Software-defined white-space cognitive systems: implementation of the spectrum sensing unit
2 likes891 views
The document discusses the implementation of a spectrum sensing unit for software-defined white-space cognitive systems, focusing on the detection of TV white-spaces and regulatory requirements for secondary user (SU) spectrum sensing. It details the IEEE 802.22 spectrum sensing model, autocorrelation techniques for DVB-T signals, and performance metrics including detection times. The authors conclude that their real-time module meets sensitivity requirements and outline plans for future enhancements to continuous monitoring of TV white-spaces.
Software-defined white-space cognitive systems: implementation of the spectrum sensing unit
- 1. Software-defined white-space cognitive systems: Implementation of the spectrum sensing unit Castelldefels, October 6th 2011 Sergio Benco, Floriana Crespi, Andrea Ghittino, Alberto Perotti Integrated Networks Laboratory (INLAB), CSP s.c.a r.l. - ICT innovation, TURIN (ITALY)
- 2. Outline TV White-Spaces Spectrum Sensing IEEE 802.22 Spectrum Sensing model DVB-T CP autocorrelation Spectrum Sensing Threshold calculation Performance Conclusions and future work Software-defined white-space cognitive systems 2
- 3. Spectrum sensing for TV White-Spaces TV White-Spaces represent the area (space domain) and the portion of the spectrum (VHF and UHF bands) where the broadcast signal strength falls below the sensitivity level of Primary User (PU) receivers Regulatory bodies are currently discussing about Secondary User (SU) spectrum sensing requirements in order to avoid interference to DVB-T receivers Interference issues can be faced through: SU geo-location and PU database queries Cognitive Pilot Channel (CPC) SU autonomous sensing (cooperative or not) Software-defined white-space cognitive systems 3
- 4. IEEE 802.22 Spectrum Sensing model PU protection contour (Dkm) Sensitivity range of the PU Rx PU ITU-R: PRPU = -92dBm @ 132km (RX) ERP TX = +90dBm, height: 500m, 615MHz) SU PU (TX) Keep-out region (Rkm) Range at wich the Desired/Undesired SU (D/U) ratio falls below 23 dB SU spectrum sensing requirements: ● PU Rx characteristics: F/B = 14 dB; D/U = 23 dB ● At PU Rx: PRSU ≤ PRPU – D/UdB + F/BdB ● At PU Rx: PRSU ≤ -101 dBm At SU Rx: Sens. ≤ -115 dBm A SU must detect a PU Tx at a range of: Rkm + Dkm Software-defined white-space cognitive systems 4
- 5. DVB-T spectrum sensing: CP autocorrelation Ns Symbol samples NCP CP samples N0CP N0 d N1CP N1d Nd Data samples K Number of symbols Ns K −1 i+kN s +N cp −1 CP correlator: See R xx (i) = ∑ ∑ x (n) x (n+ N d ) ˙ k=0 n=i+kN s References (1)(2) CP correlator test: DVB-T sensing module parameters max∣R xx (i)∣ Modes 8k (6817 subcarriers) T CP = i ⩾ γ 2k (1705 subcarriers) Avg∣R xx (i)∣ < CP lengths 1/4, 1/8, 1/16, 1/32 i∈J Channel bandwidth 8 MHz Sampling rate 12.5 MS/s (12.5 MHz) Software-defined white-space cognitive systems 5
- 6. DVB-T spectrum sensing: applied threshold False Alarm max∣R xx (i)∣ Probability (PFA ) i ̂ θ ⩾ γ T CP = = obtained through Avg∣R xx (i)∣ ∣R xx∣ < ̄ i∈J Monte-Carlo simulations over 1000 trials γ = γ ⋅ Avg∣R xx (i)∣ ⇒ ̂ P FA ⩽ 0.1 i ∈J J =N ∖Q N ={n∈ℕ : 0 ⩽ n < N s } ̂ ̂ Q={q∈ℕ : θ− N CP ⩽ q < θ+N CP } The threshold is adaptive w.r.t. the actual average correlation level plus a fixed margin that depends on PFA Software-defined white-space cognitive systems 6
- 7. DVB-T spectrum sensing over K symbols Symbol synchronization permits to obtain a coherent combining and average over K subsequent DVB-T symbols thus achieving a processing gain of about 5 dB for each 10 dB increase in K 1 symbol 10 symbols 100 symbols SNR = -15dB SNR = -15dB SNR = -15dB AWGN channel AWGN channel AWGN channel Software-defined white-space cognitive systems 7
- 8. DVB-T OFDM sensing: performance The detection time Tdet of this real-time module is calculated at the target sensing performance (PFA=0.1, PD=0.9) for a given SNR: SNR (PD=0.9) = -17 dB Symbols = 100 Tdet = 112.00 ms + Tproc SNR (PD=0.9) = -12 dB Symbols = 10 Tdet = 11.20 ms + Tproc Tch move time = 2000 ms Tsensing = Tch move time – 2Tdet Software-defined white-space cognitive systems 8
- 9. Conclusions and future work ● The DVB-T spectrum sensing based on CP autocorrelation offers a good trade-off between complexity and effectiveness ● The first attempts to exploit TV white space have raised the problem of high sensitivity requirements for the SU spectrum sensing unit ● We have developed a real-time module for OFDM spectrum sensing that approaches the requirements for the IEEE 802.22 WRAN spectrum sensing unit ● Future work will provide a SU network able to continuously monitor the TV White-Spaces through a CP-based spectrum sensing module using the GNURadio/USRP2 platform Software-defined white-space cognitive systems 9
- 10. References (1) D. Danev, E. Axell, and E. G. Larsson, “Spectrum Sensing Methods for Detection of DVB-T Signals in AWGN and Fading Channels”, In Proc. IEEE PIMRC, pp. 2721-2726, Dec. 2010 (2) V. Gaddam, M. Ghosh, “Robust Sensing of DVB-T Signals”, New Frontiers in Dynamic Spectrum, 2010 IEEE Symposium on , vol., no., pp.1-8, 6-9 April 2010 (3) S. Shellhammer, V. Tawil, G. Chouinard, M. Muterspaugh, M. Ghosh, “Spectrum sensing simulation model”, IEEE 802.22-06/0028r10, Sept. 2006 (4) C.R. Stevenson, C. Cordeiro, E. Sofer, G. Chouinard, “Functional Requirements for the 802.22 WRAN Standard”, IEEE 802.22- 05/0007r46, September 2005 Software-defined white-space cognitive systems 10
- 11. Contacts Sergio Benco Consulting Engineer, Integrated Networks Laboratory (INLAB) R&D dept. mail: sergio.benco@csp.it cell: +39 329 0118356 tel. +39 011-4815164 CSP innovation in ICT Registered and Central Offices Environment Park - Laboratori A1 via Livorno 60 - 10144 Torino Operational Offices Villa Gualino - Viale Settimio Severo 63 10133 Torino Tel +39 011 4815111 Fax +39 011 4815001 E-mail: marketing@csp.it www.csp.it Software-defined white-space cognitive systems 11