doc.: IEEE 802.15-<doc#> <month year> doc.: IEEE 802.15-<doc#> May 2007 Project: IEEE P802.15 Working Group for Wireless Personal Area Networks (WPANs) Submission Title: Adapting the IEEE 802.15.3 MAC to WPAN systems employing directional antennas Date Submitted: 7 May 07 Source: Khusro Saleem, Bryan Beresford-Smith (NICTA Ltd) Address Building 193, University of Melbourne Melbourne, VIC, 3010, Australia Voice: +61 3 8344 8407 FAX: none, E-Mail: khusro.saleem@nicta.com.au Re: Adapting the IEEE 802.15.3 MAC to WPAN systems employing directional antennas Purpose: This is a discussion only document. No action is recommended. Notice: This document has been prepared to assist the IEEE P802.15. It is offered as a basis for discussion and is not binding on the contributing individual(s) or organization(s). The material in this document is subject to change in form and content after further study. The contributor(s) reserve(s) the right to add, amend or withdraw material contained herein. Release: The contributor acknowledges and accepts that this contribution becomes the property of IEEE and may be made publicly available by P802.15. NICTA Ltd <author>, <company>

Khusro Saleem, Bryan Beresford-Smith NICTA Ltd May 2007 Adapting the IEEE 802.15.3 MAC to WPAN systems employing directional antennas Khusro Saleem, Bryan Beresford-Smith NICTA Ltd NICTA Ltd

doc.: IEEE 802.15-<doc#> <month year> doc.: IEEE 802.15-<doc#> Outline May 2007 Overview Benefits of directional antennas Challenges at the MAC layer Proposed amendments to the IEEE 802.15.3 MAC layer Conclusions and further work NICTA Ltd <author>, <company>

doc.: IEEE 802.15-<doc#> <month year> doc.: IEEE 802.15-<doc#> Overview May 2007 IEEE 802.15.3 TG3c is developing a mm-wave-based alternative PHY for the existing WPAN standard IEEE 802.15.3-2003 Signal propagation at mm-wave frequencies has high path loss Signals also experience greater attenuation through most materials High TX power not expected to overcome attenuation Directional antennas are a cost-effective solution to this problem How do we adapt IEEE 802.15.3-2003 to work effectively with directional antennas? NICTA Ltd <author>, <company>

Directional antennas - benefits <month year> doc.: IEEE 802.15-<doc#> Directional antennas - benefits May 2007 Antenna elements are small in the mm-wave spectrum and can be integrated into the package This in-turn allows for beam-forming thereby increased antenna gain and spatial re-use A side benefit is increased resilience to multi-path fading and interference NICTA Ltd <author>, <company>

Challenges at the MAC layer <month year> doc.: IEEE 802.15-<doc#> Challenges at the MAC layer May 2007 Neighbor discovery PNC-DEV communication DEV-DEV communication Hidden node problems Exposed node problems Spatial reuse NICTA Ltd <author>, <company>

doc.: IEEE 802.15-<doc#> <month year> doc.: IEEE 802.15-<doc#> Adapting IEEE 802.15.3 May 2007 The following proposals enhance the current standard whilst ensuring Interoperability with DEVs that only employ omni-directional antennas. NICTA Ltd <author>, <company>

Adapting IEEE 802.15.3 – Main change <month year> doc.: IEEE 802.15-<doc#> Adapting IEEE 802.15.3 – Main change May 2007 MAC beacons are transmitted by the PNC omni-directionally To overcome the resulting antenna gain asymmetry in the PNC-DEV links, the MAC beacons are transmitted at a lower rate The lower rate transmission rate results in a lower SNR requirement Select the lower rate, such that the lower SNR requirement overcomes for the antenna gain asymmetry in the PNC-DEV links We should end up with the same probability of correct frame reception We can readily adapt this to any PHY specification with the addition of a few PHY-SAP primitives This represents a form of cross-layer optimization NICTA Ltd <author>, <company>

Adapting IEEE 802.15.3 – Main change <month year> doc.: IEEE 802.15-<doc#> Adapting IEEE 802.15.3 – Main change May 2007 For example, consider the current 2.4GHz PHY specification in IEEE 802.15.3-2003 For 64-QAM-TCM, to achieve FER≤8%, we need SNR≥20dB Assume we transmit MAC beacon as per the MAC header at 22Mb/s using DQPSK To achieve FER≤8%, SNR≥13dB ΔSNR=7dB Loosely speaking, we can lose about ~7dB in antenna gain Yes, there will be a reduction in throughput; but the beacon is a small fraction of the superframe anyway NICTA Ltd <author>, <company>

Adapting IEEE 802.15.3 – Main change <month year> doc.: IEEE 802.15-<doc#> Adapting IEEE 802.15.3 – Main change May 2007 How do we address Neighbor Discovery? The PNC MAC beacon is transmitted omni-directionally and DEVs scan using directional beams By transmitting the PNC MAC beacon omni-directionally, neighbor discovery is a trivial process And by lowering the data rate we overcome the asymmetric antenna gain This approach is interoperable with omni-directional DEVs as well The beacon must be lengthened to facilitate Angle-Of-Arrival (AOA) estimation NICTA Ltd <author>, <company>

Adapting IEEE 802.15.3 – PNC-DEV communication <month year> doc.: IEEE 802.15-<doc#> Adapting IEEE 802.15.3 – PNC-DEV communication May 2007 So what does the PNC-DEV link look like? The PNC is omni and the DEV is directional or omni (depending on capability) So when the directional-capable DEVs are expecting the beacon they point their beams at the PNC The beam pointing direction (AOA) is learned each time a PNC beacon is received NICTA Ltd <author>, <company>

Adapting IEEE 802.15.3 – DEV-DEV communication <month year> doc.: IEEE 802.15-<doc#> Adapting IEEE 802.15.3 – DEV-DEV communication May 2007 So what does the DEV-DEV link look like? The DEV-DEV link could be omni-omni, omni-directional, or directional-directional depending on device capability The main question is how to two DEVs employ a directional-directional link? This requires a training sequence for AOA estimation, so lets reuse our PNC-DEV link communication model again NICTA Ltd <author>, <company>

Adapting IEEE 802.15.3 – DEV-DEV communication <month year> doc.: IEEE 802.15-<doc#> Adapting IEEE 802.15.3 – DEV-DEV communication May 2007 How do two DEVs establish a directional-directional link? After successful stream creation, the two DEVs undergo an AOA estimation sequence during the first allocated CTAP Assume DEV1 initiated the stream request DEV1 transmits an DEV-beacon omni-directionally, just like the PNC beacon, at a lower rate DEV2 scans directionally and receives this and estimates the AOA DEV2 then acknowledges this DEV-beacon directionally which is then received by DEV1 and used to estimate its AOA The training sequence is complete DEVs use this training phase to ascertain if there is sufficient gain in the link to meet the stream throughput and latency requirements. NICTA Ltd <author>, <company>

Adapting IEEE 802.15.3 – DEV-DEV communication <month year> doc.: IEEE 802.15-<doc#> Adapting IEEE 802.15.3 – DEV-DEV communication May 2007 PNC DEV1 DEV2 Create stream procedure Omni beacon AOA estimation Directional special-ACK AOA estimation Directional-directional link NICTA Ltd <author>, <company>

Adapting IEEE 802.15.3 – Communicating in the CAP <month year> doc.: IEEE 802.15-<doc#> Adapting IEEE 802.15.3 – Communicating in the CAP May 2007 How do we communicate during the CAP? This is where we should exploit spatial reuse Borrow ideas from [1], employ Directional RTS/CTS Directional NAV (DNAV) AOA caching Beam lock and unlocking Combat hidden node problem NICTA Ltd <author>, <company>

Conclusions and further work <month year> doc.: IEEE 802.15-<doc#> Conclusions and further work May 2007 A simple set of amendments to the existing IEEE 802.15.3-2003 standard are proposed Ensuring compatibility with omni-directional DEVs Many outstanding issues Impact on throughput Modulation format (PHY dependent) Details training sequence for DEV-DEV AOA training phase Need simulation results for overall MAC behaviour with these modifications NICTA Ltd <author>, <company>