doc.: IEEE /1229r1 Submission November 2009 Alexander Maltsev, IntelSlide 1 Application of 60 GHz Channel Models for Comparison of TGad Proposals Date: Authors:

doc.: IEEE /1229r1 Submission November 2009 Alexander Maltsev, IntelSlide 2 Abstract This contribution presents an approach for evaluation of TGad proposals using 60 GHz WLAN channel models.

doc.: IEEE /1229r1 Submission November 2009 Alexander Maltsev, IntelSlide 3 Introduction The evaluation methodology (EVM) document of TGad [1] has to define methods to calculate various PHY and MAC characteristics of a TGad proposal so that different proposals can be fairly compared. Calculation of these characteristics should involve usage of 60 GHz channel models like, for example, [2]. Some properties of 60 GHz WLAN systems (and channel models that may be applied for their evaluation) are different from legacy WLAN systems operating in 2.4 and 5 GHz bands. The main difference is that 60 GHz WLAN systems require to use high-gain steerable directional antennas. Hence, modifications in the evaluation process for 60 GHz WLAN systems are required in comparison with legacy WLAN systems. This presentation makes a proposal for part of the EVM that should address the process of calculating candidate system characteristics using 60 GHz WLAN channel models.

doc.: IEEE /1229r1 Submission Typical Characteristics Used for Proposals Comparison Typical criteria for comparison of different proposals in a WLAN system (see, for example, [3]) include a set of PHY and MAC characteristics. PHY characteristics are usually PER vs. SNR curves for different operation modes and different modulation and coding schemes of the system. MAC characteristics are usually system throughput (or “goodput”) for different network configurations and traffic models. Calculation of the MAC throughput characteristics requires knowledge of PHY PER vs. SNR characteristics and, in addition, a path loss model for the scenario of interest November 2009 Alexander Maltsev, IntelSlide 4

doc.: IEEE /1229r1 Submission 60 GHz Channel Model The channel model proposed in [2] can be used for calculation of all types of PHY and MAC system metrics for comparison of different TGad proposals. Channel model [2] provides complex amplitudes of different rays taking into account the attenuation of the signal between the transmitter and receiver along the rays in a real scale. Hence, each ray has a part of information about the propagation loss of the channel and a part of information about the impulse response. Note that this approach is different from the traditional channel modeling approach in the WLAN bands of 2.4 and 5 GHz where separate models are generated for path loss and channel impulse response. In order to calculate PER vs. SNR curves and develop a path loss model, decoupling of the channel impulse response and path loss characteristics of the general channel model is required. November 2009 Alexander Maltsev, IntelSlide 5

doc.: IEEE /1229r1 Submission Decoupling CIR and Path Loss Characteristics Channel impulse response (CIR) and path loss characteristics may be decoupled in the 60GHz channel model if the antenna models and beamforming algorithms are fixed. There are many options for antenna and beamforming algorithms to be used in 60 GHz WLAN systems. Parameters of the antennas and beamforming algorithms used will significantly influence evaluation results for different proposals. For a fair comparison of proposals, it is suggested to use two sets of antennas and beamforming algorithms parameters: –First set of standard parameters defined by the EVM, that shall be used for all proposals. –Second set of parameters specific for a given proposal, that may define different types of antenna technologies and beamforming algorithms. The first set of parameters has to be defined by the EVM [1] and a path loss model for the standard antennas and beamforming parameters has to be developed in the channel modeling document [2]. November 2009 Alexander Maltsev, IntelSlide 6

doc.: IEEE /1229r1 Submission Standard Set of Antennas and Beamforming Parameters Many existing 60 GHz specification include both omni and directional PHY modes to help solving the directional MAC problem. To address this issue the standard antenna models are proposed to include: 1.Isotropic radiator (as defined in [2]) 2.Basic steerable antenna model (as defined in [2]) with 30 0 antenna beamwidth To steer directional antennas, algorithm adjusting TX and RX antenna patterns towards the most powerful channel ray is suggested as the standard beamforming algorithm. The description of the algorithm is to be provided in [2]. Hence, evaluation of TGad proposals using the standard antennas and beamforming algorithm should include calculation of PER vs. SNR curves for: 1.Omni-to-omni communications 2.Omni-to-directional communications; 3.Directional-to-directional communications. November 2009 Alexander Maltsev, IntelSlide 7

doc.: IEEE /1229r1 Submission Normalization of CIR Characteristics Running link level simulations requires the channel impulse response (CIR) to be normalized (in average or instantaneous senses). Instantaneous normalization of the CIRs obtained after the beamforming application is proposed to be used for PHY simulations. The instantaneous normalization approach will remove flat fading effects of the channel model –This would be approximated by beam searching procedure that reestablished link in case of deep fade November 2009 Alexander Maltsev, IntelSlide 8

doc.: IEEE /1229r1 Submission Summary of PHY Characteristics Evaluation for TGad Proposals If the suggested approach adopted then PHY evaluation of a TGad proposal should include the following results: 1.PER vs. SNR curves for AWGN channel (reference curves) 2.PER vs. SNR curves for omni-to-omni communications in LOS scenario 3.PER vs. SNR curves for omni-to-directional communications in NLOS scenario 4.PER vs. SNR curves for directional-to-directional communications in NLOS scenario In addition, impact of different PHY impairments on PER vs. SNR curves has to be estimated November 2009 Alexander Maltsev, IntelSlide 9

doc.: IEEE /1229r1 Submission Path Loss Model Development Once the antenna model and beamforming algorithms are fixed, a path loss model may be derived. A path loss model for directional-to-directional communications with the basic steerable antenna model and maximum power ray beamforming algorithm is derived in [4] and is included in the channel model document [2]. The same approach as in [4] is planned to develop omni-to-omni and omni-to- directional path loss models and include them in [2] to have all required path loss models for the standard antenna models and beamforming algorithm. Also a similar approach may be used to derive a path loss model for beamforming algorithms and antenna models of TGad proposals. November 2009 Alexander Maltsev, IntelSlide 10

doc.: IEEE /1229r1 Submission Example of TX and RX Antenna Patterns Adjustment Toward the Most Powerful Ray for Directional-to- directional and Omni-to-directional cases directional-to-directional communication omni-to-directional communication Slide 11 November 2009

doc.: IEEE /1229r1 Submission Average Path Loss vs. Distance for NLOS Scenario Slide 12 November 2009

doc.: IEEE /1229r1 Submission Path Loss Model Summary for Directional-to-Directional Communications [4] November 2009 Alexander Maltsev, IntelSlide 13 If antenna system parameters and beamforming algorithms are fixed then it is possible to derive an average path loss model using the standard form: Here A and n are parameters specific for the scenario and antenna system, f is the carrier frequency in GHz, R is the distance between TX and RX in m. The normal in dB SF model may be used together with the average path loss model. SF standard deviation is specific for the scenario and antenna system parameters. The path loss model parameters for the conference room scenario, basic steerable antenna model and maximum ray beamforming algorithm are: ScenarioA, dBnSF std. dev., dB Conference room LOS Conference room NLOS

doc.: IEEE /1229r1 Submission Conclusions Channel impulse response (CIR) and path loss characteristics is decoupled in the 60GHz channel model for EVM. This contribution suggests to evaluate every TGad proposal for the standard set of antenna models and beamforming algorithms, which are common for all proposals, and beamforming algorithms and antenna models, which are specific for individual TGad proposals. The suggested approach will allow a fair comparison of TGad proposals. The standard (common for all proposals) antenna models are suggested: 1.The isotropic radiator [2], for omni-directional antennas. 2.The basic steerable antenna model from [2], with 30 0 antenna beamwidth. The algorithm steering antennas towards the direction of the most powerful ray is suggested as the standard beamforming algorithm. Path loss models for the standard approach have been developed for directional-to-directional communications in [4] and will be developed for other cases (omni-to-omni and omni-to-directional) and included in [2]. November 2009 Alexander Maltsev, IntelSlide 14

doc.: IEEE /1229r1 Submission November 2009 Alexander Maltsev, IntelSlide 15 References 1.IEEE doc /0096r8. TGad evaluation methodology, E. Perahia, July, IEEE doc /0334r3. Channel models for 60 GHz WLAN systems, A. Maltsev et al, July, IEEE doc /814r31. IEEE TGn Comparison Criteria, Adrian P Stephens, July, IEEE doc /0553r1 Path Loss Model Development for TGad Channel Models, A. Maltsev et al, May, 2009.