Cognitive Radio Enabling Opportunistic Spectrum Access (OSA): Challenges and Modelling Approaches

Transcription

1 Cognitive Radio Enabling Opportunistic Spectrum Access (OSA): Challenges and Modelling Approaches Xavier Gelabert Grupo de Comunicaciones Móviles (GCM) Instituto de Telecomunicaciones y Aplicaciones Multimedia (iteam) Universitat Politècnica de València (UPV) <xavier.gelabert@ieee.org> 1

2 Outline Introduction Cognitive Radio Dynamic Spectrum Access Fixed Spectrum Access DSA Models DSA Opportunity Space DSA Architecture Modelling Opportunistic Spectrum Access Basic Concepts Opportunity Detection Opportunity Allocation Opportunity Use Concluding Remarks 2

3 INTRODUCTION 3

4 Cognitive Radio Mitola s View [Mitola1999]: The point in which wireless PDAs and the related networks are sufficiently computationally intelligent about radio resources and related computer to computer communications to detect user communications needs as a function of use context, and to provide radio resources and wireless services most appropriate to those needs. Wikipedia [ CR is a paradigm for wireless communication in which a network/wireless node changes its transmission or reception parameters to communicate efficiently avoiding interference with licensed or unlicensed users ( ) Other definitions [Mitola1999] J. Mitola, G. Q. Maguire, "Cognitive radio: making software radios more personal," Personal Communications, IEEE, vol. 6, no. 4, pp , August

5 Cognitive Radio What is a Cognitive Radio? [Sayaweera2010] Definition not yet agreed upon by anyone: A device? A Cognitive Radio A methodology? Cognitive Radio Networks Different communities seem to have their own definition: Hardware/RF community: CR is essentially an extension of SDR (Software Defined Radio). Antenna community: CR is smart/intelligent antennas & adaptive beam forming. PHY MAC Layer communication researchers: CR synonymous with dynamic spectrum access and sharing (DSA and DSS) Computer and IT: CR is a device/system with machine learning capabilities. Networking and protocol designers: CR is a device/system that performs cross layer optimizations. Information theorists: CR is channel with side information. Actually, CR should be a combination of all this and more [Sayaweera2010] A short course on DSS CR, June

6 Cognitive Radio Solely in the MAC/Networking community there are many confusing terms [Zhao2007] Spectrum pooling Dynamic spectrum access Dynamic spectrum allocation Spectrum Sharing Spectrum commons Spectrum property rights Cognitive Radio Opportunistic spectrum access Spectrum overlay Spectrum underlay [Zhao2007] Q. Zhao, B.M. Sadler, A Survey of Dynamic Spectrum Access: Signal Processing, Networking, and Regulatory Policy, IEEE Signal Processing Magazine, May,

7 Cognitive Radio Our Motivation behind CR Commonly accepted that the wireless spectrum is a scarce resource in wireless communications. The static assignment of the spectrum results in inefficient usage of the spectrum. The utilization of the spectrum can be as low as 10%. Dynamic access to the spectrum (DSA) can be considered as a method for solving spectrum scarcity. Spectrum Opportunities 7

8 Cognitive Radio CR Capabilities and Functionalities Spectrum Awareness (SAw): Quickly and robustly detect the presence of incumbent (preemptive) users to avoid causing interference. Dynamic Spectrum Access (DSA): Based on SAw, CRs will access the spectrum on an opportunistic basis. Spectrum opportunities in time, frequency, space Dynamic Spectrum Sharing (DSS): CRs must be aware of other CRs, coexistence. Spectrum Agile (SAg): CRs should provide seamless operation over multiple channels. SAw DSA 8

9 DYNAMIC SPECTRUM ACCESS 9

10 Fixed Spectrum Assignment Fixed Spectrum Assignment (FSA) limitations: Spectrum licensee cannot be changed. Service type within licensed spectrum cannot be changed. Spectrum license is location invariant. Granularity of spectrum usage is fixed. Spectrum is protected from use by unlicensed users. Fixed allocation No sharing Strict Use 10

11 Dynamic Spectrum Access (DSA) CR enables the usage of temporally unused spectrum: Dynamic Spectrum Access (DSA) If this band is further used by a licensed user, CR moves to another spectrum hole or stays in the same band Alters its transmission power level or modulation scheme to avoid interference. [Akyildiz2008] [Akyildiz2008] Cognitive Radio Networks Course, Barcelona June

12 DSA Models Goal: Relax FSA limitations Fixed allocation No sharing Strict Use [Zhao2007] Q. Zhao and B. M. Sadler, A survey of dynamic spectrum access, IEEE Signal Processing Mag. vol. 24, no. 3, [Buddhikot2007] M. M. Buddhikot, Understanding dynamic spectrum access: Models, taxonomy and challenges, DySPAN

13 DSA Models DSA: Exclusive Use Model Radio spectrum is licensed to a user/service to be exclusively used under a certain rule. may not fully utilize allocated spectrum in time/space. Spectrum access rights can be granted (spectrum trading) to cognitive radio users optimize spectrum use. Long Term Exclusive Use Service change Fixed-Type Flexible Type Exclusive Use Model Allocation granularity Non-real-time secondary market Dynamic Exclusive Use (secondary markets) Real-time secondary market for homogeneous multi-operator sharing Heterogeneous multi-operator sharing [Buddhikot2007] Dynamic Exclusive Use Spectrum Property Rights Dynamic Spectrum Allocation [Zhao2007] 13

14 DSA Models DSA: Shared Commons Open sharing among peer users as the basis for managing a spectral region. Success of wireless services in ISM band. Uncontrolled commons Spectrum not owned by any entity. Maximum transmit power constraint. CR users suffer from high interference levels Uncontrolled Controlled Managed commons Considers spectrum to be controlled jointly by a group of CRs. Requires: management protocol scalable mechanism Enforcement of spectrum access rules: Avoid CR missbehaviour. Private commons Spectrum owner specifies technology and protocol for the CR user access. CR user may receive a command from spectrum owner. transmission parameters. CR user may sense and access the spectrum 14

15 DSA Models DSA: Hierarchical Access Model Radio spectrum can be simultaneously shared between primary (licensed) user, and a secondary (unlicensed/cr) user. CR users can opportunistically access the radio spectrum if it is not occupied or fully utilized by primary users. Two different models Spectrum overlay [Opportunistic Spectrum Access (OSA)] E.g. FDMA, TDMA, OFDMA systems. Spectrum underlay E.g. CDMA, UWB systems. Tx. primary users Tx. secondary users Tx. primary users Tx. secondary users Overlay Frequency Underlay Frequency 15

16 DSA: Opportunity Space DSA exploitable domains Frequency Domain A frequency hopping problem. Time Domain A time scheduling problem. Geographical Space Domain A spatial re use problem. Hidden terminal problem Other domains Code (spread spectrum systems) Angle, Polarization Codes Tx. primary users Tx. secondary users Time 16

17 DSA: Architecture Infrastructure based: CR users transmit data through central controller (or Base Station) (a) Directly: single hop communication. (b) Between multiple relaying BSs: multi hop communication. Infrastructure less (Ad-Hoc): communication among unlicensed users on a direct peer to peer basis or, alternatively, through other unlicensed users acting as relays. [(c)] Which entity is responsible to manage spectrum access decisions: Centralized: the decision of spectrum access is made by a central controller Distributed: decision is made locally by each of the unlicensed users 17

18 MODELLING OPPORTUNISTIC SPECTRUM ACCESS 18

19 OSA: Some Basic Concepts Basic Concepts Primary Network Existing network infrastructure with access rights to spectrum band. E.g. cellular systems, TV broadcast networks. Primary User Has the licensed to operate in a certain spectrum band within the primary network. Secondary Network Does not have license to operate in a desired band. Thus, spectrum access on an opportunistic manner Secondary Has no spectrum license Additional functionalities required to access licensed band. OSA Problem Decomposed Spectrum Opportunity Detection: which bands can I access? Allocation: who can access and when? Exploitation: How do we access? Primary=licensed Secondary=cognitive, unlicensed PU Spectrum Collision SU PU Primary Base Station (PBS) Secondary Base Station (SBS) PU Spectrum frequency Spectrum Opportunity SU C channels SU SU overlay 19

20 Opportunity Detection To access the radio spectrum opportunistically an unlicensed user must ensure that the spectrum is not occupied by any licensed user... time Spectrum awareness models: Geo Located Database (Geo DB): CR users report their location to a DB which provides information on available bands. Beacon (Pilot) Reception: transmitted to unlicensed users over a common channel to inform on available bands. Spectrum Sensing: Unlicensed user devices perform sensing to identify available bands. 20

21 Opportunity Detection Geo Located Databases Challenges: Device location accuracy? Who will maintain (update) the database? What availability is needed for the database? What about dynamic use of spectrum? Beacon Reception Challenges: Who provides the beacon signal? How is the tx. information kept up to date? What spectrum is used for the beacon? Separate beacons for separate frequency bands or one beacon for all the bands? 21

22 Opportunity Detection Spectrum Sensing Decomposed Centralized Sensing Distributed Sensing Decision Spectrum Use Local Sensing (device centric) Cooperative Reactive Sensing (on demand) Proactive Sensing Energy Detector Matched Filtering Spectral Correlation (Cyclostationarity) Wavelet based sensing Waveform based sensing... Sensing dissemination Sensing Method Spectrum Sensing Hardware Approach Sensing Entity Standards employing sensing Single Radio Sensing Dual Radio Sensing Internal Sensing (at the terminal) External Sensing (e.g. sensor network) IEEE k IEEE Bluetooth 22

23 Opportunity Detection Some Challenging Problems in Spectrum Sensing Channel uncertainty channel dynamic variations due to fading and shadowing. Noise uncertainty unknown noise power to calculate the detection sensitivity of an unlicensed user. Aggregated interference uncertainty many CR users in the same CR network interfering with each other. Sensing interference limit Unknown location of licensed receiver (cannot compute interference on receiver). Unaware of licensed receiver if passive. Partially solved through Cooperative Sensing 23

24 Opportunity Detection Some Design Trade offs in Spectrum Sensing Optimizing the spectrum sensing period longer observation periods increased sensing accuracy but lower system throughput Cooperation vs. communication overhead sensing results collection and dissemination ΔT T sens T sens... Data Tx Data Tx... time Reactive (on-demand) vs. proactive sensing Reactive: lower overhead but larger spectrum access delays Proactive: larger overhead but lower spectrum access delays Rate and reliability tradeoff due to multi channel access Improved data rate multiple streams are transmitted over multiple bands Enhanced reliability when same data stream is transmitted redundantly over multiple bands Increased overhead required to sense and maintain spectrum usage statistics. Optimal decision is required for channel selection (e.g. how many and which bands to select for sensing). 24

25 Opportunity Detection Performance Characteristics of Spectrum Opportunity Detector Spectrum sensors perform binary hypotheses (H 0 or H 1 ) over a given band: ε = Pr( Y > λ H 0 is true) False Alarm probability Spectrum under use δ = Pr( Y < λ H 1 is true) Missed Detection probability Spectrum collision Y : decision statistic. λ : the decision threshold. Values of ε and δ mainly depend on: The time bandwidth product: m=t W The average SNR (γ). Improved ε and δ by means of sensing cooperation. 25

26 Opportunity Detection Energy Detection Pros. Not require prior information on licensed users. Simple implementation. T 0 γ = 10 db (SNR) Cons. Susceptible to the uncertainty of noise power Can only detect the presence of the signal (cannot differentiate the type of signal) Missed-Detection Probability (δ) 26

27 Opportunity Allocation Process of deciding which CR user will use which opportunity and for how long Similar to MAC problem in traditional wireless networks. Tradeoff: minimizing overlooked spectrum opportunities. avoiding collisions with primary users. Adopt/Adapt traditional MAC approaches Contention based: e.g. Aloha, CSMA Example: Extend RTS/CTS handshake to negotiate spectrum opportunities. Contention free: e.g. TDMA, FDMA, OFDMA Time Domain Scheduling... 27

28 Opportunity Use Waveforms that can use discontinuous portions of the spectrum (NC )OFDM becomes an attractive option Control transmission parameters to ensure that the transmission does not affect existing users Band of operation and bandwidth Transmission power control Modulation level Coding Beamforming (directional antennas) FFT Pruning

29 FINAL REMARKS AND CONCLUSIONS 29

30 Concluding Remarks Cognitive Radio (CR) notion with multiple/contradictory definitions: CR is the enabling technology for Dynamic Spectrum Access DSA advocates for the temporary use of non utilized licensed bands Spectrum sharing can be achieved in time, frequency, space etc. Spectrum sharing models: Exclusive Use, Open Sharing, Hierarchical Access Model. Architecture: Infrastructure/ less, Multi /Single hop, Centralized/Distributed decision. Detection of spectrum opportunities based on the detection of primary signals Geo located DB, Beacon Signals, Sensing. Opportunity Allocation as an extension of the MAC problem in traditional networks Spectrally agile wireless transceiver necessary for improving spectrum efficiency: NC OFDMA 30

31 Future Intelligent systems Higher computational capability More flexibility Harvesting more and more radio spectrum Digital dividend opportunity and more standards to come Number of publications including the term «cognitive radio». Source: Engineering Village, Accessed 27th October

32 THANK YOU FOR YOUR ATTENTION! Funding provided by the Ministerio de Industria, Turismo y Comercio through project Avanza I+D (ref. TSI ) CORAGE: Cognitive Radio Generation is hereby acknowledged. 32