, USA)
This paper will discuss design and implementation of MICRONET, a testbed that enables realistic, over-the-air software-defined radio (SDR), cognitive radio (CR), and dynamic spectrum access (DSA) experimentation and demonstration over short distances and at low cost. The system can be extended to enable larger scale experimentation at moderate cost, making it well suited to limited-resource educational applications, moderately funded research initiatives, and proof-of-concept demonstrations of experimental systems or operational concepts. Through use of a similar interface, the testbed prepares users for remote use of more capable testbeds such as Virginia Tech's CORNET for large-scale and high-bandwidth applications while providing local users hands-on access to hardware that can be spatially reconfigured. MICRONET's modular design uses PCs with modest performance, low-cost commodity digital video broadcast (DVB) tuners/receive-only SDR RF front ends, and optionally, moderately priced SDR transceivers. Relatively short transmit-receive distances facilitate unlicensed operation under FCC Part 15 rules, and use of laptop computers enables the system to be set up and taken down quickly for use during class time or demonstrations if no dedicated space is available. Use of open-source software allows packaging of system software as well as SDR toolkits and demonstration applications on live DVDs or bootable flash drives for turn-key installation on user-supplied hardware. Details of the system design and implementation are discussed in addition to example DSA and position location experiments. Example experiments and demonstration applications that use the testbed will be described, including position location, diversity and switched-beam antenna systems, and dynamic spectrum access.
Opportunistic Spectrum Access Learning Proof of Concept
Clement Robert (Heriot Watt University, United Kingdom); Christophe Moy (SUPELEC/IETR, France); Honggang Zhang (Université Européenne de Bretagne (UEB) and Supelec & Zhejiang University, France)
UCB (Upper Confidence Bound) is a family of algorithm coming from the machine learning community to answer learning problems that can be modeled as MAB (Multi-Armed Bandit). Theoretical capabilities of UCB algorithms as a means for a cognitive radio equipment to learn about spectrum opportunities have been shown in [1]. It has been explained in [2] that this kind of approach is efficient in a context of high uncertainty, e.g. where the cognitive system has a priori no knowledge about the environment conditions. This paper shows results of a wireless implementation of learning algorithms in lab conditions (USRP + GNU Radio + Simulink). UCB is used by a secondary user to derive the best channel to select in an OSA (Opportunistic Spectrum Access) Scenario build around a couple of USRP N210 platforms. One platform is playing the role of the primary network and generates carriers with a pre-defined probability of occupancy through an OFDM modulation scheme made in GRC environment (GNU Radio Companion). Another platform is playing the role of the secondary user, implemented with Simulink. This proof-of-concept validates in real conditions the UCB capabilities for OSA context. [1] Wassim JOUINI, Damien ERNST, Christophe MOY, Jacques PALICOT, "Upper confidence bound based decision making strategies and dynamic spectrum access", International Conference on Communications, ICC'10, Cape Town, South Africa, 26-29 May 2010. [2] Wassim JOUINI, Christophe MOY, Jacques PALICOT, "Decision making for cognitive radio equipment: analysis of the first 10 years of exploration", EURASIP Journal on Wireless Communications and Networking 2012, 2012:26
Evaluation of Fast Frequency Hopping Modem Using Emulation-Based Test Framework for Cognitive Radios
Jared Thompson (Air Force Institute of Technology, USA); Ethan Hennessey (Air Force Institute of Technology, USA); Kenneth Mark Hopkinson (Air Force Institute of Technology, USA); Mark D. Silvius (Air Force Institute of Technology, USA)
This paper introduces a test framework for evaluating the performance of dynamic spectrum access (DSA) cognitive radios (CR) using an emulated radio environment. The test framework is designed to evaluate the capability of CRs to both avoid causing harmful interference to others while minimizing received interference. First, in developing this framework, the authors review existing testbeds and performance metrics for evaluating DSA systems. Second, the authors present an initial case study, applying this new test framework to the evaluation of an adaptable fast frequency hopping (FFH) modem for a DSA CR. The modem leverages knowledge of the spectrum using a radio environment map (REM). The REM is used by the modem to alter the chip sequence and avoid interference. The REM is created using a simple energy detector. The FFH modem is constructed in a MATLAB model and evaluated over a range of signal-to-noise ratios for a single interferer. The simulation results can be used for a comparison with test results using a hardware emulator as presented with the test framework.
A Flexible and Extensible Cognitive Radio Test System (CRTS)
Junsung Choi (Virginia Tech, USA); Sayantan Guha (Virginia Tech, USA); Ferdinando Romano (Virginia Tech, USA); Tonya Smith-Jackson (North Carolina A&T State University, USA); Carl B. Dietrich (Virginia Tech & Wireless @ Virginia Tech, USA)
The multitude of possible approaches to a given cognitive radio application demands an ability to compare the functionality of different cognitive radio systems under a variety of operational conditions. We will describe a step towards a comprehensive cognitive radio testing capability. The cognitive radio test system (CRTS) can evaluate the performance of multiple cognitive radios under a variety of stressful environments. We will detail our prototype, including communication and link quality feedback between a virtual or SDR-based radio transmitter and receiver, and the ability for the transmitter to switch among multiple predefined operating modes based on the feedback it receives from the radio receiver. Modular implementation using C++ enables efficient implementation of machine learning and optimization techniques, testing of multiple adaptive controllers or cognitive engines via a common interface with the hardware or virtual radios they control, and testing with multiple signal environments that include different combinations of noise and interfering signals, and propagation characteristics. Signal environment characteristics including temporal behavior are parameterized to allow automated generation of multiple operational scenarios of varying complexity to challenge and evaluate cognitive engine performance, while hardware modifications such as introduction of filters, preamplifiers, or attenuators enable assessment of the effects of RF front end hardware on cognitive radio performance. Plans for extending the system from the initial simulations and cabled measurements through over-the-air tests through a series of progressively more realistic but less controllable use cases, including use with indoor and outdoor testbeds and live operation within a wireless network are discussed.
The Wireless Innovation Forum does not endorse or recommend the content of any of papers and presentations contained within these proceedings, rather these items are listed strictly as a service for the community. The information and materials contained in these proceedings are provided "as is" without warranty of any kind, either express or implied, including without limitation, any warranty of accuracy, adequacy or completeness of the information and materials, title, non-infringement of third party rights, merchantability, fitness for a particular purpose, etc.
