SDR'12-WInnComm-Europe Papers and Tutorials
Jump to: Thursday | Friday
Jump to: Tutorial 1.1 | Tutorial 1.2 | Tutorial 1.3 | Session 2.1 | Session 2.2 | Session 2.3
Wednesday, June 27
Tutorial 1.1 (Netherlands I)
A Rapid Graphical Programming Approach to SDR Design & Prototyping with LabVIEW and the Universal Software Radio Peripheral
Filip Langenaken (National Instruments, Belgium)
The Universal Software Radio Peripheral (USRP) has proven a popular hardware platform for host-based software defined radio prototyping with GNURadio. While the approach is popular, the learning curve associated with Linux, GNURadio, Python, and C++ can be challenging for some potential users. To improve ease-of-use and enable the platform for a broader set of users, National Instruments has recently released support for the USRP hardware to Microsoft Windows and the LabVIEW graphical programming environment. This session will show how you can design and prototype communications algorithms with LabVIEW in Windows. Live demonstrations will show how you can define a custom communication protocol using LabVIEW-based graphical programming and establish a live digital link between two USRPs.
Tutorial 1.2 (Belgium I)
Using CREW: a federated platform for experimentally-supported research on cognitive networks and spectrum sensing
Stefan Bouckaert (Ghent University - IBBT, Belgium); Ingrid Moerman (Ghent University - IBBT, Belgium); Sofie Pollin (IMEC / UC Berkeley, USA); Asier Alonso (TECNALIA, Spain)
The aim of the FP7-call5 project CREW (Cognitive Radio Experimentation World) is to create an open federated testbed for the evaluation of cognitive radio and cognitive networking strategies. Experiments can be executed at 5 CREW test sites across Europe, located in Belgium, Germany (x2), Ireland and Slovenia, offering a diverse set of wireless technologies including heterogeneous ISM, heterogeneous licensed cellular, wireless sensor, and heterogeneous outdoor wireless networks. Since the start of the project in October 2010, the baseline testbeds of the testbed owners in the consortium have been extended with advanced sensing hardware developed by other partners in the consortium. Additionally, federation functionality such as a benchmarking framework allowing the comparison of different cognitive solutions, a common data format to store and share the experiments and experimental results, and mix-and-match interfaces that can be used to combine cognitive components from different partners were developed. Finally, a portal website was created (www.crew-project.eu/portal), which offers a single point of entry to experimenters, and provides information on the different facilities in the CREW testbed and how to access them. Early 2012, three additional partners joined the CREW project for the duration of one year after being selected as part of an open call for experimenters, to execute their experiments on top of the CREW testbed offered by the eight core partners of the consortium. While the former partners are using the current CREW testbed for the execution of their experiment, the other partners plan to improve the functionality offered by the federation in the coming months. =Tutorial program= In the first part of the tutorial, an overview of the federated CREW facility will be given, and the goal and use of selected CREW federation functionalities such as the common data format or the benchmarking framework will be explained. [ Presenter: Ingrid Moerman, IBBT. Duration: 30' ] The second and most important part will illustrate how the CREW facility can be used to evaluate cognitive radio/networking solutions, by presenting three concrete use cases that have been investigated using the CREW federation: -The CREW federation houses a broad range of spectrum sensing equipment, from off-the-shelf sensor nodes such as WiSpy devices, to specialised sensing engines that are custom-built by partners of the consortium. We demonstrate the technical background behind, the advantages of, and the experimentation possibilities enabled by integrating advanced sensing components in a cognitive networking testbed. [ Presenter: Sofie Pollin, imec. Duration: 20' ] -In past experiments performed by the CREW consortium, a common data format and benchmarking methodologies have been used to compare the different sensing devices that are present in the CREW consortium. Through this example, it will be explained to the audience how the benchmarking and common data aspects of the CREW federation contribute to making fair comparison of different cognitive solutions a possibility. Additionally, the results of the sensing device comparison experiment are briefly presented. [ Presenter: Stefan Bouckaert, IBBT. Duration: 20' ] -Testimonial from an open call partner. Tecnalia is one of the three partners that was accepted as part of the first open call for experimenters. In this talk, it will be discussed how CREW will be used to carry out an experiment focussing on the assessment of benefits of optimized linear collaborative multiband spectrum sensing in cognitive radio networks. [ Presenter: Asier Alonso Muñoz, TECNALIA-Telecom. Duration: 20' ] Before finishing with a Q&A, in a final part of the tutorial, we will show how the audience can learn more about the CREW federation and/or access the federated testbed, by demonstrating the use of the federated portal. Furthermore, we will share preliminary information on the second open call, which allows selected experimenters to become part of the CREW project and to get funded for executing their cognitive radio and cognitive networking experiments. [ Presenter: Ingrid Moerman, IBBT. Duration: 30' ] For the audience of the conference, this tutorial provides the opportunity to learn about the CREW project and about the existing possibilities of experimental research related to cognitive radio and cognitive networks in general. For the members of the CREW consortium, it is an interesting opportunity to meet the cognitive radio and cognitive networking experimenter community, whose comments and questions will help to shape the future of the CREW federation in a demand-driven way. =The presenters= Ingrid Moerman is a professor at Ghent University since October 2000. In 2006 she joined IBBT, where she is coordinating several interdisciplinary research projects, including the FP7-CREW project. She is currently leading the research and education on mobile & wireless communication networks within the 'Internet Based Communication Networks and services (IBCN)' research group, one of the research groups of IBBT. Her team has about 25 researchers. Stefan Bouckaert received his M.Sc. and Ph.D. degrees in electro-technical engineering from Ghent University, Belgium, in 2005 and 2010, respectively. He is currently a post-doctoral researcher at IBBT-IBCN, where he is mainly involved in the CREW project, working on the topics of experimentally-supported research and benchmarking. In January 2012, he became also responsible for the business development of the IBBT iLab.t research centre (http://ilabt.ibbt.be). Sofie Pollin received a degree in electrical engineering in 2002 and a Ph.D. degree in 2006 (with honors) from the Katholieke Universiteit Leuven, Belgium. Since October 2002 she has been a researcher at the Wireless Research group of imec. In the summer of 2004 she was a visiting scholar at National Semiconductor, Santa Clara, California. In the summer of 2005 she was a visitor at UC Berkeley. In 2006, she started as a post-doctoral researcher at UC Berkeley working on coexistence issues in wireless communication networks. Currently, she is a principal scientist at imec working on cognitive radio and software defined radio. Asier Alonso received a degree in electrical engineering from the University of Deusto in 2004, and also holds a master in computer architecture (University of Deusto 2008). Since 2005 he works as a full time researcher at the Telecom Unit of Tecnalia (before Robotiker), in the fields of Software Defined Radio, programmable logic, and signal processing.
Tutorial 1.3 (Germany)
Developing GNU Radio Signal Processing Blocks
André Selva (Universidade Estadual de Campinas, Brazil); André L. G. Reis (Unicamp - State University of Campinas, Brazil); Karlo Lenzi (Universidade Estadual de Campinas, Brazil); Luís Meloni (Universidade Estadual de Campinas, Brazil); Silvio E. Barbin (University of Sao Paulo, Brazil)
GNU Radio is a popular toolkit for the development of SDR applications. Despite offering a large library of signal processing blocks, one may find useful to develop new blocks that fits exactly with the application requirements. Furthermore, complex applications, like Digital TV transceivers or LTE networks, for example, demand more complex digital signal processing blocks than the ones available in the GNU standard library. Thus, it is essential to know how to build custom blocks on GNU Radio if we desire to make full use of the GNU Radio features and Software Defined Radio platforms. This works aims to present an in-depth study of how to create new blocks in GNU Radio, clarifying dubious points generated from the lack of documentation available on the subject, which is based mainly on forums discussions and tutorials, and to serve as a future reference for new developers. To best understand this process and to cover all stages of development involved in creating blocks, we will present an step-by-step of how to build a byte interleaver, commonly used in wireless communication systems, such as WiMAX, LTE and DTV.
Session 2.1 - Spectrum Sensing
Chair: Vincent J Kovarik, Jr (Prismtech, USA)
16:15 - 18:15 (Netherlands I)
Spectrum Sensing in the Vehicular Environment: An Overview of the Requirements *Top Paper*
Haris Kremo - Presenting (Toyota InfoTechnology Center, Japan); Rama K Vuyyuru (Toyota Info Technology Center, USA); Onur Altintas (Toyota InfoTechnology Center, Japan)
This paper overviews the challenges related to spectrum sensing in the vehicular environment, with emphasis on sensing in the TV licensed band. In the vehicular environment the cognitive radio can help to: 1) satisfy capacity demand for Intelligent Transportation Systems (ITS) applications; and 2) offload time insensitive applications from the ITS dedicated spectrum. However, neither sensing, nor geolocation database lookup alone can provide sufficient incumbent protection. Collaboration among the sensors to take advantage of spatial diversity is difficult due to the rapidly changing network topology. Nevertheless, mobility provides the opportunity to use time diversity at each sensor. We also discuss the influence of sensing subsystem design on the vehicular cognitive network medium access (MAC) sublayer. Whenever applicable, we compare sensing requirements for vehicular cognitive networks to the requirements provided in the IEEE 802.22 standard.
On the Performance Assessment of Heuristically-driven Linear Collaborative Spectrum Sensing within the FP7-ICT CREW Project (Presentation Only)
Asier Alonso (TECNALIA, Spain); Javier Del Ser (TECNALIA, Spain); Sergio Gil-Lopez (TECNALIA, Spain)
Spectrum sensing and decision are essential tasks of any cognitive radio system. As such, many different schemes have been devised so far in order to perform these task in the most accurate and reliable way, which can be a priori classified as: 1) conventional non-collaborative spectrum sensing, where a decision is taken at every cognitive node of the network in isolation; 2) collaborative spectrum sensing, where the spectral measurements registered by different nodes are combined - either in a centralized or distributed fashion - so as to produce a decision with higher reliability than the case where the decision is taken based on a single measurement (see rightmost plot in Figure 1). As for the latter, a number of collaborative spectrum sensing techniques pre-process the spectral measurements of every compounding sensing node of the network so as to produce a binary (hard) local decision on their occupancy, followed by a hard-decision fusion approach (e.g. OR, AND) that generates the final spectral occupancy metric. Technical advantages of these hard-decision based techniques lie on their simplicity for their implementation in conventional digital hardware. However, as was stated in the theory of evidence developed by Arthur P. Dempster and Glenn Shafer , binary local decisions can be easily outperformed, in terms of credibility and reliability when applied to outcomes of a same event, by soft fusion techniques where the unprocessed outcomes of the said event are input to a unique soft test. This interesting result motivates the upsurge of soft-decision based fusion techniques applied to spectrum sensing for cognitive radios. Among the broad portfolio of such combining approaches, in this work we have concentrated on the so-called linear statistics combination, which will be hereafter denoted as LSC . LSC hinges on linearly combining the unprocessed spectrum measurements captured by cognitive radio nodes by means of a set of configurable coefficients, based on whose result a decision is taken. When the LSC is applied at each compounding band of a broad bandwidth, the resulting scheme is rather coined as multiband LSC. The values of such coefficients are usually set equal to each other in the conventional implementation of the LSC approach. However, as explained by the authors in , collaborative multiband LSC spectrum sensing can be optimized in terms of the achievable throughput over such a band by making use of evolutionary optimization strategies. Taking into account this state of art, this work gravitates on assessing the benefits of optimized linear collaborative multiband spectrum sensing in cognitive radio networks with respect to its non-optimized counterpart. Four specific goals stem further from this general objective: To incorporate the LSC decision scheme as well as its heuristic optimization framework recently introduced in  in a real testbed. To benchmark hard-decided collaborative spectrum sensing techniques against genetically-optimized and nonoptimized LSC approaches. To quantize the performance gains entailed by the application of Harmony Search (HS, see ) heuristics to the optimization of the LSC coefficients under a maximumthroughput criterion. To investigate and implement the data format, mechanisms and access procedures through which the sensed information is stored in a shared database and thus made network-wide available. From a general perspective, our technical contribution is twofold: on the one hand, HS heuristics are applied for the first time to the optimization of the linear coefficients involved in collaborative spectrum sensing. As such, this algorithm inspires from the improvisation process of musicians, i.e. the process by which such musicians (who may have never played together before) refine - through variation and check - their individually improvised notes resulting in an aesthetic harmony played by the entirety of musicians in the orchestra. Due to its outperforming behavior over genetic approaches and its simplicity, HS has been so far applied to a number of applications and problems related to wireless communication networks, such as the Capacitated P-median Problem , multicast routing , multiuser detection , , distributed radio resource allocation  and OFDMA subcarrier power allocation . To the knowledge of the authors, this is the first contribution in the literature dealing with HS for the optimization of collaborative spectrum sensing schemes, which is expected to outperform its genetic counterpart in terms of aggregate throughput and/or speed of convergence. On the other hand, heuristically-optimized collaborative spectrum sensing is put into practice through its implementation on the test beds compounding the open federated experimentation platform provided by the CREW (Cognitive Radio Experimentation World) project , funded by the EC under the 7th Framework Programme. Specifically, a multiband energy level sensing functionality capable of capturing, formatting and forwarding spectrum measurements to a database is developed so as to allow the fusion center to access all the sensed spectrum information, hence modelling the conventional allocation of exclusive radio resources devoted to the circulation of cognitive control information (leftmost diagram in Figure 1). We will outline the results and conclusions regarding the performance of the previously mentioned heuristic algorithms obtained in the various tests carried out within the CREW real testbed, along with a comparison study with hard-decided AND and OR combination strategies. see attached extended abstract for more detail
Performance Evaluation of a Spectrum-Sensing technique for LDACS and JTIDS Coexistence in L-Band *Top Paper*
Giulio Bartoli (University of Florence, Italy); Romano Fantacci (University of Florence, Italy); Dania Marabissi (Università di Firenze, Italy); Luigia Micciullo (University of Florence, Italy); Claudio Armani (SELEX Elsag, Italy); Roberto Merlo (SelexElsag spa, Italy)
This paper deals with a cognitive approach able to guarantee the coexistence of new data link for air-ground aeronautical communications LDACS and military JTIDS systems. Future LDACS shall coexist with current systems operating in the same frequency band for this reason coexistence issues must be carefully investigated. In particular JTIDS transmissions can affect the LDACS performance acting as disruptive impulse noise. JTIDS exploits frequency hopping to protect information, hence its interference on LDACS system cannot be foreseen and avoided. In addition the bandwidth of the two signals results to be completely overlapped in case of collision. The disruptive effects of JTIDS interference on LDACS can be mitigated if the collisions can be detected and hence suitable processing techniques can be activated. This paper proposes a method to detect the presence of JTIDS interference exploiting an energy detection spectrum sensing technique based on sliding windows. The performance of the proposed method is presented in terms of missed detection probability of the JTIDS interference and error rate of the LDACS system showing a good behavior.
A Set of Methodologies for Heterogeneous Spectrum Sensing (Presentation Only)
Wei Liu (IBBT, Belgium); Sofie Pollin (IMEC / UC Berkeley, USA); Peter Van Wesemael (IMEC, Belgium); Danny Finn (Dublin University, Trinity College & CTVR Telecomunications Research Centre, Ireland); Christoph Heller (EADS Innovation Works, Germany); Mikolaj Chwalisz (Technische Universität Berlin, Germany); Daniel Willkomm (Technische Universitaet Berlin, Germany); Nicola Michailow (Technische Universität Dresden, Germany); Zoltan Padrah (Jozef Stefan Institute, Slovenia); Ingrid Moerman (Ghent University - IBBT, Belgium); Stefan Bouckaert (Ghent University - IBBT, Belgium)
Cognitive radio has received tremendous amount of attention in the academic world. As a key enabler for cognitive radio, the spectrum sensing technology faces many challenges. The real-life wireless communication often happens in a heterogeneous environment, hence, how to combine sensing results obtained with heterogeneous devices is one of the biggest challenges. In order to have a valid heterogeneous sensing system, many issues need to be , such as the calibration among heterogeneous devices, post processing for the obtained data in different formats, what is the most efficient way for combining data, and many other aspects. In this work we present a set of methodologies that have been derived in the scope of the CREW project to deal with some common issues encountered for heterogeneous sensing. The FP7 project CREW (www.crew-project.eu) targets the development of a federated testbed for cognitive radio systems by physically and virtually interconnecting radio equipment of the individual project partners. By combining the spectrum sensing devices from each project partner, we are able to form a sensing platform with many popular heterogeneous devices. These devices include dedicated integrated sensing hardware (imec sensing engine), USRP software-defined radios (SDRs), small, low power sensor nodes (TelosB), off-the-shelf, low cost USB spectrum analyzers (WiSpy) as well as high cost, high precision spectrum analyzers. The presentation will start with a series of heterogeneous and distributed sensing experiments that we performed as background information. The aim of those experiments was originally to compare the performance of different sensing devices. However, during those experiments we have gradually formed a system to properly process distributed data and calibrate heterogeneous devices. The derived methodologies are covered next. The main focuses here are the measurement we performed to compensate for the hardware heterogeneity, and the common data format we derived to overcome the software difference.Finally we also propose some principles we adopted to detect outliers in the heterogeneous distributed sensing experiment. This work is not about analyzing results from any particular heterogeneous sensing experiment, but rather to learn from all the experiments from a methodology point of view. It can serve as a reference for future work in heterogeneous sensing.
Session 2.2 - SCA Implementations
16:15 - 18:15 (Germany)
A component-based architecture for protocol design and development in SDR frameworks (Presentation Only)
Maurizio Colizza (University of L'Aquila, DEWS, Italy); Marco Faccio (University of L'Aquila, Italy); Claudia Rinaldi (University of L'Aquila, Italy); Fortunato Santucci (University of l'Aquila, Italy)
The increasing interest in software defined radio (SDR) as enabling technology for defining and developing advanced wireless systems, e.g. mobile ad-hoc networks (MANET) with high degree of adaptivity and ability to (re)configure in application scenarios, motivates research efforts in developing methods and tools for supporting a complete and sound design flow, that encompasses i) waveform/protocol specification, ii) thorough validation through accurate simulations and early stage testing, and then iii) rapid code development on selected target platform. Despite those needs, it can be observed that state-of-the-art approaches still lack significantly in several components, with critical drawbacks in those environments where performance estimation and assessment is particularly challenging (MANETs are still an example). A (non exhaustive) list of weaknesses can be provided as follows, 1. simulators are not typically conceived to offer the opportunity to reuse the developed code for subsequent implementation on target device; 2. merging of measurement code and business code is not typically addressed; 3. tools for sound and easy support of tracking projects requirements into the developed code are not available; 4. tools for supporting automatic generation of reports that rely on qualitative and quantitative performance assessment are not satisfactory; 5. high level cross verification for performance analysis (e.g. logic trigger) is not addressed. With the motivation of progressing in the depicted technical framework, our research group is involved in several projects, e.g. ARTEMIS PRESTO and FP7 NoE HYCON2 both co-funded by the EC: they have quite a broad spectrum of activities, but they both try to overcome limitations in methodologies that relate to our context. Specifically, the PRESTO project is mainly targeted: 1. to improve test-based embedded software development and validation, while considering the constraints of industrial development processes; 2. to establish functional and performance analysis with platform optimisation at early stage of the development. The project also intends to explicitly consider some industrial development constraints: simplified use of tools, smooth integration in current design processes, framework of tools that is flexible enough to adapt to different process methodologies, design languages and integration test, platform modelling for early comparison of results with real scenarios and fast prototyping. Through the WP6, HYCON2 also pursues research advances in developing methods and tools for analysis and design in the broad range of complex and networked embedded systems. After providing an appropriate description of state-of-the-art, the present paper is intended to report on our research activity, that is focused on defining and developing a set of tools that support the sound design, appropriate verification/test and development of embedded software for SDR systems. Specifically, we are defining a workflow whose qualifying features are as follows: 1. the design of a system or subsystem in a network/protocol stack is model-based; 2. the amount of manually written code (firstly for simulation) is minimized, while the code is usually obtained from the model through a set of procedure for automatic code generation; 3. the probes for measurement may be placed in the model; them may be automatically switched off when the model is used to produce code for target device; the model holds true independently of the target device; 4. when a probe for measure is selected, the generated track can be automatically added to a technical report; The suite is intended to provide the designer with the abilities: 1. to create a protocol model through the composition of library component; 2. to generate code for Simulation and Test in a network simulator starting from the model; 3. to generate code for a target device, starting from the model; 4. to integrate protocol models with application related models, e.g. those encountered in the the context of networked control systems. The paper will report on already achieved results in terms of developed models and simulation environments.
Prototyping SCA API's Using a Generic Reasoner API (Presentation Only)
Durga Suresh (Northeastern University, USA); Mieczyslaw Kokar (Northeastern University, USA); Jakub Moskal (VIStology, Inc., USA)
API's are abundant in the realm of the SDR and there are many different API's developed for different protocol layers, each with a specific purpose and particular hardware and software needs. Those API's are then implemented within a common SCA architecture, leading to a great advantage of interoperability among various radios and portability to other platforms. The standard practice of developing an API for an SDR is by first describing it in UML. While UML tools provide some methods for syntactically constraining the development of a specification of a system, they don't support the capability of verifying or enforcing the semantic constraints. Consequently, the semantic interpretation of the constraints imposed by an API is done by humans. This paper discusses the potential uses of languages with formal semantics (e.g., OWL) in the development of the SDR API's. In particular, it investigates the use of the concepts from the cognitive radio ontology (CRO) to express an API and then using a reasoner to analyze the specification, e.g., checking its logical consistency, querying the specifications of the API's and even testing their behaviors using a generic implementation of an API for the formal language (OWL API), i.e., prototyping an API. This prototyping approach is considered with respect to such metrics as the amount of effort of the prototype implementation, efficiency of the prototype API, the impact on standardization and other possible uses of such a mapping.
Component-based Approach to Decomposing SCA Components
Toby McClean (PrismTech, Canada)
The Software Communications Architecture provides a component-based framework for developing software defined radio waveforms, platforms and services. A component (resource, device or service) in the SCA is a relatively coarse grained software entity which is encapsulated and may be independently deployed. One of the objectives of the SCA-Next initiative is the capability to design and implement SCA components as a hierarchical assembly of smaller fine-grained components. In this presentation we show a methodology and tooling that uses a standardized software component specification language to provide this capability within the constructs of the SCA component model. By applying this methodology an SCA component is designed and implemented as the assembly of software components. Each of the software components can be further refined as an assembly of software components. Within an SCA component each of these components is an independent block of functionality. This independence gives the designer many options including the ability to parallelize parts of a SCA component. State of the art code generation techniques are then used to generate and implement SCA compliant components that can be deployed as part of a SCA waveform or logical platform. Benefits * Increased reusability due to the finer granularity of functionality; * Increased productivity as platform specific and SCA specific code is generated; * Better separation between structure and behavior in the design and implementation of a SCA component; and * Ability to declaratively allocate pieces of functionality within a SCA component to threads.
Portable Software Communications Architecture (SCA) Waveform Design for FPGAs (Presentation Only)
Andrew Foster (PrismTech Limited, United Kingdom)
Field Programmable Gate Arrays (FPGAs) are a key enabling technology in Software Defined Radio (SDR) development. They have been typically used to implement physical layer processing functions such as IF (Intermediate Frequency) up/down conversion and crypto functions. However, the latest generation of FPGAs are increasingly being used to support processing tasks normally associated with Digital Signal Processors (DSPs) and General Purpose Processors (GPPs). Traditionally in radio design FPGAs have been considered part of the modem hardware with the consequence that frameworks such as the Software Communications Architecture (SCA) that deal with software level control functions have until recently not adequately addressed the integration and portability issues that SDR developers must deal with in order to use FPGAs effectively within their SDRs. This presentation will examine various approaches for developing portable SCA waveform applications using FPGAs. It will include a discussion on the MHAL (Modem Hardware Abstraction Layer) / MOCB (MHAL On Chip Bus) standards as recommended by SCA 2.2.2 and contrast them with the latest FPGA interface standards proposed in the new SCA Next specification. Finally the presentation will examine the benefits of implementing SCA waveform components directly in hardware on the FPGA using the latest FPGA tools and middleware designed to support rapid and portable development in an SCA Next compliant manner.
Session 2.3 - SDR Implementations 1
Chair: Harald Kröll (ETH Zurich, Switzerland)
16:15 - 18:15 (Belgium I)
Software Radio Spectrum Analyzer
Jérôme Parisot (SUPELEC, France); Emilien Le Sur (SUPELEC, France); Christophe Moy - Presenting (SUPELEC/IETR, France); Daniel Le Guennec (IETR/Supélec-Campus de Rennes, France); Pierre Leray (IETR/Supelec Campus de Rennes, France)
We propose in this paper to describe how a spectrum analyzer has been implemented using USRP platforms with MathWorks Simulink development environment. We call it a software radio spectrum analyzer. Moreover, we plan to include smart signal processing including automatic detection of different characteristics, such as standard recognition, inter-channel interference, etc. We show then that software radio is an enabling technology to improve usual equipments (here lab equipment) and create a new generation of cognitive equipments (measurement equipments here). This work is in particular done by last year students before graduating for engineering diploma. The use of USRP platforms makes students face real signals and apply signal processing theory with the constraints of reality. The results will be shared with the community as it is done for previous work shown at last Wireless Innovation Forum SDR'11 conference .  Adrien LE NAOUR, Olivier GOUBET, Christophe MOY, Pierre LERAY "Spread Spectrum Channel Sounder Implementation with USRP Platform" Wirelless Innovation Forum Conference, SDR'11, Washington DC, USA, 29 Nov-2 Dec. 2011
A Software Defined Radio Approach for Digital Television ISDB-T Transmitters
André L. G. Reis (Unicamp - State University of Campinas, Brazil); André Selva (Universidade Estadual de Campinas, Brazil); Karlo Lenzi (Universidade Estadual de Campinas, Brazil); Luís Meloni (Universidade Estadual de Campinas, Brazil); Silvio E. Barbin (University of Sao Paulo, Brazil)
Although most countries already have Digital Television (DTV) available, there are several others migrating their infra-structure from analog to digital. One of these countries is Brazil, which intends to accomplish a full switch up to 2016. This initiative is seen as of extreme importance by local authorities since Brazil will host the FIFA World Cup in 2014 and the Olympics Games in 2016, events that will be seen by millions of viewers around the globe. The Brazilian DTV is based on the Japanese ISDB-T standard for the modulation scheme, which is an OFDM system capable of supporting services from mobile to full-definition. Since the relevance of the subject, this paper presents a software defined-radio approach to implement a full ISDB-T transmitter using GNU Radio and USRP, covering a complete study from modeling to implementation. This papers aims not only to discuss and clarify several misinterpretations of the ISDB-T standard, but also to present an efficient implementation on a SDR architecture, serving as reference to new developers and as a roadmap for manufactures.
Vehicle Power Line Communication (VPLC) implementation with USRP2 Platforms
Fabienne Nouvel (INSA, France)
This paper deals with the implementation of an embedded power line communication system for vehicle (VPLC) using USRP platforms and GNU Radio environment. This platform allows a modular design of VPLC. Many configurations of the PHY and MAC layers of can be tested in a real PLC environment, without developing them on a specific board. In conjunction with USRPs platform and daughter cards, the signal processing performed in software is transmitted over a real channel. Several frequency bands can be tested by using different daughter cards without changing the signal processing. Today the car manufacturers have to face with an increase of electronic nodes (ECU) connected each other. The ECUs already use networks like CAN and Flexray, but the number of specific wired always increase. One solution to reduce the amount of wires would be to use the PLC technology that is currently being developed for domestic networks to transmit information or at least some of it, over the 12V power distribution system found in cars. In our system, SDR can be considered to be a "wired" communication system, where some of its functional components, such as modulations, coding, synchronisation, etc., are implemented with software components. This makes it possible to configure the signal according to the requirements of the application and the characteristics of the communication channel. The software generated signal is then applied to the USRP platforms that are connected to a host computer through USB or Ethernet connections.
IEEE 802.15.4 transceiver for the 868/915 MHz band using Software Defined Radio
Rafik Zitouni - Presenting (ECE de Paris & Université de Paris Est, France); Stefan Ataman (ECE, France); Marie Mathian (ECE, France); Laurent George (Ece Paris, France)
In this paper, we present an implementation of the PHY specifications of IEEE 802.15.4 standard for the frequency band 868/915MHz using Software Defined Radio (SDR). This standard is defined for low power, low-data rate and low cost wireless networks. The specifications are used by Zigbee technology for various applications such as home automation, industry monitoring or medical surveillance. Several hardware PHY 868/915 MHz band IEEE 802.15.4 transceivers implementation have been already reported on ASIC and FPGA  . SDR is one possibility to realize the transceiver with a high flexibility and reconfigurablity . The whole emitter and receiver chain have been defined in software using the GNU Radio and the USRP (Universal Software Radio Peripheral) platform from Ettus Research. Two new blocks have been added to the GNU Radio project, one for the Direct Sequence Spread Spectrum and the second for the reconstruction of the packets. The experimentations have been performed in a noisy environment and the PER, BER and SNR have been computed. The obtained results closely follow the well-known theoretical limits. Overview of IEEE 802.15.4: IEEE 802.15.4 standard specifications have been adopted for low-power, low-cost sensor networks with high reliability. Wireless sensor networks use these specifications to define the physical and a MAC sub layer. For the physical layer, there are 3 principal frequency bands with 49 channels, 16 channels in 2450 MHz for the ISM Industrial Scientific Medical band (operated at raw data rate of 250 kb/s), 30 for North America and 3 channels in the 868 MHz band . The 915MHz and 868MHz bands are used at 40 kb/s and 20 kb/s, respectively. The specification of 2011  allow us to use different modulation techniques and data rate for the specified channels. The BPSK is one of the modulation techniques used in the 915/868 MHz band. These frequencies are interesting as they allow longer range communications, compared to the 2450 MHz band. GNU Radio and USRP: For the implementation of our transceiver we used a SDR composed of the (Universal Software Radio Peripheral) USRP1, controlled by the GNU Radio Software. The USRP1 is a low cost hardware platform for SDR communications. It is mainly composed of a USB 2.0 Cypress FX 2 interface, 4 Analogical/Digital Converter ADC/DAC connected to the Front End via a daughter board, and an Altera Cyclone EP1C12 FPGA that connects all these components. The signal processing is accomplished by the CPU of the computer executing the GNU Radio software. The GNU Radio is an open source platform that allows us to create the flow graphs to process the stream transmitted and/or received by the USRP. Each flow graph is defined by a number of blocks developed in C++ language and connected by the Python script. Overview of the implementation: Our software transceiver is based on the IEEE 802.15.4 specifications for the 868/915 MHz band. Our implementation is similar to the one in the 2450MHz band, described in . The emitter is firstly composed of the packet generator. The size of the packets constructed and transmitted by the emitter is 133 bytes. The packets are modulated with BPSK and transformed into chips by the blocks we have developed. Each bit is represented by a 15-chip Pseudo Noise sequence. At the receiver side, the non-coherent receiver decodes the stream by doing the opposite operations. After the BPSK demodulation, the last block built has to transform the chips to bits and after that to reconstruct the packets. Results: The BER and SNR are calculated by changing the signal amplitude. Thus, the Packet Error Rate can be calculated. The experimentations are performed in indoor environment with unpredictable noise sources. In a first step, a continuous bit stream of 1s is sent by the modulator. The purpose is to easily compute at the receiver side the BER in a window of 1000 bits. The SNR is calculated by the proposed block in the GNU Radio. In the second step, a packet generator block is added, thus the computation of PER is accomplished by using the CRC-16 without correction. The values of BER and SNR are closely following the well-known theoretical limits. The emitter and the receiver are implemented and run on two different computers for the computation of the PER. The fraction of packets correctly received is up to 0.8 (80%). Several problems are encountered and impact the performances of the software transceiver. The computer has to be powerful enough to transform the received signal into a continuous and uninterrupted stream. The carrier of the USRP daughter-board is inaccessible and imposes that the receiver is non-coherent. Furthermore, the USRP buffer overflows when decoding the data in real time. Thus, the buffer keeps the old data received in the last experiment and makes it hard for the receiver to be synchronized through packet reception. Conclusion and further work: This paper reports the SDR implementation of IEEE 802.15.4 standard specifications in the 915/868 MHz band and presents the experimental results obtained with it. Our development and the obtained results prove the flexibility of this platform and its practical feasibility. As a further work, we intend to build a software multimode 2450MHz and 915/868 MHz transceiver. References:  Sabater, J.; Gomez, J. M. & Lopez, M. , "Towards an IEEE 802.15.4 SDR transceiver", in 'ICECS' , IEEE, , pp. 323-326 . 2010.  Nam-Jin Oh, Jinho Ko, and Sang-Gug Lee, "A CMOS 868/915-MHz Direct Conversion ZigBee Single-chip Radio", IEEE Communications Magazine, Vol. 43, No. 12, pp. 100-109, Dec. 2005.  Ulversoy, T. "Software Defined Radio Challenges and Opportunities ", IEEE Communication Surveys & Tutorials, Vol 12, Issue 4. pp. 531-550, Nov 2010.  802.15.4-2011 IEEE Standard for Local and metropolitan area networks--Part 15.4 Low-Rate Wireless Personal Area Networks (LR-WPANs). 2011.  T. Schmid, "GNU radio 802.15.4 en- and decoding," UCLA NESL Technical Report, Sept. 2006.
Fixed-Point Aspects of MIMO OFDM Detection on SDR platforms *Top Paper*
Alongside with the increasing importance of mobile, wireless communication, the amount of standards in this field is increasing constantly. Therefore, more flexible platforms, implementing the different standards in software, hence called Software Defined Radios (SDR), are needed. Since state-of-the-art wireless standards like LTE and IEEE 802.11n make use of multiple-input multiple-output (MIMO) antenna configurations and orthogonal frequency-division multiplexing (OFDM) modulation, it is highly relevant to investigate a flexible MIMO OFDM application on an SDR platform. In LTE as well as IEEE 802.11n compliant receivers, MIMO detection causes a major part of the computational complexity. Hence, the efficient implementation of MIMO detectors on SDR platforms is of vital importance. In this paper, aspects of MIMO detection specific to SDR are discussed, using the IEEE 802.11n standard as a reference. Suitable algorithms for QR decomposition of the channel matrices for real time processing are presented, along with consecutive detection as well as SINR calculation. Due to the fixed bitwidths of available datatypes, numerical stability is one major issue when developing an SDR application, especially when targeting high antenna configurations (e.g. 4 × 4). This paper emphasizes how the algorithms have to be adapted to ensure algorithmic performance with limited 16 bit precision. As a proof of concept, an implementation of the presented algorithms on the maturing P2012 platform, an embedded many-core platform with the option for single instruction multiple data (SIMD) processor extensions by ST Microelectronics, is presented. Timing performance is benchmarked to proof the real time applicability. To demonstrate numerical stability, error correction performance is measured and compared to a floating point reference.