1. www.syrres.com IWCE 2005 1 Title Modern Spectrum Management Techniques for the Public Safety Professional VenueIWCE 2005 DateApril, 7 2005 National Public Safety Telecommunications Council Syracuse Research Corporation

2. www.syrres.com IWCE 2005 2 This is the third year in a row that “Modern Techniques” has been presented. We are going to build upon our past sessions –More complex every year –This year less tutorial, more “What is Possible” This year: –Shortspacing to 700 MHz television: making the case for mobile and portable operations –4.9 GHz modeling techniques, with interference and performance prediction –Packing and mining – the ultimate techniques…. –Some “teasing” A quick note: Thanks to NYS and NPSTC for supporting much of the research behind the development of these techniques! Overview

3. www.syrres.com IWCE 2005 3 PART I: Shortspacing to 700 MHz Television: Making the Case for Mobile and Portable Operations

4. www.syrres.com IWCE 2005 4 Previous Studies have indicated that 700 MHz public safety channels might be available nearly everywhere –NYS, NPSTC, Guard Band Managers, etc –But would require a significant waiver process A roadmap to achieve these waivers was presented at the NPSTC meetings last Fall –Focus was on the New York City area –Overview of entire process with details on base station to TV interference NPSTC is developing guidelines that define all of this –Don’t get overwhelmed, there will be a how-to guide available from NPSTC documenting all this (and more) by mid-2005 Overview

5. www.syrres.com IWCE 2005 5 These are complex waivers –Although some precedent exists, some concepts that will be utilized are “new” –The scale and location of these waivers can be massive (especially in areas with many TV stations) FCC Waivers

6. www.syrres.com IWCE 2005 6 Stations Pertinent to 64/69 Operational Pairing

7. www.syrres.com IWCE 2005 7 Stations Pertinent to 64/69 Operational Pairing – Primary Interference Sources to TV (2) Interference to TV reception by adj-channel PS towers within or near the grade B (3) Interference to TV reception by adj-channel PS mobile/portable units within or near the grade B (1) Interference to TV reception by co-channel PS towers near the grade B

8. www.syrres.com IWCE 2005 8 Interference to Television, Case (3) Interference to TV reception by adjacent channel PS mobile and portable units within or near the grade B For a 64/69 pairing, need to ensure that WFUT TV-68 maintains a set D/U ratio within its service area. –Interferers are roaming mobile and portable units –No precedent on this, need to sell the FCC on this, balanced against needs of PS

9. www.syrres.com IWCE 2005 9 Interference to Television, Case (3) For these cases we will need to compute the size of the interference “holes” around roaming public safety units –Interference range to TV reception from active mobile/portable units We will assume free space propagation from the mobile units, with 25 W ERP at 1.8 m AGL We will assume free space propagation from the portable units, with 0.32 W ERP at 1.6 m AGL

10. www.syrres.com IWCE 2005 10 x Adj-Channel TV Station Grade-B FIGURES ARE NOT TO SCALE - EXAGERATED DIMENSIONS x Public Safety Mobile or Portable Unit d: “Hole Size” “d” is the free space loss distance required to reduce the Mobile/Portable ERP level to the TV signal level minus the required D/U protection level. It is a function of D (the desired TV signal level) and a variable protection level D/U (itself a function of frequency offset from the TV signal). Illustration of the ‘Hole” Concept

11. www.syrres.com IWCE 2005 11 Mobile Interference Effects Clearly, mobile interference is a statistical phenomena –Random process with temporal, spatial, and frequency dependence Best way to capture its effects is to: –(1) Simulate the mobility of the operations –(2) Integrate the mobile interference effects against population distribution –(3) Perform Monte Carlo trials to determine the characteristics of the interference distributions

12. www.syrres.com IWCE 2005 12 Mobile Interference Effects: Process Set Mobile Distribution Uniform or Correlated –Roads or Area Density –County or Region –Number of Units ERP –Mobile or Portable Frequency Offset Compute U Compute D-U Free Space Loss Mobile ERP Worst case, min FSPL to (x,y) Compute D Longley Rice 3-sec grid Compute population grid 3-sec From FCC data (DTV program) Compare to Required D-U Stanks look-up Function of offset,  f Function of D Compute Effects %Grade B j = j=j+1 j=1 DO UNTIL j=J Preprocessing

13. www.syrres.com IWCE 2005 13 What is the D/U? DATA SOURCE: Receiver Susceptibility Measurements Relating to Interference Between UHF Television and Land Mobile Radio Services, FCC/OET TM 87-1, April 1986. The D/U ranges from -2 dB to – 43 dB, depending upon the desired signal strength (D), and the offset from the adjacent channel TV signal. The average is as follows: Public Safety 64/69 Pairing Adjacent Channel TV (63 and 68) Desired to Undesired Thresholds

14. www.syrres.com IWCE 2005 14 Mobile Interference Effects – Receive Power Levels Receive Power (dBm) at 9.1 m AGL WFUT Grade B

15. www.syrres.com IWCE 2005 15 D/U Requirement at  f = 9 MHz, From Stanks Tables dB Derived From: FCC/OET TM87-1, Receiver Susceptibility Measurements Relating to Interference Between UHF Television and Land Mobile radio Services project No. EEB-84-4, April 1986

16. www.syrres.com IWCE 2005 16 Example: Interference from 10 Units/County NOTE: This free space interference is very conservative. A two-slope power law would reflect reality – and reduce the interference significantly

17. www.syrres.com IWCE 2005 17 Example: D/U from 10 Units/County D/U (dB)

18. www.syrres.com IWCE 2005 18 In this case: 0.06 % of Grade B Population Affected Example: Grade B “Holes” from 10 Units/County WFUT Grade B

19. www.syrres.com IWCE 2005 19 Convergence of Initial Results

20. www.syrres.com IWCE 2005 20 11 Total 110 Total 22 Total 55 Total In this case, the trend indicates that even in the worst case conditions, there would need to be over 350 simultaneously active mobile units per county, or over 3,800 total simultaneously active mobile units in operation before even 1% of the Grade B population is affected Trends: % Grade B Population Affected

21. www.syrres.com IWCE 2005 21 PART II: 4.9 GHz Modeling Techniques: Interference and Performance Prediction

22. www.syrres.com IWCE 2005 22 4.9 GHz Broadband Characteristics 4.9 GHz broadband coverage characteristics differ greatly from narrowband voice below 1 GHz –Typically Time Division Duplex (TDD) –Coverage is short range –Reflections and multipath both help and hinder performance –Packet data with lost data retransmitted, not continuous voice stream.

23. www.syrres.com IWCE 2005 23 4.9 GHz Interference Scattering and Multipath

24. www.syrres.com IWCE 2005 24 4.9 GHz Interference Coexistence of deployments can be complicated – even when the systems are shared! Scattering and Multipath Scattering and Multipath

25. www.syrres.com IWCE 2005 25 Performance Scenario Generated by SRC/NSO for NPSTC Purpose was to show that COTS emmisions masks could support PS capacity requirements at an “extreme” incident scene, with no on-scene spectrum management required Multifaceted Simulation: –Complex incident, in a worst case environment –Extremely detailed coverage models –Service-specific traffic profiles –Use of 802.11-type PHY –Use of mesh routing, with node retransmission overheads –Accounting of MAC overhead –Accounting of both RF and packet collisions –Retransmission of lost packets See the complete story on-line at NPSTC’s web site: –Under FCC Docket 00-32 (August 19, 2004) –http://npstc.org/documents/fccpositions.htmlhttp://npstc.org/documents/fccpositions.html

26. www.syrres.com IWCE 2005 26 Incident Area Chemical Storage Tanks Car Bombing Secondary Remote Detonated Explosive

27. www.syrres.com IWCE 2005 27 Police Fire EMS Mobile Command Centers Explosion Scene Bomb Squad Robot and Control Triage Area

28. www.syrres.com IWCE 2005 28 Police Fire EMS Bomb Squad POLICE Incident Routing and Information Flow

29. www.syrres.com IWCE 2005 29 Received Level (dBm)

30. www.syrres.com IWCE 2005 30 Received PHY Data Rate (Mbps) Packets

31. www.syrres.com IWCE 2005 31 Simulation Results

32. www.syrres.com IWCE 2005 32 Mobile Command Access Points Mesh Routing Points Reverse Link Video From Robot

33. www.syrres.com IWCE 2005 33 PART III: Packing and Mining: The Ultimate Techniques….

34. www.syrres.com IWCE 2005 34 “Frequency Mining” involves finding frequency resources at defined site locations –Defined process, based upon FCC rules and/or coordination guidelines –Repetitive and time consuming –Focus here is on automated processes “Frequency Packing” involves assigning frequencies system wide (or region wide) in some type of optimal fashion –System design process –Extremely complex optimization problem –Very time consuming –Focus here is on automated processes Mining and Packing

35. www.syrres.com IWCE 2005 35 The processes themselves are quite simple, however an effective automated implementation can be very complex Data Needs: –Sites and site capacity targets, based upon system design –FCC license data –RPC allotment data Coverage Model for MLS, Reliability, and Reliability Degradation –In this example, Longley Rice with ITS 3-sec terrain and no LULC losses Defined Reliability Parameters –Ex. CPC f = 18 dB, IF ENBW = 12.5 kHz,  ln = 8 dB, NF = 10 dB Upon examination of results, a decision can be made at to which spectrum plan to use –Ultimately a trade off between coverage and capacity Mining and Assignment Process

36. www.syrres.com IWCE 2005 36 Co/Adj Interference External Co/Adj Interference Internal Possible Assignments Use matrix “maps” in the optimization process Soft or hard constraints are applied (arrays can be “fuzzy” logic) Assignments Made General Assignment Methods Frequency

37. www.syrres.com IWCE 2005 37 Pure Tile based Optimization (OUCH!) Depth corresponds to # of sites Height and width correspond to tile analyses Generating the input arrays is a large computational task! Integrating through the depth gives MLS and S/(  I+N)

38. www.syrres.com IWCE 2005 38 Mining and Packing Process Reuse Step Over 1,200 fully linked html pages with integrated text and graphics FCC Data SWN Data DHAAT Results. 800 PS Pool Input Data Assignment Maps (Computed) All Available Freqs. Available Freqs. SWN Capacity Needs ASSIGN, NxF Binary Matrix Assignments NxF Matrix C, Nx1 Vector SWN_Data, Nx1 Structure FCC_Data, Mx1 Structure FREQS, Fx1 Vector Selections (Computed and Optimized) Documentation Reliability Degradation CO,NxN“Fuzzy” Logic Matrix FCC Data Site Data DHAAT Results. 800 PS Pool Input Data Assignment Maps (Computed) All Available Freqs. Available Freqs. Capacity Needs ASSIGN, NxF Binary Matrix Assignments NxF Matrix C, Nx1 Vector Site_Data, Nx1 Structure FCC_Data, Mx1 Structure FREQS, Fx1 Vector Selections (Computed and Optimized) Documentation Reliability Degradation CO,NxN“Fuzzy” Logic Matrix

39. www.syrres.com IWCE 2005 39 Reliability Degradation? Q: What the heck is “Reliability Degradation”? A: It is a a reality-based measure of actual interference effects It is based upon TSB-88 concepts –Communications reliability –Tile based interference assessment –Equivalent interferer combination –Technology to technology ACCPR effects –Protection afforded only where service area exists, not over an entire IMAGINARY contour –Design to  S/(I+N), not simple contour intersections Maximizes reuse, while offering accurate interference assessments As an aside; think about this prior to rebanding. –Baseline your current reliability –Compare your post rebanding reliability before accepting the assignments. –Frequency coordinators do not go to this level of detail

40. www.syrres.com IWCE 2005 40 Two Close Sites, 16-km Separation Example Interference Determination

41. www.syrres.com IWCE 2005 41 Example Interference Determination Power Levels, Site 1

42. www.syrres.com IWCE 2005 42 Example Interference Determination Power Levels (40 and 5 dBu), Site 1

43. www.syrres.com IWCE 2005 43 Power Levels, Site 2 Example Interference Determination

44. www.syrres.com IWCE 2005 44 Example Interference Determination Power Levels (40 and 5 dBu), Site 1

45. www.syrres.com IWCE 2005 45 Example Interference Determination MLS Areas

46. www.syrres.com IWCE 2005 46 Example Interference Determination MLS Power

47. www.syrres.com IWCE 2005 47 Example Interference Determination Reliability Degradation Site 1 to Site 2

48. www.syrres.com IWCE 2005 48 Example Interference Determination Reliability Degradation Site 2 to Site 1

49. www.syrres.com IWCE 2005 49 Balancing Capacity with Coverage

50. www.syrres.com IWCE 2005 50 Spectrum Engineering Example - Processing Scale of Inputs –>1,000 Sites –> 120,000 incumbent licensed site/channels –244 paired channel possibilities Three unique spectrum plans corresponding to Region 2, Region 7, and the Non-Border Region Mining and Packing Computations –A > 1,000 site MLS study –Approximately 200,000 individual tile based multi-site reliability studies –Examination of over 121 MILLION distance and DHAAT combinations Spectrum Plan Output Documentation –Auto generation of over 1,200 fully linked html pages with integrated text and high resolution graphics

51. www.syrres.com IWCE 2005 51 The bottom line is time and money This automated process is: –Much more accurate than typical techniques –Much more efficient than typical techniques See below, and compare 2-3 weeks of labor (automated) to as much as 64 MAN-YEARS of labor to achieve the same thing (manual) –This allows us to handle new and larger problems –This is how repacking before rebanding can be achieved Time Savings

52. www.syrres.com IWCE 2005 52 Contact For Further Information Sean O’Hara Business Area Manager Analysis, Communications and Collection Systems Syracuse Research Corporation, Systems Technology Center Vice Chair NPSTC Technology Committee Co-Chair NPSTC Broadband Working Group Co-Chair NPSTC Border Issues Working Group Vie-Chair Software Defined Radio Forum Spectrum Efficiency and Cognitive Radio Working Group (315) 452-8152 ohara@syrres.com