Spectrum Management and Cognitive Radio

Spectrum Management and Cognitive Radio Alessandro Guidotti Tutor: Prof. Giovanni Emanuele Corazza, University of Bologna, DEIS Co-Tutor: Ing. Guido Riva, Fondazione Ugo Bordoni

The spectrum scarcity problem The ever increasing spectrum demand for new wireless communication systems cannot be satisfied anymore, thus leading to an overall spectrum scarcity perception The Command & Control static spectrum allocation method is completely ineffective Possible solutions: Releasing frequency bands thanks to novel and more efficient transmission techniques (e.g. DVB-T) Allowing smart devices to dynamically access licensed bands

Research activity DVB-T Mobile & DVB-T 470 MHz 790 MHz 862 MHz The digital switch-over process is releasing some very valuable bands that can be reallocated to other services WRC-07 allocated the 790-862 MHz band to mobile services starting from June 2015, and allowed EU countries to utilize this allocation before 2015 subject to technical coordination with neighbouring countries New interference issues: Co-channel interference between neighbouring countries/regions, one of them using this band for mobile service and the other one for broadcast television Adjacent channel interference within a given territory

Research activity DVB-T Mobile & DVB-T 470 MHz 790 MHz 862 MHz White Space: part of the spectrum available for radiocommunication application at a given time in a given geographical area on a non-interfering / non-protected basis with regard to other services with a higher priority [CEPT] CR: a radio in which communications systems are aware of their environment and internal state and can make decisions about their operating behaviour based on that information and predefined objectives [IEEE] Several technological challenges must be overcome Spectrum Sensing Spectrum Management Spectrum Mobility

Research activity DVB-T Mobile & DVB-T 470 MHz 790 MHz 862 MHz CEPT SE43 Technical and operational requirements for the possible operation of cognitive radio systems in the White Space of the frequency band 470-790 MHz Hidden Node Problem Amount of spectrum potentially available as White Space Geolocation Database Collaboration with CREAT-NET, NET&SERV Dept., Trento: Cooperative and distributed energy detection Cognitive Satellite Terrestrial Radios (54-862 MHz & 3.1-10.6 GHz bands)

Hidden Node Problem Path1: rooftop mounted DTT receiver in LOS condition with respect to the DTT TX Path2: CR attempting to detect the DTT signal at street level, shielded by the surrounding buildings Path3: the CR could erroneously conclude that the specific DTT channel is available and hence it might cause harmful interference to the DTT receiver path3 path2 path1 Hidden Node Margin: difference between filed strengths received at the location of the CR and the DTT receiving antenna Methodology 1: Ray Tracing simulations performed in several environments and validated against measurements Methodology 2: analytical study based on Recommendation ITU-R P.1546 HNM (50,50) ( DTT ) ( CR) E med (50,50) Emed sha G

Hidden Node Problem HNM for DTT Service Protection Urban environment 4km 2 area with 1765 test points +24.7=44.46 db Simulation Parameters M1: Frequency = 690 MHz Bandwidth = 8 MHz ERP TV = 81.11 dbm Distance = 3.4 4.7 km UK Ofcom results Densely Urban 90% of test points 95% of test points 99% of test points 18.5 db 22.4 db 29.2 db Urban 28.1 db 30.2 db 32.5 db Rural 14.9 db 15.6 db 16.6 db SE43 single-device sensing threshold range: [-91,-165] dbm M.Barbiroli, C.Carciofi, A.Guidotti, D.Guiducci, Evaluation and analysis of the Hidden Node Margin for Cognitive Radio System operation in a real scenario, submitted to EuCAP2011 Italian Contribution to SE43, Calculation of the Hidden Node Margin in a real Italian Scenario, August 2010

Cooperative Energy Detection The energy detector performs the detection of the signal by comparing the energy of the received signal with a predefined threshold s(t) + X x(t) i Pre-filter A/D (1/N)*Σ N 2 > < E H 0 H1 n (t) h(t) Single-device energy detector is extremely simple from a computational point of view, but the performance are highly susceptible to noise and fast/slow fading Possible solution: let several devices cooperate among them in order to increase the detection performance

Cooperative Energy Detection Cooperative spectrum sensing with p- out-of-l rule: if at least p CRs among the L cooperating devices detect the primary signal, then the fusion center concludes that the channel is occupied Simulations are performed on AWGN, Rayleigh and Lognormal channels Fusion Center DTT Transmitter Performance evaluation: Prob{H0 H0} vs Prob{H1 H1}

Probability of Correct Primary User Detection Cooperative Energy Detection 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 Theoretical ROC [L=6 - SNR=-3 db - N=2] p=1 p=2 p=3 p=4 p=5 p=6 Single-Device W.Zhang (ICC 2008) computed the optimal value of p When Pd and Pf have the same order, p opt L/2 When Pf<<Pd, p opt 1 (OR rule) When Pf>>Pd, p opt L (AND rule) 0.1 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Probability of Correct Secondary User Authorization

Probability of Correct Primary User Detection Probability of Correct Primary User Detection Probability of Correct Primary User Detection Cooperative Energy Detection 1 1 0.9 0.9 0.8 0.8 0.7 0.7 0.6 0.6 0.5 0.5 0.4 0.4 0.3 0.3 0.2 0.2 0.1 ROC Theoretical on Lognormal ROC [L=6 Channel - SNR=-3 db [L=6 - N=2] - SNR=-3 db - N=2] p=1 p=2 p=3 p=4 p=5 p=6 Single-Device Single-Device p=1 p=3 p=6 0.1 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 Probability of Correct Secondary User Authorization 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Probability of Correct Secondary User Authorization 1 W.Zhang (ICC 2008) computed the 0.9 optimal 0.8 value of p 0.7 0.6 0.5 0.4 0.3 0.2 0.1 ROC on Rayleigh Channel [L=6 - SNR=-3 db - N=2] Single-Device p=1 p=3 p=6 When Pd and Pf have the same order, p opt L/2 When Pf<<Pd, p opt 1 (OR rule) When Pf>>Pd, p opt L (AND rule) 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Probability of Correct Secondary User Authorization There is a significant gain obtained thanks to cooperation among CRs With higher values of L the performance improve, but the gain is less evident for increasing L The problem is the traffic overhead due to the sensing report to the Fusion Center (higher for large values of L, of course) A possible solution could be an hybrid cooperative/distributed architecture (under study)

Cognitive Satellite Terrestrial Radios The future satellite ground terminals need to integrate and coexist with the spectrally crowded terrestrial wireless systems The cognitive functionalities are proposed for satellite uplink and terrestrial transmissions The satellite downlink cannot adopt such capabilities due to many practical limitations in the controlling satellite transmissions Application Satellite-802.22 for WRAN (54-862 MHz) Satellite-UWB for PANs (3.1-10.6 GHz) 3D-spatial domain frequency reuse: since the satellite is seen under a high elevation angle (depending on the latitude), proper design of the antenna radiation pattern can allow to reduce mutual interference

Cognitive Satellite Terrestrial Radios Simulations were performed with an energy detection sensing Due to the directional gain pattern of the parabolic antenna, the Terrestrial Transmitter detection performance becomes worse when reduces For the same reason, the detection performance becomes worse for increasing values of θ On the one hand, the TT detection becomes harder for increasing θ or decreasing, but on the other hand in such cases the directional antennas interfere less with the TT S.Kandeepan, L.De Nardis, M.-G. Di Benedetto, A.Guidotti, G.E.Corazza, Cognitive Satellite Terrestrial Radios, to appear in GLOBECOM2010

LTE / DVB-T interference The digital switch-over process is releasing some very valuable bands that can be reallocated to other services WRC-07 allocated the 790-862 MHz band to mobile services starting from June 2015, and allowed EU countries to utilize this allocation before 2015 subject to technical coordination with neighbouring countries We performed a study of the co-channel interference problem, proposing a methodology to take into account the mutual interference between a LTE mobile network and a DVB-T system Monte Carlo simulations (100 iterations) Aggregated interference from 42 LTE BSs towards 200 Yagi receiving antennas Interference from a DVB-T transmitter towards 42 LTE BSs (uplink) Performance evaluation DVB-T: an acceptable guard distance was defined as the distance between the two networks such that <1% of performance degradation is obtained LTE: uplink average throughput reduction A.Guidotti, D.Guiducci, M.Barbiroli, C.Carciofi, P.Grazioso, Coexistence and mutual interference between mobile and broadcasting systems, submitted to VTC Spring 2011

LTE / DVB-T interference Please note that Link Adaptation behaviour was taken into account The effect of the 42 BSs has been evaluated for several distances between the DVB- T coverage area edge and the mobile area center The guard distances are the distance such that the DVB-T system suffers less than 1% of coverage degradation Deploying the LTE system at the computed guard distance, the effect of the DVB-T transmitter on this network has thus been evaluated

LTE / DVB-T interference There might be situations where more interest is focused on the protection of the mobile service The proposed performance parameter to be evaluated is the reduction in the throughput provided to the users Having defined the maximum acceptable throughput reduction (e.g. 10%), the DVB-T system parameters that guarantee the protection of the mobile service can be easily identified

Further work The dependence of the Hidden Node Margin values on the geometrical properties of the scenario (e.g. street orientation and width) is currently under study The performance of the Cooperative Energy Detection is being evaluated in correlated lognormal channels, as well as the implementation of the hybrid cooperative/distributed architecture The analysis of adjacent channel interference between mobile and broadcasting services still has to be analyzed Analytical characterization of the interference originated by a CR network on a primary network will be analyzed using stochastic geometry (6 months visit at SUPELEC, Paris)

Bibliography M.Barbiroli, C.Carciofi, A.Guidotti, D.Guiducci, Evaluation and analysis of the Hidden Node Margin for Cognitive Radio System operation in a real scenario, submitted to EuCAP2011 Italian Contribution to SE43, Calculation of the Hidden Node Margin in a real Italian Scenario, August 2010 Italian Contribution to SE43, Estimate of the amount of spectrum available as White Space in north west of Italy, January 2010 S.Kandeepan, L.De Nardis, M.-G. Di Benedetto, A.Guidotti, G.E.Corazza, Cognitive Satellite Terrestrial Radios, to appear in GLOBECOM2010 A.Guidotti, D.Guiducci, M.Barbiroli, C.Carciofi, P.Grazioso, Coexistence and mutual interference between mobile and broadcasting systems, submitted to VTC Spring 2011 Italian Contribution to SE43, Translation of the information provided to the geolocation database into elements of authorization to the WSD, SE43, August 2010 Italian Contribution to CPG PT-D, WRC-12 AI 1.17 Broadcasting vs mobile: theoretical analysis of mutual interference

THANK YOU!