What You Will Learn Understanding the Issues in Software Defined Cognitive Radio Jeffrey H. Reed Charles W. Bostian Virginia Tech Bradley Dept. of Electrical and Computer Engineering
Basic Concepts of Software Defined Radio (SDR) Basic Concepts of Cognitive Radio (CR) and its relationship to SDR. How Cognitive Radios are Implemented Analyzing Cognitive Radio Behavior and Performance Regulatory Issues in Cognitive Radio Deployment Cognitive Radio Applications in Interoperability and Spectrum Access Current Research Issues 2
1
Acknowledgements
Students who contributed to this presentation:
Albrecht
Johannes Fehske
Thomas Rondeau
Bin Le
James Neel
David Scaperoth
Software Defined Radio – Basic Concepts and Relationship to Cognitive Radio
Kyouwoong
Kim
David Maldonado
Lizdabel Morales
Youping Zhao
Joseph Gaeddert
3
2
Software Defined Radio (SDR)
Software Defined Radio Levels (1/2)
Termed coined by Mitola in 1992 Radio’s physical layer behavior is primarily defined in software Accepts fully programmable traffic & control information Supports broad range of frequencies, air interfaces, and application software Changes its initial configuration to satisfy user requirements
Highest Level of Reconfigurability
Completely
flexible modulation format, protocols and user functions
Flexible bandwidths and center frequency, i.e., RF front end is also configurable
Adapts to different network and air interfaces
Open architecture for expansion and modifications
5
6
3
Advantages of SDR
Software Defined Radio Levels (2/2)
Lowest
Level of Reconfigurability
Radio
not easily changed
Preset signal bandwidth and center frequency
Few and preset modulation formats, protocols, and user functions
7
Reduced content of expensive custom silicon Reduce parts inventory Ride declining prices in computing components DSP can compensate for imperfections in RF components, allowing cheaper components to be used Open architecture allows multiple vendors Maintainability enhanced 8
4
Drawbacks of SDR
Applications for SDR
Power consumption (at least for now) Security Cost Software reliability Keeping up with higher data rates Fear of the unknown Both subscriber and base units should be SDR for maximum benefit
Military
Full
Connectivity
Sensor Networks
Better Performance
Commercial
Lower
Cost – subscriber units Cost – base unit
Lower Cost – network
Better performance
Lower
Regulatory
Stretch
Build
9
expensive spectrum in innovation mechanisms 10
5
How is a Software Radio Different from Other Radios?- Design
How is a Software Radio Different from Other Radios? - Application
Conventional Radio Supports a fixed number of systems Reconfigurability decided at the time of design May support multiple services, but chosen at the time of design
Software Radio Dynamically support multiple variable systems, protocols and interfaces Interface with diverse systems Provide a wide range of services with variable QoS
Cognitive Radio
Can create new waveforms on its own Can negotiate new interfaces Adjusts operations to meet the QoS required by the application for the signal environment
11
Conventional Radio
Traditional RF Design Traditional Baseband Design
Software Radio
Conventional Radio + Software Architecture Reconfigurability Provisions for easy upgrades
Cognitive Radio
SDR + Intelligence Awareness Learning Observations
12
6
How is a Software Radio Different from Other Radios? - Upgrade Cycle Conventional Radio Cannot be made “future proof” Typically radios are not upgradeable
Software Radio Ideally software radios could be “future proof” Many different external upgrade mechanisms Over-the-Air (OTA)
Cognitive Radio SDR upgrade mechanisms Internal upgrades Collaborative upgrades
Cognitive Radio Concepts
13
7
What is a Cognitive Radio?
Cognitive Radio
are set by their operators
Term coined by Mitola in 1999 Mitola’s definition:
Fixed radios
Adaptive radios
can adjust themselves to accommodate anticipated events
Software radio that is aware of its environment and its capabilities Alters its physical layer behavior Capable of following complex adaptation strategies
“A radio or system that senses, and is aware of, its operational environment and can dynamically and autonomously adjust its radio operating parameters accordingly” Learns from previous experiences Deals with situations not planned at the initial time of design
Cognitive radios
can sense their environment and learn how to adapt
15
Beyond adaptive radios, cognitive radios can handle unanticipated channels and events. Cognitive radios require: • Sensing • Adaptation • Learning Cognitive radios intelligently optimize their own performance in response to user requests and in conformity with FCC rules.
16
8
Cognitive radios are a powerful tool for solving two major problems:
Cognitive radios are machines that sense their environment (the radio spectrum) and respond intelligently to it.
Like animals and people they
1) Access to spectrum (finding an open frequency and using it)
• seek their own kind (other radios with which they want to communicate) • avoid or outwit enemies (interfering radios) • find a place to live (usable spectrum) • conform to the etiquette of their society (the Federal Communications Commission) • make a living (deliver the services that their user wants) • deal with entirely new situations and learn from experience 17
18
9
Cognitive radio platforms are a powerful tool for solving two major problems: 2) Interoperability (talking to legacy radios using a variety of incompatible waveforms)
Levels of Radio Functionality Level
19
Capability
Comments
0
Pre-programmed
A software radio
1
Goal Driven
Chooses Waveform According to Goal. Requires Environment Awareness.
2
Context Awareness
Knowledge of What the User is Trying to Do
3
Radio Aware
Knowledge of Radio and Network Components, Environment Models
4
Capable of Planning
Analyze Situation (Level 2& 3) to Determine Goals (QoS, power), Follows Prescribed Plans
5
Conducts Negotiations
Settle on a Plan with Another Radio
6
Learns Environment
Autonomously Determines Structure of Environment
7
Adapts Plans
Generates New Goals
8
Adapts Protocols
Proposes and Negotiates New Protocols
Adapted From Table 4-1Mitola, “Cognitive Radio: An Integrated Agent Architecture for Software Defined Radio,” PhD Dissertation 20 Royal Institute of Technology, Sweden, May 2000.
10
Relationship between the Cognition Cycle and the Levels of Functionality
What is a cognitive radio? Cognitive radio
Cognition Cycle
Infer from Context An enhancement on the Orient Infer from Radio Model Establish Priority traditional software radio Normal Pre-process Select Alternate concept wherein the Goals Parse Stimuli Urgent Immediate radio is aware of its Plan environment and its capabilities, is able to Learn Observe New independently alter its States Decide physical layer behavior, States and is capable of User Driven Generate “Best” Autonomous (Buttons) Waveform following complex Act adaptation strategies. Outside Adapted From Mitola, “Cognitive Radio for Flexible Mobile Multimedia Communications ”, IEEE Mobile Multimedia Conference, 1999, pp 3-10.
World
Allocate Resources Initiate Processes 21 Negotiate Protocols
Level
Infer from Context
0 SDR 1 Goal Driven 2 Context Aware 3 Radio Aware 4 Planning 5 Negotiating 6 Learns Environment 7 Adapts Plans 8 Adapts Protocols
Infer from Radio Model
Orient
Establish Priority Pre-process Parse Stimuli
Observe
User Driven Autonomous (Buttons)
Outside World
Immediate
Select Alternate Generate Normal Goals Normal Urgent
Plan
Learn
New States
Decide
States
Determine “Best” Plan Determine “Best” Generate “Best” Waveform Allocate Resources Known Waveform Initiate Processes Negotiate Protocols Negotiate
Act
22 Adapted From Mitola, “Cognitive Radio for Flexible Mobile Multimedia Communications ”, IEEE Mobile Multimedia Conference, 1999, pp 3-10.
11
FCC Motivation for Cognitive Radio
Cognitive Radio Advantages
Currently the FCC is refarming licensed bands such as the TV Bands Long-term vision
Eliminate
rigid, coarse spectrum allocations
Switch to demand-based approach
All of the benefits of software defined radio Improved link performance
Improved spectrum utilization
Improve relative spectral efficiency
Need new protocols for long-term vision of the FCC
Inter-network interference avoidance
Maximizing utilization of available bandwidth
High speed internet in rural areas High data rate application networks (e.g., Video-conferencing)
Significant interest from FCC, DoD
23
Fill in unused spectrum Move away from over occupied spectrum
New business propositions
Supporting
Adapt away from bad channels Increase data rate on good channels
Possible use in TV band refarming 24
12
Cognitive Radio & SDR
Cognitive Radio Drawbacks
All the software radio drawbacks Significant research to realize
Information
collection and modeling processes
Learning processes
Hardware support
Decision
SDR’s impact on the wireless world is difficult to predict
Some believe SDR is not necessary for cognitive radio
Cognition is a function of higher-layer application
Cognitive radio without SDR is limited
Regulatory concerns Loss of control Fear of undesirable adaptations
“But what…is it good for?” Engineer at the Advanced Computing Systems Division of IBM, 1968, commenting on the microchip
Underlying radio should be highly adaptive Wide QoS range Better suited to deal with new standards
Need
some way to ensure that adaptations yield desirable networks
25
Resistance to obsolescence
Better suited for cross-layer optimization 26
13
Policy-based Radio
Types of Software Defined Cognitive Radios
Policy-Based Radio Reconfigurable Radio Cognitive Radio
A radio that is governed by a predetermined set of rules for choosing between different predefined waveforms The definition and implementation of these rules can be:
during
the manufacturing process configuration of a device by the user;
during over-the-air provisioning; and/or
by over-the-air control
during
Analogous to rules of what to order from a menu
“I’ll
27
have GSM today” 28
14
Reconfigurable Radio A radio whose hardware functionality can be changed under software control Reconfiguration control of such radios may involve any element of the communication network Analogous to rules of what to order from a menu and permit substitutions to the order
“I’ll
Technology Challenges in SDR
have GSM today with the 802.11 FEC” 29
15
Behind the Converters: The Software Architecture
Radio Architecture
Superhetrodyne RF Signal
Amplify Mixer Filter
IF Signal
Amplify Mixer Filter
Baseband Signal
Software Defined Radio RF Signal
Rx Tx
Amplify Mixer Filter
IF Signal
Analog To Digital Converter
Digital Signal Processing
31
Nature of Architecture Depends on Applications: Commercial vs. Military Benefits of a Good Architecture
Clear way to implement system
Reuse --- modularity
Quality control and testing
Portability – one radio to another
Upgradability
Outsourcing/managing development
Language independence
More potential for Over-the-Air Programming
Standardized interfaces Middleware-based architectures are commonly used 32
16
Implementing a SDR with the GNU Radio USRP - Universal Software Radio Peripheral
USRP
GNU Radio software - core s/w - user made s/w
4 ADC’s: •12bits per second, 64MSps, •Analog Input BW over 200Mhz
4 DAC’s •14bits per second, 128MSps
Transmit Channel RF Interface
Receive Channel RF Interface
GNU Radio S/W available at www.gnuradio.org
Courtesy:http://www.gnu.org/software/gnuradio/doc/expl oring-gnuradio.html
33
34
17
Challenges in SDR Design
Hardware
Significant effort in computing HW Advance DSP Designs Flexible RF and antennas Flexible ADCs Tradeoff of performance and flexibility
Physical devices on which processing is performed or interface to the “real world”
Software Glue holding together system
Network Functionality and ultimate value to the end-user
Integration of components into single design flow Tradeoff of performance and flexibility
Testing and validation
Technology in SDR partitioned into three basic pieces
Hardware
Software
Technology Challenges of SDR
FCC hardware/software certification Smoothing of design cycle Reduce overall time-to-market
35
Advances needed in all three arenas 36
18
Hardware
Flexible RF
Significant effort to date in computing HW
Non-traditional
computing platforms
Advanced DSP designs
High data rate FEC remains problematic
Places
fundamental limits on the signal characteristics
Emphasis on computing HW alone can be myopic
Other
RF is one of the main limiting factors on system design
Truly
critical areas that require significant further
Flexible (or software controlled) RF Flexible ADC Antennas
37
flexible SDR requires flexible RF
Difficult task
work
BW, SNR, linearity
RF is fundamentally analog and requires different a different approach for the management of attributes
One method for achieving this is through the use of MEMS 38
19
MEMS (Micro Electro Mechanical Systems) Designs for RF Front Ends
ADC Challenges
ADC is the bound between analog and digital world SDR requires the tuning of ADC characteristics
Number
E-tenna’s Reconfigurable Antenna
Sampling
Tunable antenna with narrow fixed bandwidth Patch antenna connected by RF switches
of bits
Support adequate SNR and dynamic range
Idealized MEMs RF Front-end for a Software Radio
rate
Prevent over-sampling (waste power)
ADC technology trends are not necessarily compatible with these needs
Use MEMS filter banks to create tunable RF filters J.H. Reed, Software Radio: A Modern Approach to Radio Design, Prentice-Hall 2002.
39
40
20
ADCs Getting Better Exponentially
ADC: Improving Even When Considering Power
P = 2B ⋅ f s
F=
B bits fs sample rate
2B ⋅ f s Pdiss
Pdiss is power dissipation
1994 ~ 2004 a leap of Analog to Digital Converter (ADC) technology Regression curve fit shows exponential increasing trends Trends are quite different for different ADC structures Bin Le, Tom Rondeau, Jeff Reed, Charles Bostian, “Past, Present, and Future of ADCs,” IEEE Signal Processing Magazine, November 2005 41
Power-to-sampling-speed ratio favors less number of comparators The choice in selecting an ADC is tied to application requirement Bin Le, Tom Rondeau, Jeff Reed, Charles Bostian, “Past, Present, and Future of ADCs,” IEEE Signal Processing Magazine, to be published, November, 2005 42
21
Integration of Hardware
Software Operating Environment
DSP share traits with GPP
Similar
programming methods
Similar computing concepts
SCA
Even though implementation may be wildly different
FPGA and CCM do not share these traits with GPP
Technology to date has been largely derived from existing PC paradigm
GPP-centric
structure 3.0 Hardware Supplement is an attempt to rectify this problem
SCA
Completely
Portability
Standardized structure for the management of HW and SW components
different programming paradigm is an extremely difficult problem
Several challenges remain
Power
management
Integration of HW into structure 43
44
22
Software Architectures
“The sheer ease with which we can produce a superficial image often leads to creative disaster.” Ansel Adams [1902-1984], American artist (photography)
So How Do You Make a Software Radio?
You have some hardware
And you want to run some waveforms
Poor architectural design leads to significant inefficiencies
Architectures provide multiple benefits
Clear way to implement system
Generally component-based
Standard technology interface
Standard semantic -- API
Software or hardware components
Standardized interfaces Common technology like middleware
Architectures becoming more prominent
Software Communications Architecture (SCA) $14B to $27B for SCA radio work by DoD Cluster 5 contract up to $1B for embedded & handheld prototypes Maintain awareness of activity: big money for SDR 45
GSM,
IS-95, or some other technology that the hardware is powerful enough to support 46
23
What kind of software is needed? (1/4)
What kind of software is needed? (2/4)
Something to manage hardware
Configure
More
Set devices to known state
i.e.: Make sure NCO is available and ready
Initialize
associated devices
Some standardized way of storing relevant information
cores
Make sure programmable devices are ready
Set memory pointers in DSP
Set FPGA to known state
than just short-term memory
Store configuration files Store last state of the machine Store user-defined attributes Identity Permissions
Store functional software
Should
be able to map any kind of storage device to
this
47
Dynamic RAM, hard drive, FLASH, other 48
24
What kind of software is needed? (3/4)
What kind of software is needed? (4/4)
Some way of structuring the waveforms
Install
functional software in appropriate core
Generate a start event
Standardized
way of structuring “applications” so that the radio can “run” them
In a Windows machine, these are .exe files
It
has to be generic enough for it to fit well with machines other than GPPs
Something to actually “run” waveforms
Something to keep track of what is available and what can and cannot be installed
Ideally,
Needs to be able to interface with functional software
49
this will bind the whole thing together
50
25
Domain Manager Keep track of what’s there (installed)
Processor-centric structure
Resources Capabilities e.g., Start and stop, test, describe
Standardized interface for components
Open-source implementations available
OSSIE
SCARI
51
C++ by MPRG
CORBA Adapter
Software IDL CORBA OS
Non-CORBA Software (Legacy)
Java by Communications Research Centre
Non-CORBA Software (Legacy)
Black Red Software
AP
Port Connections between resources
Seamless handling of HW and SW
I
Secure
Security Boundary
CORBA Adapter
Management Objects File System Configuration Files
I AP
Application Manage waveform operation
Non-secure
API
Devices Boot up and maintain HW
I AP
FileSystem Manager Store working environment, bit images, properties, etc.
Application Factory Manage collection of resources to create waveform
Software Communications Architecture (SCA)
I
Device Manager Keep track of HW in the system
AP
Fundamental Composition of the SCA
Hardware Hardware Trans. Security
52
26
Is the SCA Suitable for Commercial Implementations?
Summary of Trends
Maybe
No
Current version is GPP-centric, hence heavy Irrelevant capabilities decrease its effectiveness Focus on waveform portability has limited appeal Static nature not well suited for cognitive radio No provisions for power management
Yes Basic architectural principles are sound SCA 3.0 is a first step in dealing with GPP-centric communications within the radio Significant momentum ($$$ and time) within defense industry Being adopted by several other nations’ defense establishments 53
SDR need is driven by two principal factors
New applications
Increased number of protocols to support Potential cost reductions
Cognitive radio, collaborative radio & advanced roaming
ADC is no longer the key bottleneck Flexible RF products starting to come to market Software architecture critical
Reconfigurable hardware needed
Additional technology supporting architectural approach available General-purpose hardware approach is likely to be unable to keep up with wireless bandwidth growth Component-based reconfigurable hardware architectures present powerful solution Multi-core processors show promise 54
27
SDR Market Today
Military
JTRS
program created multi-billion dollar SDR market neXt Generation (XG) Communications project
International derivatives of JTRS/SCA (EU, Canada, etc)
DARPA
Cognitive Radio Implementation
Commercial
Digital
RF processors (TI Bluetooth and GSM) base station implementations (Vanu)
SDR handsets probably within 3 years as low power processors become available
Multi-standard
Regulatory
Recent
FCC directive to ensure code and RF compatibility
55
28
The VT Cognitive Engine
The VT Cognitive Engine
Simple Concept
Simple Concept Radio TX
Radio Channel TX Statistics Radio RX Radio Parameters
Channel Statistics “Old Knobs Settings”
“Knobs and Meters”
Cognitive Engine
“Meters”
Radio RX
“Old Knobs Settings”
Cognitive Engine “New Settings”
“Optimized Solution”
“New Settings”
Radio Parameters “Knobs and Meters”
57
58
29
Layer
Knobs and Meters
Meters (observable parameters)
Knobs (writable parameters)
MAC
Frame error rate Data rate
Source coding Channel coding rate and type Frame size and type Interleaving details Channel/slot/code allocation Duplexing Multiple access Encryption
PHY
Bit error rate SINR Received signal power Noise power Interference power Power consumption Fading statistics Doppler spread Delay spread Angle of Arrival
Transmitter power Spreading type and code Modulation type Modulation index Pulse shaping Symbol rate Carrier frequency Dynamic range Equalization Antenna directivity
Computational power Battery Life
CPU Frequency scaling
Other
Sample tabulation of knobs and meters by layer (adapted from Prof. Huseyin Arslan) 59
The VT Tiered Approach to Cognition
Modeling System
Take in surrounding radio environment and user/network requirements Remember models and apply Case-based Decision Theory to determine best course of action to take Use Genetic Algorithms to update and optimize the new radio parameters
Monitor feedback from radio to understand system performance
Penalize knowledge base for poor performance 60
30
Software Architecture - Theory
The Cognitive Engine
“Intelligent agent” that manages cognition tasks in a Cognitive Radio Independent entity that oversees cognitive operations Ideal Characteristics:
Awareness Sensing and Modeling
Intelligence
(Accurate decisions) (Consistent decisions)
Awareness (Informed decisions)
Adaptability (Situation dependent decisions)
Efficiency (Low overhead decisions)
Excellent QoS (Good decisions)
Reliability
Learning Building and retaining Knowledge
Radio Hardware
Adapting Evolution and Optimization
Tradeoffs exist between these characteristics 61
62
31
Software Architecture - Theory Scenario Synthesizing
Radio
Case identified
Link condition
Reasoning
Radio hardware
Success memorized Radio Hardware
Case report
Apply experience Bad trail overwritten
Performance Estimation
Case-based Decision Making
API
Modeling System
Cognitive System Module
Policy Domain User preference Local service facility
Strategy instruction
Link Configure Optimization
Security User data security System/Network security
WSGA Initialization Objectives Constraints 63
WMS
CE-user interface
User/policy
CE-Radio Interface
Knowledge Base
Cognitive System Controller wavfrm Policy Policy Model
Security
Decision Maker
Sec
Selector
Environment Observation
Software Architecture – Limited Functionality
Knowledge Base Short Term Memory Long Term Memory
64
32
Software Architecture: Full Functionality
Some Approaches to Cognitive Engine
Radio
Channel Probe
CE-Radio Interface
API
Genetic Algorithms Markov Models Neural Nets Expert Systems Natural Language Processing Fuzzy Logic
Cognitive System Module WMS
Cognitive System Controller
User Domain User preference Local service facility
Security
Modeling System
User Model
Chob Resource Monitor
User data security System/Network security
Policy Domain
Evolver Policy Model
Radio
Uob
Actual Meters
|(Simulated Meters) – (Actual Meters)| Simulated Meters
Reg
User preference Local service facility
Decision Maker Security X86/Unix Terminal
DCH = max{S CH • U CH } DU = max{SU • U U }
Initial Chromosomes WSGA Parameters Objectives and weights
Knowledge Base Short Term Memory Long Term Memory WSGA Parameter Set Regulatory Information
Open issue on what are the appropriate cognitive engine techniques
System Chromosome
65
66
33
Some Approaches to Signal Classification
GA’s and biological metaphor The WSGA uses a genetic algorithm, which operates on chromosomes.
Cyclic spectrum analysis Statistical characterization of signal parameters Eigenstructure techniques Model-based approaches Vector space (I-Q plane) approaches
The genes of the chromosome represent the traits of the radio (frequency, modulation, bandwidth, coding, etc.). The WSGA creatively analyzes the information from the CSM to create a new radio chromosome.
67
68
34
Analyzing the Performance of a Network of Cognitive Radios
Analyzing Performance in a Cognitive Radio
70
35
Ways of Analyzing Performance
For the Cognitive Radio Detection of Primary Users (PU), SW Platform, QOS of PU, Position Location
throughput
Latency
These depend on link performance measures:
PHY Layer, e.g.:
Voice
quality
Video quality
For the network of Cognitive Radios
Quantifying
Parameters
Data
QOS,
Cognitive Radio Performance Evaluation: QoS
the impact of the use of CR in a
network
Game Theoretic Approach See www.mprg.org/gametheory
MAC, network-layer, e.g.:
71
Bit error rate (BER) Signal to noise ratio (SIR) Signal to interference and noise ratio (SINR) Received signal strength Frame error rate (FER) Packet error rate Routing table change rate
72
36
Cognitive Radio Performance Evaluation: Detection of Primary Users
Cognitive Radio Performance Evaluation: Underlying Software Radio Platform
Probability of detection (PoD) as a function of:
Number
of observed symbols
SNR
Number
of signals present (primary and secondary)
Level of cooperation, e.g., number of devices (CRs) needed to achieve a given PoD (see next slide)
Probability of false alarm
Same
Number of supported waveforms Processing power (mips, flops, #gates) Waveform-code reusability and portability
parameters as PoD
Reusable: the same code can be used in principle in a different SDR platform Portable: instantaneous plug and play
73
Delay for loading unloading waveforms RF front-end: Frequency range, Dynamic range, Sampling frequency, Sensitivity, Selectivity, Stability, Spurious response
Power consumption Size, Weight, Cost
74
37
Cognitive Radio Performance Evaluation: Position Location
Main performance measures for position location service:
Cognitive Radio Performance Evaluation: Primary users' QoS
Precision and Availability
Different technologies provide different quality of position location services:
performance degrades significantly when no clear view of sky (indoors, urban canyons) works best in rural areas (no shadowing)
Network based services
Assisted GPS (AGPS)
Time needed to vacate channel after primary user (re-) appears Negative impacts:
accuracy in general lower than AGPS works best with many base stations present (populated areas) performance doesn't degrade indoors
Hybrid services
combines advantages of both approaches AGPS whenever possible, if not available switch to network based service
75
Decreased SINR and Increased BER, FER, … results in: Decreased: Data throughput Latency Voice quality Video quality Increased Call drop rate (cell phone networks) Handover failure (cell phone networks) 76
38
Locally optimal decisions that lead to globally undesirable networks
Dynamic cognitive radios in a network
Dynamic benefits
Improved spectrum utilization Improve QoS
Many decisions may have to be localized
Adaptations of one radio can impact adaptations of others
Distributed behavior
Interactive decisions Locally optimal decisions may be globally undesirable 77
Scenario: Distributed SINR maximizing power control in a single cluster Power For each link, it is SINR desirable to increase transmit power in response to increased interference Steady state of network is all nodes transmitting at Need way to analyze networks maximum power with interactive decisions. Game theory can help. 78
39
What is a game?
Key Issues in Implementation
A game is a model (mathematical representation) of an interactive decision process. Its purpose is to create a formal framework that captures the process’s relevant information in such a way that is suitable for analysis. Different situations indicate the use of different game models. Identification of the type of game played by the cognitive radios provides insights into performance
NE3 NE3
1. 2. 3. 4. 5.
Steady state characterization Steady state optimality Convergence Stability Scalability
a2
NE2
NE1 NE1
a1 a1 a3
79
Convergence Optimality Scalability Stability Steady State Characterization As Are How these do does number initial outcomes system of devices variations desirable? impact increases, impact the system thesystem? system? steady state? Is itthe possible toconditions predict behavior in the How Do What these the is processes steady theoutcomes system states willimpacted? maximize lead change? to steady the state conditions? target parameters? many different outcomes are system possible? Do How Is convergence previously long doesoptimal itaffected? take steady to reachstates the steady remainstate? optimal? 80
40
An Analogy between a Cognitive Radio and a Car Driver
Cognitive Radio, Spectrum Policy, and Regulation
Cognitive Radio’s capabilities: 9 Senses, and is aware of, its operational environment and its capabilities 9 Can dynamically and autonomously adjust its radio operating parameters accordingly 9 Learns from previous experiences 9 Deals with situations not planned at the initial time of design
Car Driver’s capabilities: 9 Senses, and is aware of, its operational environment and its capabilities 9 Can dynamically and autonomously adjust the driving operation accordingly 9 Learns from previous experiences 9 Deals with situations not planned at the initial time of learning to drive They behave almost exactly the same!!! 82
41
“Rules of the Road” ➟ “Rules of the Cognitive Radio”
“Rules of the Road”-inspired CR Philosophy and Etiquette Insights from “Traffic Model Analogy”
POLICY AWARE
LOCATION AWARE
Traffic Law ➟ Spectrum Regulations
Primary User has higher priority over Secondary users
Precautions for certain areas, such as hospital, airplane, gas station, etc, where RF emission is highly restricted
Management by both Punishment and Encouragement
Radio emission may be prohibited at certain location or for certain type of radio
TRAFFIC Scheduling
Spectrum pooling is encouraged
Various traffic schedule methods and duplex methods for efficient and fair sharing of congested unlicensed spectrum TDD vs. FDD ➟ Dynamic Uplink/Downlink transmission in TDD mode
Parking Zone *Source of some pictures in this section: “California Drivers Handbook 2005”; “Illinois Rules of the Road 2004”
$ fine
FDD mode operation with paired spectrum 83
84
42
A traffic model analogy – Common Issues
A traffic model analogy (cont.)
It is critical that everyone drives sensibly or defensively ➟ Every CR should be aware of Hidden Node problems
Vehicle Following Distances: TWO-SECOND RULE: Use the two-second rule to determine a safe following distance.
Vehicle Following Distances for Car Drivers Hidden Node Problem
➟ Time needed to vacate channel after primary user (re-) appears for Cognitive Radios
A and C are unaware of their interference at B. Due to A, C and B cannot hear each other. 85
86
43
Learning from “Traffic model analogy” for the development of Cognitive Radio…
A traffic model analogy (cont.)
SPEED LIMIT for car driver
REM
➟ Interference Level Limit (e.g. Max. Allowed Interference Temperature)
Time (or duration) Location (x, y, z), Type of radio environment Local Policy
for Cognitive Radio City Map for Car Drivers
Profile of primary users Profile of interference Max. allowed Interference Level
➟ Radio Environment Map (REM) for Cognitive Radios 87
88
44
Spectrum Policy Challenges
Learning from “Traffic model analogy” for the development of Cognitive Radio…(cont.)
The spectrum is already allocated
True
Language and Etiquette for CR for Signaling and Negotiation
spectrum scarcity on urban areas (ISM band)
We need to deal with existing standards The standards are embedded in the hardware!
Regular conformance check against regulations 89
90
45
Spectrum Occupancy Study
Spectrum Utilization
Spectrum utilization is quite low in many bands Concept:
Have
radios (or networks) identify spectrum opportunities at run-time
Transparently (to legacy systems) fill in the gaps (time, frequency, space)
Considered Bands
ISM
Public
TV
Safety (UHF)
Spectrum occupancy in each band averaged over six locations (Riverbend Park, Great Falls, VA, Tysons Corner, VA, NSF Roof, Arlington, VA, New York City, NRAO, Greenbank, WV, SSC Roof, Vienna, VA)
dBµV/m
Source: FCC NPRM 03-0322. http://hraunfoss.fcc.gov/edocs_public /attachmatch/FCC-03-322A1.pdf
From F. Jondral, “SPECTRUM POOLING - An Efficient Strategy for Radio Resource Sharing,” Blacksburg (VA), June 8, 2004.
Lichtenau (Germany), September 91 2001
Results from Shared Spectrum Co. and Univ. of Kansas
92
46
Regulatory Trends
Regulatory Trends
In
Proceedings that are the Key Drivers:
an effort to improve radio spectrum management and promote its more efficient use, the regulatory bodies are trying to adopt a new spectrum access model. This represents a paradigm shift from hardware-embedded policy implementation to dynamic softwarebased adaptation
Harder to keep tight control! 93
Receiver Standards
Interference Temperature
ET Docket No. 03-65 NOI ET Docket 03-237 NPRM/NOI
Cognitive Radio
License-exempt Operation in the TV Broadcast Bands
Additional Spectrum for License-exempt devices below 900 MHz and in the 3 GHz Band
ET Docket No. 03-108 NPRM ET Docket No. 04-186
ET Docket No. 02-380
94
47
Policy Engine Approach
DARPA XG Program
PE needs to provide limiting operational parameters
Interpret
policy automatically
Act dynamically in response to the operating environment
Sense Sense
PE needs to authenticate the policy It will require an extremely efficient policy format
It
must handle the complexity of current policy without presenting a significant load to the CE
XG is trying to Develop the Technology and System Concepts to Dynamically Access Available Spectrum
The goal is to limit the search space before looking for a solution
Real Low Realtime, time,LowLowpower, power,wideband wideband monitoring monitoring
Adapt Adapt Transition Transition network networkto tonew new emission emissionplan plan
Autonomous Dynamic Spectrum Utilization
Characterize Characterize
Goal: Demonstrate Factor of 10 Increase in Spectrum Access
Rapid Rapidwaveform waveform determination determination
React React Formulate FormulateBest Best Course Courseof ofAction Action
Rely
on CE to do the reasoning about spectrum sharing
Source: DARPA XG Program
95
96
48
XG Program Aspects Measurements Measurements
Policy-Based Controls Policy Policy-Based Controls
XG Products
Temporal, Spectral, Dimensional, Energy Characteristics
XG XG Behaviors Behaviors
Initial Initial XG XG Implementation Implementation
Control of Features, Priorities, Allocations, Exclusions,…
The BIG Question: FCC Certification At all costs, the FCC must avoid “an epidemic situation in the unlicensed area.”
Military & Civil Communications and Sensor Applications
Transition to Military Use
The Primary Product XG Program is Not a New Radio,, but a Set of Advanced Technologies for Dynamic Spectrum Access
97
FCC likes to operate from “established engineering practices.” The SDR and CR communities must defined these. Open source radios are a particular problem because their operating parameters are not necessarily bounded. 98
49
Proposed Approach
People seeking certification must explain how their software will respect parameter limits specified in FCC rules.
Bios/OS
Submitted software must be accompanied by flow charts, code, and an explanation of how it works. Applications
Policy Engine
Software certification should not be more difficult to achieve than hardware certification.
Cognitive Engine
99
100
50
How can CR improve spectrum utilization?
Example of a Possible Cognitive Radio Application
Allocate the frequency usage in a network. Assist secondary markets with frequency use, implemented by mutual agreements. Negotiate frequency use between users. Provide automated frequency coordination. Enable unlicensed users when spectrum not in use. Overcome incompatibilities among existing communication services.
102
51
How can CR improve network management efficiency?
Present practice characterizes service demand in a network statistically By using cognitive radio, time-space characterization of demand is possible Cognitive Radio
How can a CR enhance service delivery?
Learns plans of the user to move and use wireless resources Expresses its plans to the network reducing uncertainty about future demand
For
example: actual position, native language, habits, travel, etc.
The network can use its resources more efficiently
103
Wireless communications in general and cognitive radio in particular have great potential to generate personal user information
Enhanced services can be provided using this information CR interacts with the network on user’s behalf 104
52
Example of Cognitive Radio in Cellular Environment
CR in a Cellular System
Good signal
3. Signal Base Station
Transition in signal
Request Decrease In Call Drop Threshold
2. Evaluate Alternatives
4. Adapt Network
Do Nothing Increase Coding Gain Increase Transmit Power Vertical Handoff Decrease Call Drop Threshold
1. Observe and Analyze Situation
Cognitive radio is aware of areas with a bad signal Can learn the location of the bad signal
Has “insight”
Radio takes action to compensate for loss of signal
Note Daily Drive Home at 5:30 (GPS Aided) Recall Brief Coverage Hole
Actions available:
105
Bad signal
Power, bandwidth, coding, channel
Radio learns best course of action from situation 106
53
Supplements Cellular System
Cellular systems are plagued with coverage gaps Cognitive radio can enhance coverage around these gaps by:
Current Research Efforts in Cognitive Radio
Learning
the areas of coverage gaps
Learning the best PHY layer parameters
Taking action prior to getting to the area
Sharing this knowledge with other cell phones
Coverage gaps are found very rapidly
Alert
cellular system of gap, so provider can remedy situation 107
54
Universities Participating at DySPAN
Bar-Ilang Univ. Georgia Tech Mich. State Univ. Michigan Tech MIT Northwestern Univ. Ohio Univ. Rutgers Univ. RWTH Aachen Univ. Stanford Univ.
Univ. of Calif. Berkeley Univ. of Cambridge Univ. of Col. Univ. of MD Univ. of Pittsburg Univ. of Toronto Univ. of Warwick Universitaet Karlsruhe University of Piraeus Virginia Tech
DARPA
109
55
DARPA neXt Generation Program: Motivation - Problems
DARPA neXt Generation Program: Research Goals
Spectrum Scarcity
1.
Spectral
resources are not fully exploited
Opportunities exist in space, time, frequency
Current static spectrum allocation prevents efficient spectrum utilization
Development of technologies that enable spectrum agility
Deployment difficulty
Different
policy regimes in different countries of communication networks tedious
Of particular interest in military applications
Deployment
2.
Development of standards for a software based policy regime to enable policy agility
Unless otherwise stated, all the information in this description of the DARPA XG program is based on the XG Vision rfc, available online: http://www.darpa.mil/ato/programs/xg/
111
Sensing and characterization of the (RF-) environment Identification of unused spectrum ("opportunities") Allocation and exploitation of opportunities
Explained in more detail on the next slides 112
56
1.
Decoupling of policies from implementation
2.
Define abstract behaviors, e.g., "Channel can be vacated within t sec." Policies implement (dictate) behaviors Protocols instantiate behaviors
Traceability
All behaviors must be traceable to policies:
3.
Each operational mode a device is capable of is tied to a specific policy which allows it
Software based
XG Operation Sensing Loop Message Flow
DARPA neXt Generation Program: Concepts of Policy Agility (1)
RF Info Acquisition
RF Resource Request
Develop Options
Radio Platform RF Transmit Plan
Spectrum use policies have to be machine understandable Policy constraints can be implemented "on-the-fly" via software downloads
System Strategy Reasoner Select s Opportunitie
Transceiver 113
Policy Reasoner Process Request
ine Determ ities n tu Oppor nal ditio
Accredited Policy
d or A ints /No a Yes Constr
Policy Engine
114
57
DARPA neXt Generation Program: Concepts of Policy Agility Machine understandable policies will enable software downloads "on-the-fly"
XG Accomplishments
Collected And Analyzed RF Environment For Many Scenarios
Developed Low-Volume, High-Performance Sensor
Provides Needed Capability For Rapid Wideband Sensing Next Phase To Explore Integration With JTRS C-1
Policy Language And Radio Interface Defined
Used As Basis For Phase 2 Design Evaluations
Policy Language RFC V1 Composed And Released Extensible To Future “Cognitive” Technology
Three Feasible Designs For Interference Avoidance, Network Operation, And Rendezvous
Demonstrated Feasibility And Performance Of Adaptive Spectrum Technologies In Midst Of Phase 3 Source Selection
Will Select At Least One Design For 2-year Prototype Development And Demonstration Effort
Figure drawn from XG Vision RFC 115
116
58
XG Sensor
XG Sensor Focuses on Capabilities and Features Needed for JTRS C-1 Transition
Significantly
smaller footprint (more than 3X volume reduction)
XG – Phase 2 Significant Findings
Understanding of Temporal Characteristics Is Necessary Need to Detect Below Noise Floor
Interference Avoidance Policies Specific to Detected Signal
Difference in Detecting Known vs. Unknown Signals in Noise Affects How Aggressively XG Can Access Spectrum
Allocation Tables Provide A Priori Knowledge of Expected Signal Types, Especially Fixed and Broadcast
Continuous
Only 1 filter for 30 MHz – 1 GHz
XG Necessary for Which
LowPerformance power devices reduced power XGSensor Sensor Performance Necessary forXG XGImplementation Implementation WhichIs Is Not Yet Available for Military Communications Not Yet Available for Military Communications to 1 W average 117
Degree of A Priori Knowledge of Signals Provides Significant Performance Enhancement
RF card is 2X2 inches
frequency coverage 30 MHz – 2.5 GHz (vs. 6 bands)
All Signals are Not Created Equal
Policy Reasoning Necessary for Range of Incumbent Signal Protection
Commercial Services Are Sensitive to Effects of Interference at Many Levels, Including Reception Quality, BER, and Increase in Transmitter Power
Military Signals Are Inherently Hardened and Tolerant of Interference
Agile Systems Can Even Move If Interference Occurs 118
59
Phase 3 Development and Demonstration Activities
Build XG Technologies in Prototype Radio
Integrate The Radio, Adaptation Algorithms, Sensor Components, Policy-based Controls, And Radio Software into SCA Traceable Prototype
Continue Developing Key Policy Control Technologies Conduct Early Incremental Field Demos
Build Confidence in XG Capabilities Though A Series of Demos
Implement Networks Of Spectrum-agile Radios Which Dynamically Adapt To Changing Spectrum Environments 10x More Spectrum Without Interference To Non-XG Radios Demonstrate And Validate The XG Prototype’s Capabilities In Representative Military And Urban RF Environments.
E2R
Increase capability and environmental complexity at each demo
Transition to Military Program of Record In FY07
119
60
E2R Participants 1/2
E2R Research in Europe
Academic Partners Eurecom: Institut Eurecom I2R KCL:Centre for Telecommunications Research (CTR) - King's College London UoA: University of Athens TUD: Dresden University UoKarlsruhe: University of Karlsruhe, Communications Engineering Lab UPRC: University of Piraeus Research Center UNIS: University of Surrey
E2R = End-to-End Reconfigurability
Efficient,
advanced & flexible end-user service
provision
Tailoring of application and service provision to user preferences and profile
Efficient
spectrum, radio and equipment resources utilization
Enabling technologies for flexible spectrum resources
Multi-standard
platforms
Operator R&D Partners DoCoMo: DoCoMo Communications Laboratories Europe GmbH FT: France Telecom R&D TILAB: Telecom Italia S.p.A. TID: Telefonica I+D
A single hardware platform shared dynamically amongst multiple applications
121
Source http://e2r.motlabs.com/
122
61
E2R Participants 2/2 Manufacturer Partners MOTO: Motorola Labs ACP: Advanced Circuit Pursuit AG ASEL: Alcatel SEL DICE: Danube Integrated Circuit Engineering Nokia: Nokia GmbH PMDL: Panasonic UK PEL: Panasonic European Laboratories GmbH SM: Siemens Germany SMC: Siemens Mobile Communications SpA THC: Thales Communications TRL: Toshiba Research Europe Limited MIL: Motorola Israel Ltd Regulator partners DiGITIP UPC: UPC RegTP
Berkeley Wireless Research Center
123
62
Berkeley Wireless Research Center • • •
Designing a cognitive radio to improve spectrum utilization Radio searches for feasible region and optimal waveform for transmission (environment sensing) Avoiding of Interference with primary spectrum users by: - Measuring spectrum usage in time, frequency, and space - Having statistical traffic models of primary spetrum users
•
Rutgers Winlab
A cognitive radio test bed is currently being built
• The six system functions are split between physical and data link layer • Two control channels: - UCC for group management (group announcement) - GCC used only by members of a certain group
•From R.W. Brodersen, A. Wolisz, D. Cabric, S. M. Mishra, D. Willkomm "Corvus: A Cognitive Radio Aproach For Usage of Virtual Unlicensed Spectrum", July 29th 2004 125
63
WINLAB Rutgers University Benefits Design of info-stations for emergency and disaster relief applications • Use of customized commercially available hardware, e.g. 802.11 wireless •
Increases the total information available for rescue workers Tailors the information with regard to specific needs and available bandwidth Coordinates communication of different rescue groups at one site
Virginia Tech’s CWT
From: http://www.winlab.rutgers.edu/pub/docs/focus/Infostations.html
127
64
National Science Foundation Grant CNS-0519959 “An Enabling Technology for Wireless Networks – the VT Cognitive Engine” Develop and test a prototype system for using cognitive techniques to allow WiFi-like unlicensed operation in unoccupied TV channels.
Virginia Tech’s MPRG
Investigate the behavior of networks containing both legacy radios and cognitive radios that can interoperate with them.
National Institute of Justice Grant 2005-IJ-CX-K017 “A Prototype Public Safety Cognitive Radio for Universal Interoperability.” Build a prototype cognitive radio that can recognize and interoperate with three commonly used and mutually incompatible public safety waveform standards
129
65
Some SDR and Cognitive Radio Research at VT
SCA core framework
Open source effort
Role of DSPs
Power Management
Integration of testing into the framework
Rapid prototyping tools Smart antennas
Smart antenna API
Networking performance
Experimental MIMO systems
Cooperative radios
Distributed MIMO
Distributed Applications Cognitive radio networks
Game theory analysis of cognitive networks
Learning Techniques Test Beds
UWB SDR
Low Power SCA
Distributed PCs
Public Safety Radio Demo
131
CR Test-bed under development MWOL
Cordless Phone
Bluetooth TV station AP (Data Collection Node)
REM online updating
Observations
Distributed Measurement Collaborative Processing
Ethernet AP (Data Collection Node)
Neighbor WLANs
AP (Data Collection Node)
Interference Detection, Classification, Location
Arbitrary Waveform Generator
Analysis and decision Actions
OSSIE Framework
Tektronix TDS694C: Digital Real-time Oscilloscope Tektronix RSA3408A: RealTime Spectrum Analyzer 132
66
Public Safety - Interoperability
The Future of Cognitive Radio
Focus on multi-agency interoperability since 9/11/2001 Cognitive radio technology can improve interoperability by enabling devices to bridge communications between jurisdictions using different frequencies and modulation formats. Such interoperability is crucial to enabling public safety agencies to do their jobs. Example: National Public Safety Telecommunications Council (NPSTC) supported by U.S. DOJ’s AGILE Program
134
67
IEEE 802.22
IEEE Project 1900 (P1900)
WRAN system based on 802.22 will make use of unused TV broadcast channels Interoperable air interface for use in spectrum allocated to TV Broadcast Service Allows Point to Multi-point Wireless Regional Area Networks (WRANS) Supports a wide range of services
Data,
voice and video small and medium enterprises
Small office/home office (SOHO) locations
The IEEE P1900 Standards Group was established in 1Q 2005 jointly by the IEEE Communications Society (ComSoc) ComSoc) and the IEEE Electromagnetic Compatibility (EMC) Society. The objective of this effort is to develop supporting standards related to new technologies and techniques being developed for next generation radio and advanced spectrum management.
Residential,
135
136
68
IEEE P1900.1 Working Group:
IEEE P1900.2 Working Group:
Objective document: “Standard Terms, Definitions and Concepts for Spectrum Management, Policy Defined Radio, Adaptive Radio and Software Defined Radio.” Radio.” Purpose: This document will facilitate the development of these technologies by clarifying the terminology and how these technologies relate to each other. 137
Objective document: “Recommended Practice for the Analysis of InIn-Band and Adjacent Band Interference and Coexistence Between Radio Systems.” Systems.”
Purpose: This standard will provide guidance for the analysis of coexistence and interference between various radio services.
138
69
IEEE 802.11h
IEEE P1900.3 Working Group:
Objective document: “Recommended Practice for Conformance Evaluation of Software Defined Radio (SDR) Software Modules.” Modules.” Purpose: This recommended practice will provide guidance for validity analysis of proposed SDR terminal software prior to physical programming and activation of SDR terminal components.
139
802.11h helps WLANs share spectrum How?
801.11h
implements two methods to help spectrum sharing: Dynamic Frequency Selection (DFS) Transmission Power Control (TPC)
DFS is used to select the appropriate spectrum for WLAN
TPC is used to manage WLAN networks and stations for reduction of interference, range control (setting borders for WLAN), and reduction of power consumption (e.g., beneficial in laptop use). 140
70
IEEE 802.15.3a
Multiband OFDM for Personal Area Network
Wireless
Hurdles in CR
USB2.0 (480Mbps) at 5 meters distances
Cognitive Radio - Plausible Application to UWB Regulation
Very
fast spectrum sculpting via OFDM technology with wide bandwidth 528MHz
FCC Development Policies
Software Flexibility
Real-life Functionality
QoS Support
QoS
can be supported by controlling the number of sub-carriers
141
The process and rules governing how frequencies and waveforms are selected and approved for use by cognitive equipment must be addressed.
Interface with policy updates CR devices are smart enough to understand user request and surrounding environments
Network Availability for CR
Network needs to announce their availability to CR
Flexible or Reconfigurable Hardware Requires a language and protocols for initial interfacing with software and validation for existing devices as policies change across time and space Software Architectures
More dynamic than SCA
142
71
Predictions for Future Evolution Adaptive spectrum allocation
SDR with high ASIC content Factory reprogrammable
Limited reconfiguration by user
Early cognition Reprogrammable for fixed number of systems 2005
Just Remember This...
Increased use of reconfigurable hardware
“The best way to predict the future is to invent it.”
Cognitive radios
Alan Kay, Author Mid-level cognition
2007 Time
2010 143
144
72
Jeffrey H. Reed
Jeffrey H. Reed
Willis G. Worcester Professor of ECE and Deputy Director, Mobile and Portable Radio Research Group (MPRG) Authored book, Software Radio: A Modern Approach to Radio Engineering IEEE Fellow for Software Radio, Communications Signal Processing and Education Industry Achievement Award from the SDR Forum Highly published. Co-authored – 2 books, edited – 7 books. Previous and Ongoing SDR projects from
Contact Information:
[email protected] Electrical and Computer Engineering MPRG 432 Durham Hall Blacksburg, VA 24061 (540) 231-2972
DARPA, Texas Instruments, ONR, Mercury, Samsung, NSF, General Dynamics and Tektronix 145
146
73
Charles W. Bostian
Charles W. Bostian
Alumni Distinguished Professor of ECE and Director, Center for Wireless Telecommunications Co-author of John Wiley texts Solid State Radio Engineering and Satellite Communications. IEEE Fellow for contributions to and leadership in the understanding of satellite path radio wave propagation. Award winning teacher Previous and Ongoing CR projects from National Science Foundation, National Institute of Justice 147
Contact Information:
[email protected] Electrical and Computer Engineering Virginia Tech, Mail Code 0111 Blacksburg, VA 24061 (540) 231-5096
148
74
Backup Slides Hardware Blocks
Software Modules
149
150
75
Example: Simple AM Transmitter (1/2)
Example: Simple AM Transmitter (2/2)
Building Blocks •All Blocks are each defined as objects
Connecting Building Blocks Amp
m X
“Amp” - Gain Stage
~
“LO” - Local Oscillator
“m” - Message Signal “mix” - Multiplication Stage
FIR
“FIR” - Filter Stage
•The arrow is an object that connects the flow graph
H/W Interface
Amp
FIR
X
+1
µ
m
~ 151
152
76
Multi-Objective Optimization
Example SDR: GNU Radio
What is GNU Radio?
GNU
Radio is a set of S/W signal processing building blocks that allow users to create their own S/W radio
Why GNU Radio?
Attempts
to solve the complexity issues of both H/W and S/W of SDR
Modular (use with most any GPP)
S/W used on Windows, Linux, Mac
Multiple knobs are adjusted to tune multiple meters Complex problem to satisfy objectives like:
Requires advanced algorithms for optimization and learning. Evolutionary Algorithms offer significant benefits for this problem
153
Bit error rate Data rate Bandwidth Latency Power Battery life Many more
Stochastic search strategies Flexible and powerful 154
77
Spectrum Policy Language Design Actors and Roles Spectrum Opportunities Language Design Knowledge
design
Policy Language Designer (e.g. BBN/XG Program)
encode publish
Policy Administrator (e.g. FCC, NTIA)
Core Language Model and Representation
Complement static spectrum allocation with "Opportunistic spectrum access"
Primary
users Licensed Priority to use allocated spectrum Guaranteed QoS
Secondary users Non-licensed Can allocate unused spectrum among themselves Have to vacate bands if required by primaries
Policy Editing and Verification Tools
Area that needs improvements! Spectrum Policy
DARPA neXt Generation Program: Motivation – Proposed Solution
Machine Readable Policy Instances
query
XG System
Policy Repository
Awareness via XG Protocols and Sensing
Source: BBN Technologies Solutions LLC
155
Unless otherwise stated, all the information in this description of the DARPA XG program is based on the XG Vision rfc, available online: http://www.darpa.mil/ato/programs/xg/
156
78
DARPA neXt Generation Program: Concepts of Policy Agility (2)
DARPA neXt Generation Program: Promises
Decoupling policies, behaviors, and protocols: Separating what needs to be done from how it is implemented
1.
Flexible radio operation due to spectrum agility Simplified user control of XG systems
2.
System operation can be controlled in terms of behavior No need for technological details
Facilitated policy design
3.
Constraints can be tailored to national or institutional needs in terms of behaviors No need for technological details
Eased wireless device accreditation
4.
Traceability provides a means for an easy testing procedure of behaviors against policies
Broad and future proof standard
5.
Will be designed to be applicable to a broad range of radios Future proof design will enable extension of the standard Framework character: different technological solutions (protocols) can be accomodated to perform a particular task (sensing, identification, allocation)
The framework's four key components Figure drawn from XG Vision RFC
157
158
79
