Reconfigurable antennas for WiFi networks Daniele Piazza Founder and CTO Adant Technologies Inc
Company Overview Adant Padova, Italy Adant SF Bay Area Adant Taiwan Adant designs, licenses, and manufactures reconfigurable (smart) antenna systems for the wireless communications industry 2
Adant main markets WIRELESS NETWORKING MOBILE INTERNET HOME GATEWAY RFID Wireless ISP 3
Outline q Reconfigurable antenna system for WiFi devices q Advantages and challenges of using reconfigurable antennas in commercial WiFi devices l Network capacity maximization l Interference mitigation l Enhanced capacity in high density environments l Coordination and benefit with digital beamforming l Coordination and benefit with MU-MIMO q Conclusions 4
Reconfigurable antenna system Active RF devices shape the antenna beams and control their direction PCB metamaterial antenna integrates easily into any device SW algorithms select the optimal beam shape for maximum reliability of the wireless link 5
MIMO systems adaptive antennas The best wireless communication channel between TX/RX antennas is carefully selected among the large set of channels that can be generated by changing the antenna radiation patterns STANDARD MIMO SYSTEM Standard static antenna MIMO SYSTEM WITH RECONFIGURABLE ANTENNAS Reconfigurable antenna TX RX TX RX DIFFERENT CHANNELS maximize SINGLE channel CHANNEL enhance pre-coding diversity + SNR at suppress interference CHANNEL SELECTION techniques receiver 6
Reconfigurable HW conceptual design Active antenna Passive metallization q Passive metal elements can be connected or disconnected from a ground plane or connected and disconnected between each other to change the beam and null direction of the active element q Adant patented smart antenna technology and designs allow for best radiation efficiency, good impedance matching and lowest cost 7
Reconfigurable antenna capabilities q Each antenna in the array can generate up to 2 N independent radiation patterns where N is the number of parasitic elements in the antenna 8
2016-Q2 Adant Smart Antenna Portfolio All designs are scalable to different form factors to fit Wi-Fi base stations for indoor and outdoor applications Antennas with omnidirectional coverage Antennas with sectorial coverage 9
Integration with WiFi systems q Integration is designed to support packet by packet configuration switching q Antenna system allows for 1 µsec switching time WiFi 3x3 device ANT 1 CPU with reconfigurable antenna drivers GPIO WiFi chipset GPIO ANT 2 ANT 3 Antenna control circuitry DC bias line 10
Maximum diversity + SNR H1 RX H2 HN TX Q array configurations y = H x + n Optimal antenna selection C = max log ) det I + SNR 2 N 3 H 2 H 2 11
Configuration selection in WLAN devices q Normally the channel matrix (H)is not directly available in commercial WLAN devices q The problem of selecting the optimal smart antenna configuration can be separated in two steps: l Evaluating the cost function: access the required parameters and compute the value l Smart Antenna training: optimizing the cost function 12
Cost function in WLAN devices q Cost function must be strongly correlated to the performance metric that needs to be optimized DATA PACKET STA ACK AP Available parameters RSSI SR rate BW PL Meaning Strength of received signal per antenna Packet success rate Rate of data transmission (modulation and coding) Bandwidth of operation Packet length Example of cost function (CF) for throughput (TP) optimization CF = f SR, RSSI, rate, BW, PL ρ CF, TP 0.9 13
Test setup q Downlink and uplink throughput is measured with an omnidirectional antenna and with a reconfigurable antenna using a 3x3 BS and 2x2 client Old villa Concrete ground and ceiling Harsh environment Underground parking lot Concrete ground and ceiling Quasi LOS environment 14
Measured Performance single client Reconfigurable antenna system improvement vs static antenna system 40% of cases where gain > 35% in download;; (30%in upload) Average TP impr. Downlink = 36% Uplink = 30% Data sample of 100 measurements 15
Measured Performance single client Performance benefit is max at low SNR Throughput Improvement vs. internal omnidirectional antennas 5 GHz DOWNLINK UPLINK ALL THROUGHPUT RANGES HIGH THROUGHPUT (>60% of MAX THROUGHPUT) LOW THROUGHPUT (<30% of MAX THROUGHPUT) 1.36x 1.30x 1.15x 1.04x 1.43x 1.34 16
Measured Performance multi client Reconfigurable antenna system improvement vs static antenna system Average TP improvement = 30% Data sample of 44 measurements 17
Reconfigurable antennas for interference mitigation H1 RX H2 HN TX Q array configurations Optimal antenna selection C = max log ) det I + SNR 2 N 3 H 2 H 2 W 2 HI 18
Adant interference mitigation Dynamically change direction of radiation to maximize power at the receiver while minimizing interference 19
Interference test setup q Throughput in downlink is measured with an omnidirectional antenna and with a reconfigurable antenna using a 3x3 BS and 2x2 client q Client (STA) is exposed to a lower amount of interference with respect to the AP q Distance between AP and client is minimized to determine only the effect due to interference mitigation D > d D > d d < 5 m d < 5 m 20
Measured Performance 20 different scenarios Reconfigurable antenna system improvement vs static antenna system Average TP improvement = 3.5X 21
High density deployment q Multiple APs operating on the same frequency channel q Co-channel interference is main cause of performance degradation q Reconfigurable antennas provides coordinated interference mitigation capabilities to improve the network capacity 4 7 8 AP 2 AP 4 6 2 AP 1 AP 3 22
Aggregate Throughput [Mb/s] 600 500 400 300 200 600 Multiple base stations with implicit zyxel ruckus 500 cisco aruba coordination 400 AP with Adant reconfigurable antenna system zyxel ruckus cisco aruba 300 Each optimizes its own cost function with no 200 coordination with the other base 100 stations Aggregate Throughput [Mb/s] 400 300 200 Aggregate Throughput [Mb/s] Aggregate Throughput [Mb/s] 0 300 200 Deployment 1 1 BSS1 BSS2 BSS1+BSS2 Deployment 1 1 Deployment 1 2 Deployment 2 1 Deployment 1 700 Deployment 2 600 Deployment 3 zyxel zyxel ruckus ruckus 600 cisco 500 cisco aruba aruba 500 400 Aggregate Throughput [Mb/s] Aggregate Throughput [Mb/s] 700 600 500 400 300 200 100 0 350 300 250 200 150 100 100 100 100 50 0 BSS1 BSS2 BSS1+BSS2 ruckus cisco aruba 5 5 7 7 AP#2 AP#2 6 6 5 57 7 6 Deployment 1 and 2 68 8 1 2 2 4 4 Aggregate Throughput [Mb/s] 1 3 3 AP#1 AP#1 0 350 300 250 200 150 100 50 0 3 4 BSS1 BSS2 BSS1+BSS2 zyxel ruckus cisco aruba Deployment 2 2 5 75 7 AP#2 AP#2 3 BSS1 BSS2 BSS1+BSS2 4 6 86 8 0 BSS1 BSS2 BSS1+BSS2 Figure.5.Aggregate)throughput)of)each)BSS,)total)per)dep Deployment 3 ) We)repeated)a)very)similar)test)with)four)additional)deploym technology) may) address) the) cotchannel) interference) pro transmission)power:)to)this)end)we)increased)the)transmiss 23dBm)keeping)the)others)to)17dBm.)This)should)increase) but)only)for)the)aps)equipped)with)sa)technology.) ) As)before)we)report)in)Figure)6)the)positions)of)the)clients)fo first)two,)3t1)and)3t2,)are)on)the)left;)the)others,)4t1)and)4t2 to)indicate)how)we)moved)nodes)from)one)deployment)to)t ) 8 8 7 7 2 2 1 1 1 AP#1 AP#1 1 3 3 2 2 4 4 ) ) 23 0
Explicit coordination q Joint user-beam selection 4 AP 2 6 AP 4 AP 2 AP 4 4 7 8 3 AP 1 AP 3 AP 1 2 AP 3 A. Michaloliakos, W. C. Ao, and K. Psounis, Joint user-beam selection for hybrid beamforming in asynchronously coordinated multi-cell networks, in Proceedings of Information Theory and Applications Workshop (ITA), San Diego, California, USA, February 2016. 24
Coordination schemes performance Avg. Throughput per user [Mbps] 60 50 40 30 20 10 SU MISO MU MIMO Coor. MU Hybrid 1x Power Coor. MU Hybrid 2x Power Coor. MU Hybrid 8x Power Coor. MU MIMO 30x30m hall, 200 users, AP density: 1 AP/BS per 112.5/45/22.5 sq m (~10/7/5m between APs/BSs) User density: 1 active user per 4.5 sq m (~10% of users active) 0 8 20 40 Number of stations A. Michaloliakos, W. C. Ao, and K. Psounis, Joint user-beam selection for hybrid beamforming in asynchronously coordinated multi-cell networks, in Proceedings of Information Theory and Applications Workshop (ITA), San Diego, California, USA, February 2016. 25
Smart antenna and digital beamforming Maximum signal strength at receiver and channel selection H1 RX H2 HN TX Q array configurations y = H Bx The reconfigurable antenna selects the best wireless channel and the precoding matrix optimizes that channel for maximum SNR at the receiver 26
Smart antenna and digital beamforming Optimal combined radiation patterns to improve link of interest (power at receiver) and interference mitigation (null steering towards interfering sources) Combined beamforming and reconfigurable antenna digital beamforming only AP Optimal control of radiation nulls and radiation in the intended direction 27
28 Test setup Downlink throughput is measured with an omnidirectional antenna and with a reconfigurable antenna system using an AP with digital beamforming Old villa Concrete ground and ceiling Harsh environment Underground parking lot Concrete ground and ceiling Quasi LOS environment
Measured Performance Reconfigurable antenna system improvement vs static antenna system Average TP impr. = 32% 25% of cases where gain >40% 40% Data sample of 100 measurements 29
Advantages of Adant and TXBF Adant smart antenna system significantly improves AP performance with and without TXBF Average benefit vs. standard antenna system TXBF OFF TXBF ON 35% 32% Results based on >50 independent measurements in two different environments EXAMPLE Standard Omni Antennas Adant smart antenna system TXBF OFF 100 Mbps 135 Mbps TXBF ON 130 Mbps 170 Mbps 30
Reconfigurable antennas and MU-MIMO The reconfigurable antenna creates the wireless channel to enable the possibility of simultaneous transmission to multiple users through the proper precoding matrix H1 H N B 2 = 0 RX 1 H2 TX HN H N B 1 = 0 RX 2 31
STANDARD ANTENNA RECONFIGURABLE ANTENNA Grouping and reconfigurable antennas Client A Client B Client A Client B SCENARIO A Grouping OK Client C A P Grouping OK Client C A P Omnidirectional radiation Smart antenna configuration Client B Grouping NOT possible Client B Client A SCENARIO B A P Grouping OK Client A A P Client C Omnidirectiona l radiation Client C Smart antenna configuration 32
33 Test setup q Different combinations of three MU-MIMO clients are connected to a 4x4 commercial grade AP MU-MIMO capable based on QCA 9990 chipset q Aggregate downlink throughput is measured with an omnidirectional antenna and with a reconfigurable antenna system
Measured Performance Reconfigurable antenna system improvement vs static antenna system 45% of cases where gain >30% Average TP improvement = 30% Data sample of >40 measurements 34
Conclusions q Reconfigurable antenna systems significantly enhance WiFi networks by: l Maximizing network capacity l Mitigating interference through implicit or explicit coordination between base stations l Enhancing based band pre-coding techniques like MU-MIMO q Antenna selection speed and uplink benefit highly depend on the level of integration that can be achieved with the WiFi chipset q Additional benefits that Adant is further exploring in using reconfigurable antennas in WiFi networks are relative to the possibility of l Enhancing security of the network l Providing reliable localization services 35