November 2007 doc IEEE 802 11 072793 r

November 2007 doc. : IEEE 802. 11 -07/2793 r 0 60 GHz vs. 5 GHz Propagation Discussion Date: 2007 -11 -12 Authors: Submission 1 Vinko Erceg, Broadcom

November 2007 doc. : IEEE 802. 11 -07/2793 r 0 LOS Path Loss at 60 GHz • For LOS conditions path loss is mainly governed by the free-space propagation formula. Some papers have reported few d. B less path loss and some papers few d. B more than the free-space formula. For distances up to 10 m, for LOS conditions, it seems like a reasonable assumption (Friis formula): • At 1 m distance path loss is 68 d. B and at 10 m distance the path loss is 88 d. B. • Standard deviation of approximately 5 d. B was reported in the literature that accounts for shadowing. A margin due to shadowing has to be added if 90% coverage is to be determined, for example. The margin is a function of the path loss exponent and standard deviation. Submission 2 Vinko Erceg, Broadcom

November 2007 doc. : IEEE 802. 11 -07/2793 r 0 NLOS Path Loss at 60 GHz • Very few NLOS measurements were made at 60 GHz. • With increasing carrier frequency losses due to diffraction, reflection, and material penetration increase for NLOS conditions. • When compared to 5 GHz, we estimate that these additional losses are in the 5 -10 d. B range [1, 2]. • Standard deviation of shadowing at 60 GHz NLOS propagation is larger than at 5 GHz. Submission 3 Vinko Erceg, Broadcom

November 2007 doc. : IEEE 802. 11 -07/2793 r 0 Total Expected NLOS Path Loss at 60 GHz Compared to 5 GHz • Free space loss is additional 21 d. B. • Let’s assume that additional diffraction, penetration and reflection losses are 7 d. B. • The total additional path loss expected at 60 GHz when compared to the 5 GHz carrier frequency may be about 28 d. B. • Assuming NLOS distance path loss exponent of 4, this translates into communication range (distance) reduction by a multiplicative factor of 0. 2, i. e: Range(60 GHz) = 0. 2 Range(5 GHz). • The calculation above does not include shadowing standard deviation, it includes only median path loss. Submission 4 Vinko Erceg, Broadcom

November 2007 doc. : IEEE 802. 11 -07/2793 r 0 60 GHz NLOS Range Reduction Range factor 10 d. B Loss 15 d. B 20 d. B 25 d. B 30 d. B 0. 56 0. 42 0. 31 0. 24 0. 18 • 60 GHz compared to 5 GHz, path loss exponent = 4 (NLOS) Expected additional loss (includes path loss difference, antenna gain and PA transmit power estimates) Submission 5 Vinko Erceg, Broadcom

November 2007 doc. : IEEE 802. 11 -07/2793 r 0 60 GHz NLOS Range Reduction: Cont’d Range factor 10 d. B Loss 15 d. B 20 d. B 25 d. B 30 d. B 0. 46 0. 32 0. 22 0. 15 0. 1 • 60 GHz compared to 5 GHz, path loss exponent = 3 (NLOS) Expected additional loss (includes path loss difference, antenna gain and PA transmit power estimates) Submission 6 Vinko Erceg, Broadcom

November 2007 doc. : IEEE 802. 11 -07/2793 r 0 60 GHz LOS Range Reduction Range factor 10 d. B Loss 15 d. B 20 d. B 25 d. B 30 d. B 0. 32 0. 18 0. 1 0. 05 0. 03 • 60 GHz compared to 5 GHz, path loss exponent = 2 (LOS) Expected additional loss (includes path loss, antenna gain and PA transmit power estimates) Submission 7 Vinko Erceg, Broadcom

November 2007 doc. : IEEE 802. 11 -07/2793 r 0 PA Power • Given current PA technology developments, regulations, and safety reasons it may be expected that the consumer grade PAs at 60 GHz will have up to 10 d. B less power that the ones for the 5 GHz WLAN applications, for example (10 m. W vs. 100 m. W). • Some of the lower PA power can be compensated for by antenna gain. Submission 8 Vinko Erceg, Broadcom

November 2007 doc. : IEEE 802. 11 -07/2793 r 0 Antenna Gain at 60 GHz – Caution • High directivity antennas in a scattering environment lose gain when compared to the omnidirectional antennas. In the literature this is regarded as antenna GRF (gain reduction factor). • Directive antennas amplify only some propagation rays, others are attenuated and therefore multipath components are lost that results in a loss of power, loss of diversity, and reduced rank for MIMO systems that reduces spatial multiplexing gains. • Omnidirectional antennas amplify all rays evenly, in most cases. • Therefore, in path loss calculations, nominal (advertised) antenna gain determined in an anechoic chamber has to be reduced by GRF. • It was shown that for outdoor environments the GRF is significant (5 -7 d. B for 20 o horizontal antenna beamwidth (-3 d. B beamwidth points)). • Indoor environments experience more scattering than outdoor environments potentially resulting in larger GRF. Requires further investigation. Submission 9 Vinko Erceg, Broadcom

November 2007 doc. : IEEE 802. 11 -07/2793 r 0 LOS Delay Profiles for Different Antenna Beamwidths at 60 GHz 10 d. B 20 d. B 30 d. B Tx Omni Antenna Rx Omni Antenna Tx Omni Antenna Rx 60 o Antenna Tx Omni Antenna Rx 10 o Antenna With decreasing antenna beamwidth LOS component is amplified and reflected components are attenuated by the antenna pattern [3]. Submission 10 Vinko Erceg, Broadcom

November 2007 doc. : IEEE 802. 11 -07/2793 r 0 Blockage Effects • If the LOS component is blocked by a person, for example, it was reported in the literature that up to 20 d. B of signal power can be lost [4, 5]. • This effect can be illustrated by removing LOS component from the delay profiles on the previous slide. • The loss is more pronounced for narrow beamwidth antennas, where LOS component is more dominant. • For the omnidirectional antennas there is some empirical evidence that the loss is considerably smaller, up to 6 d. B [6]. This result seems too optimistic, should be verified by additional measurements (authors to emulate blockage have removed the strongest LOS path from the delay profile measurements, may not be accurate representation of blockage). Submission 11 Vinko Erceg, Broadcom

November 2007 doc. : IEEE 802. 11 -07/2793 r 0 Blockage Effects: Cont’d • To avoid large signal power loss in the case of the direct LOS path blockage, especially in the directional antenna case, adaptive beam antennas may be preferable. • It is not clear how the adaptive beam antennas would perform when compared to the omnidirectional antennas in the case of a blockage. • If the high reliability link is to be expected, the blockage scenario has to be carefully considered and solution found in order not to break the communication link. • Closely spaced multiple transceivers may mitigate shadowing problem, but at the higher cost and complexity. Submission 12 Vinko Erceg, Broadcom

November 2007 doc. : IEEE 802. 11 -07/2793 r 0 Discussion To get better handle on 60 GHz propagation the following important measurements should be performed that are missing in the reported and published results: • NLOS measurements (very few were reported). • GRF - omni vs. directional antenna measurements. • Diversity, MIMO measurements - as a next step. Submission 13 Vinko Erceg, Broadcom

November 2007 doc. : IEEE 802. 11 -07/2793 r 0 Discussion/Conclusion • 60 GHz frequency band may be suitable for applications limited to short range communication, mesh, or distributed (densely deployed) networks. • 60 GHz frequency band may not be suitable for applications that require similar range requirements and applications as 802. 11 a/b/g/n systems. • To cover a regular house at 2. 4 GHz or 5 GHz frequency bands one WLAN Access Point may be required, while for the same radio coverage at 60 GHz 9 units may be required (based on simple assumption that Range(60 GHz) = 1/3 Range(5 GHz)). • Additional measurements and analysis should be done at 60 GHz frequency band. Submission 14 Vinko Erceg, Broadcom

November 2007 doc. : IEEE 802. 11 -07/2793 r 0 References • [1] P. F. M. Smulders and L. MI. Correia “Characterisation of Propagation in 60 GHz Radio Channels, ” Electronics and Comm. Eng. Journal, April 1997, pp. 73 -80. • [2] H. Yang, P. F. M. Smulders and M. H. A. J. Herben “Indoor Channel Measurements and Analysis in the Frequency Bands 2 GHz and 60 GHz, ” 2005 IEEE PIMRC, pp. 579 -583. • [3] Takeshi Manabe, Yuko Miura, and Toshio Iharw “Effects of Antenna Directivity on Indoor Multipath Propagation Characteristics at 60 GHz, ” in Proceedings of IEEE PIMRC, Toronto, 1995, pp. 1035 -1039. • [4] K. Sato and T. Manabe “Estimation of Propagation-Path Visibility for Indoor Wireless LAN Systems under Shadowing Condition by Human Bodies, ” Vehicular Technology Conference, Vol. 3, 18 -21 May 1998, pp. 2109 – 2113. • [5] S. Collonge, G. Zaharia, and G. El Zein “Influence of the Human Activity on Wide-Band Characteristics of the 60 GHz Indoor Radio Channel, ” IEEE Trans. on Wireless Comm. , Vol. 3, No. 6, Nov 2004. • [6] T. Zwick, T. J. Beukema, and H. Nam “Wideband Channel Sounder With Measurements and Model for the 60 GHz Indoor Radio Channel, ” IEEE Trans on Veh Technol, Vol. 54, No. 4, July 2005. Submission 15 Vinko Erceg, Broadcom