January 2006 doc IEEE 802 22 060007 r

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Primary User Protection in 802. 22 Proposals IEEE P 802. 22 Wireless RANs Date: 2006 -01 -16 Authors: Notice: This document has been prepared to assist IEEE 802. 22. It is offered as a basis for discussion and is not binding on the contributing individual(s) or organization(s). The material in this document is subject to change in form and content after further study. The contributor(s) reserve(s) the right to add, amend or withdraw material contained herein. Release: The contributor grants a free, irrevocable license to the IEEE to incorporate material contained in this contribution, and any modifications thereof, in the creation of an IEEE Standards publication; to copyright in the IEEE’s name any IEEE Standards publication even though it may include portions of this contribution; and at the IEEE’s sole discretion to permit others to reproduce in whole or in part the resulting IEEE Standards publication. The contributor also acknowledges and accepts that this contribution may be made public by IEEE 802. 22. Patent Policy and Procedures: The contributor is familiar with the IEEE 802 Patent Policy and Procedures http: //standards. ieee. org/guides/bylaws/sb-bylaws. pdf including the statement "IEEE standards may include the known use of patent(s), including patent applications, provided the IEEE receives assurance from the patent holder or applicant with respect to patents essential for compliance with both mandatory and optional portions of the standard. " Early disclosure to the Working Group of patent information that might be relevant to the standard is essential to reduce the possibility for delays in the development process and increase the likelihood that the draft publication will be approved for publication. Please notify the Chair Carl R. Stevenson as early as possible, in written or electronic form, if patented technology (or technology under patent application) might be incorporated into a draft standard being developed within the IEEE 802. 22 Working Group. If you have questions, contact the IEEE Patent Committee Administrator at patcom@iee. org. > Submission 1 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Purpose of the Presentation • In November 2005 ten proposals were submitted to IEEE 802. 22 • Many of these proposals include techniques for protecting primary users – Primary users in this band refers to DTV and Part 74 devices – This is often referred to as “vertical sharing” – Similarly, “horizontal sharing” is sharing between systems with equal access rights • This presentation summarizes these protection techniques • This presentation does not attempt to evaluate the proposals, just to summarize • This presentation does not address other aspects of the proposals Submission 2 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Proposal Documents Company Documents Communication Research Center (CRC), Canada 22 -05 -0057 r 0, 22 -05 -0063 r 0 ETRI, Samsung Electro-magnetics, and Georgia Tech 22 -05 -0108 r 0, 22 -05 -0109 r 0 Huawei Technologies 22 -05 -0106 r 0, 22 -05 -0107 r 1 Institute for Infocomm Research 22 -05 -0093 r 0, 22 -05 -0094 r 0 Nextwave 22 -05 -0091 r 0, 22 -05 -0092 r 0 Nanotron 22 -05 -0101 r 0, 22 -05 -0102 r 0 Phillips and France Telecom 22 -05 -0103 r 0, 22 -05 -0104 r 0, 2205 -0105 r 1 Samsung 22 -05 -0099 r 0, 22 -05 -0100 r 1 ST Microelectronics and Runcom 22 -05 -0097 r 0, 22 -05 -0098 r 0 Thomson 22 -05 -0095 r 0, 22 -05 -0096 r 0 Submission 3 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 CRC Canada • The protection mechanism in this proposal is the broadcasting of which TV channels are occupied over the ATSC-DTV forward link – This mechanism is assumed to based on knowledge of the geographic location and access to a database of TV channel utilization Submission 4 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 ETRI/Samsung/Georgia Tech • Two methods of processing the received signal are suggested for generating parameter that can be used to decide which TV channels are occupied – Coarse – Multi-resolution Spectrum Sensing – Fine – Analog Autocorrelation • It is suggested that all the processing be done in the analog domain. • This is an implementation choice and could equally be done in the digital domain Submission 5 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 ETRI/Samsung/Georgia Tech Directive Antenna SW or Duplexer Transmitter (RF/IF) PHY (Baseband) Receiver (RF/IF) Omni Antenna MAC Sensing Receiver Coarse “MRSS” Low Speed ADC Fine “AAC” Cognitive WRAN Architecture Submission 6 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Multi-resolution Spectrum Sensing • It is suggested that the coefficients of the Wavelet Transform are generated by correlating the received signal with the a given wavelet basis function in the analog domain • The coefficient is converted to digital format using a lowspeed ADC • The wavelet basis function is changed to calculate another wavelet transform coefficient • These wavelet transform coefficients are then used to decide if a channel is occupied Submission 7 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Multi-resolution Spectrum Sensing X x(t) Driver Amp z(t) ADC CLK#2 w(t) v(t)*f. LO(t) y(t) CLK#1 Timing Clock MAC Wavelet Generator Block Diagram of Multi-resolution Spectrum Sensing Submission 8 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Multi-resolution Spectrum Sensing 40 -50 20 -60 0 -70 -20 -80 PSD (d. B) Power Spectrum Magnitude (d. B) • Example of spectrum analysis using MRSS -40 -90 -60 -100 -80 -110 -100 0 0. 2 0. 4 0. 6 0. 8 1 1. 2 Frequency 1. 4 1. 6 1. 8 -120 2 x 10 6 The spectrum of the wireless microphone signal Submission 0 0. 2 0. 4 0. 6 0. 8 1 1. 2 Frequency (Hz) 1. 4 1. 6 1. 8 2 x 10 6 The corresponding signal spectrum detected with the MRSS technique 9 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Analog Autocorrelation • The autocorrelation of a random process is the expected value of the product of the signal and a delayed version of the signal • The proposal suggests calculating the autocorrelation function in the analog domain. • The delay value is varied and the process is repeated • The low-speed ADC converts the autocorrelation value to digital format • The proposal claims that this technique can be used for blind detection • The proposal states that this can also be used for signal classification Submission 10 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Analog Autocorrelation Sensing Antenna x(t) Multiplication Integrate Low Speed ADC FIR Delay Td x(t-Td)) Decision Making Block Diagram of Analog Autocorrelation Submission 11 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 ETRI/Samsung/Georgia Tech • The MAC maintains several lists – – – Set of TV channels used by CPE Set of TV channels used by base station Set of five candidate TV channels Set of occupied TV channels Set of disallowed TV channels • The MAC provides a quiet period for sensing the spectrum – The lists are update after the sensing period Submission 12 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 ETRI/Samsung/Georgia Tech • The downlink channel descriptor is modified to include these lists • The MAC includes a procedure for switching channels – There seems to be an option for only a subset of the CPEs to change channels • The CPEs report their sensing results to the BS in response to a request Submission 13 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Question for ETRI/Samsung/Georgia Tech • Are MRSS and AAC intended to operate in parallel or in series? • How long does it take to calculate the wavelet transform coefficients using this analog technique? • How are the wavelet transform coefficients from MRSS used to decide if a channel is occupied or not? • Do you have any simulation results showing probability of detection and false alarm rate using MRSS? • How is the autocorrelation function used to decide whether the channel is occupied? • How is AAC used for blind detection? Don’t you need some knowledge of the signal to fix your decision rule? • Do you have any simulation results showing probability of detection and false alarm rate using AAC? Submission 14 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Huawei Technologies • The proposal suggest dividing each TV channel into sub -channels • It would be possible to populate a subset of subchannels • This is intended to reduce interference caused to wireless microphones Submission 15 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Institute for Infocomm Research • Time is allocated after the downlink sub-frame and before the uplink sub-frame for spectrum sensing • The channel is divided into sub-channels which consist of 48 data sub-carriers • To avoid interference to wireless microphones some of these sub-channels may be nulled out – Since the sub-channels may consist of non-adjacent sub-carriers several methods are proposed for nulling out a contiguous set of sub-carriers Submission 16 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Institute for Infocomm Research • The proposal also includes an option to null out some sub-carriers for sensing while transmitting on the other sub-carriers Submission 17 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Questions for Infocomm • Have you performed any tests to see if nulling out a portion of the TV channel will effectively prevent interference with wireless microphones operating in that part of the TV channel? • Have you performed any simulations to demonstrate that you can effectively listen on a nulled sub-carrier while adjacent sub-carriers are populated? • What type of signals (e. g. DTV or wireless microphones) can be detected using this approach? Submission 18 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Nextwave • The proposal states that the MAC includes a number of procedures for managing sharing of the spectrum. The listed procedures are, – Testing channels for other users including specific spectrum users – Discontinuing operations after detecting other users including specific spectrum users – Detecting other users including specific spectrum users – Scheduling for channel testing – Requesting and reporting of measurements – Selecting and advertising a new channel • The proposal does not seem to include the details of these procedures Submission 19 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Questions for Nextwave • When can you supply the details of the MAC procedures listed in your proposal? Submission 20 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Nanotron • This proposal does not include any material on protection of primary users Submission 21 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Philips and France Telecom • PHY Summary – Energy Detector – Smoothed PSD estimator – ATSC detector based on correlation with PN 511 signal • MAC Summary – New Spectrum Management (SM) Layer – Support for quiet period – Reports on spectrum measurements Submission 22 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Philips and France Telecom • Energy Detector – Estimate received signal strength indicator (RSSI) – Estimate noise power – Detector compares RSSI and scaled version of the noise power estimate Submission 23 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Philips and France Telecom • Part 74 Feature Detector – Using the FFT from the OFDMA PHY calculate the periodogram as an estimate of the power spectral density – Then calculate the mean and variance of the PSD – The detector then compares the PSD with a linear function of the mean and variance – This allows detecting non-uniformity of the PSD Submission 24 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Philips and France Telecom • ATSC DTV Feature Detector – The received signal is correlated with the know PN 511 signal – The running mean and variance of the correlator output are calculated – The detector compares the variance with a scaled version of the mean Submission 25 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Philips and France Telecom • Cognitive MAC (CMAC) – Scheduled quiet periods for spectrum sensing – A CPE can perform in-band sensing during a quiet period – A CPE can perform out-of-band sensing either during a quiet period or outside of a quiet period – There is a mechanism for base stations to attempt to synchronize their quiet periods with overlapping WRAN cells, to improve spectrum sensing Submission 26 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Philips and France Telecom • Cognitive MAC Frames – MAC Frames sent by the base station to request the CPE performs spectrum sensing and reports results – Two types of measurement reports: regular and urgent – Regular reports send measurement results to base station on a regular interval – An urgent report is sent to indicate that a primary user has been detected by the CPE in the occupied TV band Submission 27 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Philips and France Telecom • Cognitive MAC Frames – If the CPE has allocated TX time then after a quiet period the CPE reports an abbreviated report in a measurement frame specifying which TV channels are occupied – Subsequently the BS can then request more detailed measurement reports – If the CPE does not have allocated TX time then there is a Urgent Coexistence Situation (UCS) time slot that can be used to transmit the urgent abbreviated report Submission 28 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Philips and France Telecom • Cognitive MAC Frames – The UCS notification slot can be accessed either using a contention-based protocol, like CSMA/CA or it can use a contention-based CDMA mechanism using a randomly selected code – If the BS does not receive a report from the CPE it may be due to interference from an in-band primary user – Under this condition the BS shall infer that the channel is occupied by a primary user and change to another channel Submission 29 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Philips and France Telecom • Cognitive MAC Frames – There are multiple messages for reporting spectrum sensing results • MAC Frames for reporting measurement statistics • MAC Frame for reporting only the results of the detection algorithm specifying which channels are occupied and which are not • Subfields in the MAC header to indicate a UCS and which channel in which it was detected – By maintaining a list of backup channels, at BS and CPEs, the proposal supports recovery from the detection of an in-band primary user Submission 30 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Questions for Philips and France Telecom • Do you have any simulation results for the three detectors suggested in this proposal? Submission 31 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Samsung • The proposal emphasizes the need for distributed sensing • The proposal correctly identifies that the distributed sensing information must be fused together at the base station – Data Fusion: Combining processed information to make centralized decision – Decision Fusion: Combine localized decisions into a centralized decision • The authors state that the data fusion and decision fusion architectures are fairly independent of the individual decision algorithms employed at the CPEs Submission 32 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Samsung • The authors identify that the WRAN is solving a distributed Hypothesis Testing problem. • The proposal suggest two types of detectors – Energy Detector – Cyclostationary Detector • The proposal assumes that the samples of the received signal are independent Submission 33 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Samsung • The authors point out that energy detection is a good approach when you do not have any a priori information about the signal • The authors point out that energy detection is optimum for detecting Gaussian signal in additive white Gaussian noise • For data fusion in the energy detector the proposal suggests summing the individual sensed energy values and using that sum as the decision statistic Submission 34 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Samsung • For decision fusion based the authors suggest summing the individual decisions from each of the CPEs an comparing that sum-of-decisions to a threshold to calculate the centralized decision Submission 35 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Samsung • The proposal gives some background on cyclostationary processes • The authors state that the cyclostationary detector does not have the problem with noise that the energy detector has, since detection can theoretically be done irrespective of the noise power level • Energy detection is a special case of cyclostationary detector, since the power spectral density is a subset of the spectral correlation density function Submission 36 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Samsung • One of the advantages of using the cyclostationary detector is the fact that it can be used for both detection and classification • Another advantage of the cyclostationary detector is that it is extensible to be able to handle new signals that may be introduced in the band in the future Submission 37 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Questions for Samsung • Why do you believe that the data fusion and decision fusion architectures are fairly independent of the individual decision algorithms employed at the CPEs? • How can you assume that the samples of the received signal are independent? • What practical issues arise due to detection of a signal corrupted by noise using a cyclostationary detector? Submission 38 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 ST Micro and Runcom • This proposal describes the trade-offs in allocating time for quiet periods for spectrum sensing – Spectrum sensing can be performed reliably without interference caused by WRAN transmissions – The sensing period will increase latency over the WRAN network – The sensing period will decrease WRAN network throughput Submission 39 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 ST Micro and Runcom • The proposal suggests three methods of scheduling spectrum sensing – Non-overlapping with data transmissions • Sensing is only performed during quiet periods – Full-overlap with data transmissions • Sensing can be performed during data transmissions – Selective-overlap with data transmission • Sensing can be performed on selective channels during data transmission Submission 40 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 ST Micro and Runcom • Selective-overlap sensing – If a channel has enough frequency separation from the utilized channel it is possible to perform reliable sensing on that channel during data transmission Submission 41 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 ST Micro and Runcom • The proposal suggests that the WRAN use dynamic frequency hopping (DFH) • After the sensing period the BS selects one of the available TV channels and network hops to that channel for a specified time. • How long the network dwells on each TV channel can be changed dynamically • By changing operating frequency the set of TV channels that cannot be sensed during data transmission changes, which improves spectrum sensing Submission 42 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 ST Micro and Runcom • Since two neighboring WRAN cells could potentially select the same channel the proposal introduces a method of avoiding this event – One cell waits until the adjacent cell selects it new operating frequency – It decodes the message from the adjacent BS – Then it selects a different operating frequency • This method required adjacent BS to be able to decode messages from each other Submission 43 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 ST Micro and Runcom Submission 44 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 ST Micro and Runcom • Sensing reports can be in one of two forms – Bit vector reports specifying one of four states of the channel: • Occupied by a licensed system • Occupied by another 802. 22 system • Noisy • Vacant – Raw data reports transmit the raw data • The proposal describes several methods for scheduling the reporting – Poll-me – Contention Submission 45 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Questions for ST Micro and Runcom • How many TV channels need to be available for use so that dynamic frequency hopping can be used effectively? • What does the WRAN network do if there is insufficient available TV channels? Submission 46 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Thomson • The proposal describes dynamic frequency selection (DFS) as required in 802. 22 • The proposal recommends utilizing quiet periods for sensing • The proposal recommends techniques for detecting ATSC DTV signals – Sensing Method 1 – Sensing Method 2 Submission 47 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Thomson • Sensing Method 1 – – First step is to detect the pilot using filtering and PLL Using pilot down convert to baseband Use a matched filter on the PN 511 sequence As an alternative use a matched filter on the PN 63 sequence – The proposal also lists several other signal features that can be used to detect the ATSC DTV signal Submission 48 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Thomson • Sensing Method 2 – This method does not require extracting the pilot – A complex autocorrelation is performed – The detector identifies peaks in the autocorrelation signal – The autocorrelation signal may be smoothed using an IIR digital filter Submission 49 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Thomson • The proposal introduces a novel method for improving the estimate of timing an frequency by utilizing the received signal from a stronger DTV station • This method relies on the tight frequency tolerances in DTV – 1 KHz Submission 50 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Thomson • The proposal includes recommendations on the design of the omnidirectional sensing antenna – Concern is raised about the difficulty in designing an omni antenna that operates over the entire UHF band without spatial nulls – The proposal suggests using two antennas and alternating sensing. This eliminates the nulls at the cost of twice the time required for sensing • The proposal also includes a recommendation on the design of a device that can be used to detect if the antenna is not properly aligned Submission 51 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Questions for Thomson • Do you have any simulation results for the two sensing methods you proposed? Submission 52 Steve Shellhammer, Qualcomm

January 2006 doc. : IEEE 802. 22 -06/0007 r 1 Summary • This proposal attempts to summarized the portion of each proposal that addresses protection of primary users • There is a follow-up presentation on Cognitive Radio Framework for Protection of Primary Users Submission 53 Steve Shellhammer, Qualcomm