Click to enter your title
MIMO Technology for Advanced Wireless Local Area Networks
Dr. Won-Joon Choi Dr. Qinfang Sun Dr. Jeffrey M. Gilbert Atheros Communications
2005 Design Automation Conference June 15, 2005
Agenda
This presentation will give an overview of MIMO technology and its future in Wireless LAN:
Wireless Local Area Networks (WLAN)
Current standards (11a/b/g) Next-generation 11n overview and status
Beamforming Spatial Multiplexing Bandwidth Number of spatial streams
MIMO fundamentals
MIMO scalability
The Wireless LAN Explosion
The Wireless LAN / Wi-Fi market has exploded! New technology is enabling new applications: Office
Email / Info anywhere Voice over IP
Home
Internet everywhere Multimedia
Hot-spots
Hot-spot coverage Metro-Area Networks
Wireless LAN Technology Advances
Wireless LAN technology has seen rapid advancements
Standards: 802.11 .11b .11a .11g Data rates: 2Mbps 100+ Mbps Range / coverage: Meters kilometers Integration: Multiple discretes single chip solutions Cost: $100s $10s (sometimes free w/rebates!)
How can this growth continue?
Previous advances have been limited to a single
transmitting and receiving radio The next generation exploits multiple parallel radios using revolutionary class of techniques called MIMO (Multiple Input Multiple Output) to send information farther and faster
Existing 802.11 WLAN Standards
802.11b
Standard Approved Available Bandwidth Sept. 1999 83.5 MHz
802.11a
Sept. 1999 580 MHz
802.11g
June 2003 83.5 MHz
802.11n
? 83.5/580 MHz 2.4/5 GHz
Frequency Band of Operation # Non-Overlapping Channels (US) Data Rate per Channel
2.4 GHz
5 GHz
2.4 GHz
3 1 11 Mbps
24 6 54 Mbps
3 1 54 Mbps DSSS, CCK, OFDM
3/24 1 600 Mbps DSSS, CCK, OFDM, MIMO
Modulation Type
DSSS, CCK
OFDM
What Is Being Proposed for 802.11n?
Main Features PHY MIMO-OFDM
Beamforming
Spatial
Multiplexing
Extended bandwidth (40MHz) Advanced coding
MAC
Aggregation Block ACK Coexistence Power saving
Wireless Fundamentals I
In order to successfully decode data, signal strength needs to be greater than noise + interference by a certain amount
Higher data rates require higher SINR (Signal to Noise and Interference Ratio) Signal strength decreases with increased range in a wireless environment
60
Data Rate 1 Data Rate 2
Throughput
50 40 30 20 10 0 1 2 3 4 5 6 7 8
9 10 11 12
Range
Wireless Fundamentals II
Ways to increase data rate:
Conventional single tx and rx radio systems
Increase transmit power
Subject to power amplifier and regulatory limits Increases interference to other devices Reduces battery life Fixed direction(s) limit coverage to given sector(s) Subject to FCC / regulatory domain constraints
Use high gain directional antennas
Use more frequency spectrum
Advanced MIMO: Use multiple tx and / or rx radios!
Conventional (SISO) Wireless Systems
channel
Bits TX DSP Radio Radio DSP RX Bits
Conventional Single Input Single Output (SISO) systems were favored for simplicity and low-cost but have some shortcomings:
Outage occurs if antennas fall into null Switching between different antennas can help Energy is wasted by sending in all directions Can cause additional interference to others Sensitive to interference from all directions Output power limited by single power amplifier
MIMO Wireless Systems
Bits D S P TX Radio Radio
channel
Radio Radio
D S P RX
Bits
Multiple Input Multiple Output (MIMO) systems with multiple parallel radios improve the following:
Outages reduced by using information from multiple antennas Transmit power can be increased via multiple power amplifiers Higher throughputs possible Transmit and receive interference limited by some techniques
MIMO Alternatives
There are two basic types of MIMO technology:
Beamforming MIMO Standards-compatible techniques to improve the range of existing data rates using transmit and receive beamforming Also reduces transmit interference and improves receive interference tolerance
Spatial-multiplexing MIMO Allows even higher data rates by transmitting parallel data streams in the same frequency spectrum Fundamentally changes the on-air format of signals Requires new standard (11n) for standards-based operation Proprietary modes possible but cannot help legacy devices
Beamforming MIMO Overview
Consists of two parts to make standard 802.11 signals better Uses multiple transmit and/or receive radios to form coherent 802.11a/b/g compatible signals
Receive beamforming / combining boosts reception of standard 802.11 signals
Radio D S P RX
Bits TX
Radio Radio
Bits
Phased array transmit beamforming to focus energy to each
receiver
Bits TX
D S P
Radio
Radio Radio Bits RX
Benefits of Beamforming
Benefits
Power gain (applicable only to transmit beamforming)
Power from multiple PAs simultaneously (up to regulatory limits) Relaxes PA requirements, increases total output power delivered
Array gain: dynamic high-gain antenna Interference reduction
Reduce co-channel inter-cell interference
Diversity gain: combats fading effects
Multipath mitigation
Per- subcarrier beamforming to reduce spectral nulls
Multipath Mitigation
Multiple transmit and receive radios allow compensation of notches on one channel by non-notches in the other Same performance gains with either multiple tx or rx radios and greater gains with both multiple tx and rx radios
Spatial Multiplexing MIMO Concept
Spatial multiplexing concept:
Form multiple independent links (on same channel) between transmitter and receiver to communicate at higher total data rates
DSP
Bits Bit Split TX DSP
Radio Radio
Radio
DSP DSP
Bit Merge RX
Bits
Radio
Spatial Multiplexing MIMO Difficulties
Spatial multiplexing concept:
Form multiple independent links (on same channel) between transmitter and receiver to communicate at higher total data rates
However, there are cross-paths between antennas
DSP
Bits Bit Split TX DSP
Radio Radio
Radio
DSP DSP
Bit Merge RX
Garbage
Radio
Spatial Multiplexing MIMO Reality
Spatial multiplexing concept:
Form multiple independent links (on same channel) between transmitter and receiver to communicate at higher total data rates
However, there are cross-paths between antennas
The correlation must be decoupled by digital signal processing algorithms
DSP
Bits Bit Split TX DSP
Radio Radio
Radio
D S P
Bit Merge RX
Bits
Radio
Spatial Multiplexing MIMO Theory
High data rate
Data rate increases by the minimum of number of transmit and receive antennas Detection is conceptually solving equations Example of 2-by-2 system:
Transmitted signal is unknown, x1 , x2 Received signal is known, y1 , y2 Related by the channel coefficients, h11, h12 , h21, h22
y1 h11x1 h12 x2 y2 h21x1 h22 x2
Need more equations than unknowns to succeed
High spectral efficiency
Higher data rate in the same bandwidth
MIMO Scalability
Moores law
Doubling transistors every couple of years Increases number of streams Higher performance/speed Higher complexity
MIMO
MIMO is the bridge to allow us to exploit Moores law to get higher performance
MIMO Scalability
Notation
R: data rates (Mbps) Es: spectral efficiency (bps/Hz) Bw: bandwidth (MHz) Ns: number of spatial streams NR: number of Rx chains NT: number of Tx chains
MIMO Scalability
Data Rates
R = Es * Bw * Ns -> Scales with bandwidth and the number of spatial streams Example
11a/g: Es = 2.7; Bw = 20MHz; Ns=1; R = 54Mbps Spatial multiplexing MIMO
Es = 3.75; Bw=40MHz;Ns = 2; R = 300Mbps
Number of Tx/Rx chains
At least as many chains as Ns Ns = min(NR, NT)
MIMO Hardware Requirements
MIMO Transmitter (parallelism and data rate scaling)
MOD FEC Stream Split MOD Spatial Mapping
IFFT
RF
IFFT
RF
1* O(Bw*Es*Ns)
Ns * O(Bw*Es)
1* NT* NT* O(Bw*Es*Ns*NT) O(Bw*Es) Analog RF
MIMO Hardware Requirements
MIMO Receiver (parallelism and data rate scaling)
RF
FFT
Demod
MIMO Equalizer
RF
Stream Merge Demod
DEC
FFT
NR* Analog RF
NR* O(Bw*Es)
1* O(Bw*Es*NR*Ns2)
Ns* O(Bw*Es)
Ns* 1* O(Bw*Es) O(Bw*Es*Ns)
Conclusions
The next generation WLAN uses MIMO technology Beamforming MIMO technology
Extends range of existing data rates by transmit and receive beamforming
Spatial-multiplexing MIMO technology
Increases data rates by transmitting parallel data streams
MIMO allows system designers to leverage Moores law to deliver higher performance wireless systems
Circuit Implications of MIMO
Crystal
Common crystal is required Common synthesizer is preferred Allow additional flexibility
With total power limit, PA requirements relaxed With PA limit, total power increased.
Synthesizer
PA
Cross-talk/ Coupling
Need to minimize coupling between antennas
Circuit Impairments/Corrections
Timing offset
Common across multiple chains Common across multiple chains Common with common synthesizer With independent synthesizers, a new tracking algorithm may be needed. 1/f noise, I/Q mismatch, spurs, etc. Estimated and corrected for each chain
Frequency offset
Phase noise
Other impairments
Backup Slides
0.18um standard digital CMOS 7.2x7.2 mm2 die size 15x15mm2 BGA with 261 balls Ref: ISSCC05
Backup Slides
MIPS R4Kc, 16kB I and D caches 16b SDRAM interface 9b ADCs (4x) 9b DACs (4x) 180 MHz 100 MHz < 0.65 LSB INL&DNL, -48dB SNDR, 27mW <0.25 LSB INL&DNL, -51dB SNDR, 20mW
Total power, PCI mode, CPU off
Total power, MPEG-TS mode, CPU on
690 mW
1.8W
Supports 802.11 a, b, g, 20 and 40 1 to 108 Mb/s raw data rates MHz channel BW
Backup Slides
SDRAM Controller and Memory Interface MPEG-TS Local Bus MIPS Processor I2C WLAN MAC MRC/BF PCI UART Peripheral Interface
RS232 LED Control GPIOs
SDRAM and Flash
ADCs
2.4/5 GHz Rx Transceiver
Tx
Video Encoder/ Decoder
DACs
2.4/5 GHz Transceiver Rx
Tx
OFDM Mod/ Demod
Host System
Radio Control WLAN SOC
IR Interface
IR Remote Control
