Mr. Krishna Kumar.S,C-DAC Thiruvananthapuram

Download Report

Transcript Mr. Krishna Kumar.S,C-DAC Thiruvananthapuram

Centre for Development of Advanced Computing
Software Defined Radio
& Cognitive Radio:
Implementation Initiatives
KRISHNA KUMAR S.
Centre for Development of Advanced Computing
[email protected]
Centre for Development of Advanced Computing
C-DAC
A National Centre of Excellence
and premier R & D institution
under DIT, MCIT, Govt. of India
involved in the design, development and
deployment of Electronics & IT- based
solutions for human advancement
25 Feb 2012
sdr & cr: implementaion initiatives
2 of 60
Centre for Development of Advanced Computing
10 Locations
14 Centres
3000 members
25 Feb 2012
sdr & cr: implementaion initiatives
3 of 60
Centre for Development of Advanced Computing
Broadcast & Communications @ C-DAC
AM/FM Radios & Television
Alpha Numeric Information
Displays
CCTV Cameras
Monitors
TETRA
VoIP
Direct Reception System
Digital Audio
Multifunction mono/stereo
Audio Consoles
Software Defined Radio
MANET
CR
CNM
CD Players
Equalizers
NG WF SDR
Networking
Monitoring Amplifiers
Digital Audio Work Stations
1990
25 Feb 2012
TETRA-WiMax
2000
2010
sdr & cr: implementaion initiatives
2015
4 of 60
Centre for Development of Advanced Computing
C-DAC SDR programme road map
Handheld
SDR Demos- 2 SCA WFS
Study
Report
C-DAC
SDR
proposal
CR
SDR
Manpack
1st
SDR SDR
Project PoC
2004
2006
2007
2008
DIT
DIT
DIT
C-DAC
25 Feb 2012
SDR-NC
2009
SDR-Demo
2010/11
2012
2012
2013
DIT
DIT
NAVY
C-DAC
sdr & cr: implementaion initiatives
5 of 60
Centre for Development of Advanced Computing
recent/current projects in SDR/CR
SDR for Naval Communication (DRDO)
SDR Manpack (DIT)
SDR Handheld (Core funding)
Cognitive Radio Networks (DIT)
jointly with IISc.
Next Generation Waveforms for SDR
Related projects
Mobile Adhoc NETworks (MANET)
Autonomic Network Management Systems (ANMS)
25 Feb 2012
sdr & cr: implementaion initiatives
6 of 60
Centre for Development of Advanced Computing
the PoC SDR lab model
Demonstrated re-configurablity with two SCA compliant waveforms
o
TETRA UHF band & Military
o
Legacy FM Radio (VHF band) Clear Mode
25 Feb 2012
sdr & cr: implementaion initiatives
7 of 60
Centre for Development of Advanced Computing
SDR manpack: product perspective
25 Feb 2012
sdr & cr: implementaion initiatives
8 of 60
Centre for Development of Advanced Computing
agenda
introduction
SDR architecture
SDR waveforms
from SDR to CR
Spectrum Sensing Engine
application scenarios
conclusions
25 Feb 2012
sdr & cr: implementaion initiatives
9 of 60
Centre for Development of Advanced Computing
Software Defined Radio
a radio in which some or all of the radio’s
operating functions are implemented through
modifiable software or firmware
25 Feb 2012
sdr & cr: implementaion initiatives
10 of 60
Centre for Development of Advanced Computing
SDR Platform
Consists of
 Hardware,
 Firmware,
 Operating system
 Middleware
Takes different personalities, defined by
the waveform that is loaded
25 Feb 2012
sdr & cr: implementaion initiatives
11 of 60
Centre for Development of Advanced Computing
SDR platform architecture
IF to
250MHz
LO
LNA
AMP
12-16 BITS
60-125MSPS
ADC
IMG
FILTER
MXR
MEM
MEM
FPGA
DSP
+
GPP
IF
FILTER
CLK
DUPLEXER
OR
T/R SW
CLK
MXR
AMP
CLK
DAC
CLK
MMI
RF
FILTER
IF
FILTER
LO
12-16 BITS
200-1000MSPS
12 BIT / 250MSPS
LO
DDS
SYNTH
ADC
MXR
LO
IF
FILTER
CLK
25 Feb 2012
CLK
sdr & cr: implementaion initiatives
CLK
CLK
CLK
FREQ
STD
12 of 60
Centre for Development of Advanced Computing
RF transceiver
25 Feb 2012
sdr & cr: implementaion initiatives
13 of 60
Centre for Development of Advanced Computing
Hopping Synthesizer
25 Feb 2012
sdr & cr: implementaion initiatives
14 of 60
Centre for Development of Advanced Computing
Harmonic Filter Bank
25 Feb 2012
sdr & cr: implementaion initiatives
15 of 60
Centre for Development of Advanced Computing
Baseband Boards
25 Feb 2012
sdr & cr: implementaion initiatives
16 of 60
Centre for Development of Advanced Computing
agenda
introduction
SDR architecture
SDR waveforms
from SDR to CR
Spectrum Sensing Engine
application scenarios
conclusions
25 Feb 2012
sdr & cr: implementaion initiatives
17 of 60
Centre for Development of Advanced Computing
Waveform: Definition
From http://www.wirelessinnovation.org/Introduction_to_SDR
The set of transformations applied to information to
be transmitted and the corresponding set of
transformations to convert received signals back to
their information content.
Representation of a signal in space
The representation of transmitted RF signal plus
optional additional radio functions up to and
including all network layers.
25 Feb 2012
sdr & cr: implementaion initiatives
18 of 60
Centre for Development of Advanced Computing
Waveform
Can be visualized at different levels
Architecture
 Conceptual entity
 Defines and abstracts the waveform functions
 Almost independent of the platform specifics
Implementation
 Physical realization of architecture
 Closely related to platform
25 Feb 2012
sdr & cr: implementaion initiatives
19 of 60
Centre for Development of Advanced Computing
Waveform Architecture design
What it is?
What it is not?
25 Feb 2012
sdr & cr: implementaion initiatives
20 of 60
Centre for Development of Advanced Computing
Architecture design process
User
behavioural
patterns
operational
needs
Candidate
Architecture
Radio
Standards
Technical
Specifications
SIMULATION
Final
Architecture
25 Feb 2012
sdr & cr: implementaion initiatives
21 of 60
Centre for Development of Advanced Computing
Waveform Implementation
Physical realization of architecture
Closely related to platform
Implementer should know
Overall platform architecture
Availability of Computing elements
• GPP, DSP, FPGA
Other configurable resources
• clocks, vca, vco, tunable filters etc.
25 Feb 2012
sdr & cr: implementaion initiatives
22 of 60
Centre for Development of Advanced Computing
Task Partitioning among CEs
GPP
DSP
Signaling and
control
Higher layer
and MAC
functions
RULES OF THUMB
25 Feb 2012
PHY Bit level
processing
Symbol rate
processing
Soft real-time
numerically
intensive tasks –
e.g. channel
estimation
FPGA
Ideally all hard
real time PHY
functions
Tasks best
implemented using
parallel
architecture
Symbol rate
processing for
wideband systems
sdr & cr: implementaion initiatives
23 of 60
Centre for Development of Advanced Computing
What devices in a given SDR?
Device architectures are being upgraded
constantly
New FPGAs realize DSP functions using
specific architectures
New DSPs use hardware accelerators to
implement hard real time tasks
GPP performance too scales up
Blurred boundaries! Vanishing boundaries?
Platform designer priorities do matter
25 Feb 2012
sdr & cr: implementaion initiatives
24 of 60
Centre for Development of Advanced Computing
Portability & Re-configurability
Probably the most important features of SDR
Waveform should be portable across platforms [a
statement to be qualified]
Waveform should be able to configure and control
platform resources
Ensured by proper design and implementation of
Waveform and Platform
May result in sub-optimal implementation
 but that’s okay! in most cases
25 Feb 2012
sdr & cr: implementaion initiatives
25 of 60
Centre for Development of Advanced Computing
Application Program Interfaces
Key enabler in ensuring portability & reconfigurability
Abstracts low level functions
Platform provider to facilitate platform
abstraction through APIs
Waveform implementer to use APIs to
access platform features
25 Feb 2012
sdr & cr: implementaion initiatives
26 of 60
Centre for Development of Advanced Computing
GPP & DSP APIs
GPP API calls are typically POSIX calls
DSP API calls are C-function calls
API implemented as a library
API to be used while building DSP image
25 Feb 2012
sdr & cr: implementaion initiatives
27 of 60
Centre for Development of Advanced Computing
RF APIs
To abstract Radio functions
 Tuning LO
 Configure Tx. DAC
 Configure AGC
 Etc.
API calls are pre-defined messages
Processed and executed by a dedicated controller
25 Feb 2012
sdr & cr: implementaion initiatives
28 of 60
Centre for Development of Advanced Computing
FPGA: Wrapper
Equivalent of API for FPGA
Wrapper defines the platform logic
Waveform logic defines the (part of)
PHY signal processing
Waveform logic to be integrated with
wrapper
25 Feb 2012
sdr & cr: implementaion initiatives
29 of 60
Centre for Development of Advanced Computing
FPGA: Architecture
FPGA IO Ring
RF Control
TOP module
SPI Glue
ADC
DAC
ADC
FIFO
DAC
FIFO
SPI Signals
14 bit
16 bit
EMIF
EMIF
Glue
Waveform
Logic
DSP
GPIO
(Push
Buttons, LEDs,
DIP Switches,
GPIO
Headers)
UART Port
25 Feb 2012
McBSP
McBSP
Glue
Interrupt 1
Interrupt 2
UART
Glue
uPP
Glue
uPP
sdr & cr: implementaion initiatives
30 of 60
Centre for Development of Advanced Computing
FPGA: model based design
To design waveform signal processing
Can be done in a graphical way
Designer need not no low-level
architecture of the device
Can be used jointly with
Matlab/Simulink or similar simulation
environments
25 Feb 2012
sdr & cr: implementaion initiatives
31 of 60
Centre for Development of Advanced Computing
FPGA: model based design tools
Xilinx - System Generator
Altera - DSP Builder
Actel - Synplify
Lattice - ispLever DSP
Agilent system view – VHDL code
generation
25 Feb 2012
sdr & cr: implementaion initiatives
32 of 60
Centre for Development of Advanced Computing
FPGA: integration with wrapper
Combine the HDL source level
Combine the wrapper source with waveform
netlist
 Similar to adding a library
Combine at bitmap level
 Wrapper logic implemented in advance
 Waveform logic added using partial reconfiguration
25 Feb 2012
sdr & cr: implementaion initiatives
33 of 60
Centre for Development of Advanced Computing
agenda
introduction
SDR architecture
SDR waveforms
from SDR to CR
Spectrum Sensing Engine
application scenarios
conclusions
25 Feb 2012
sdr & cr: implementaion initiatives
34 of 60
Centre for Development of Advanced Computing
from SDR to CR
A natural evolution
CR, by nature, has to be an SDR
but,
an SDR with certain specific fetures.
25 Feb 2012
sdr & cr: implementaion initiatives
35 of 60
Centre for Development of Advanced Computing
CR additional requirements
A truly wideband radio front end
Support for White/Gray space detection
Spectrum sensing - hardware & software
Geo-location and Database
• IEEE 802.22
Dynamic Spectrum Management
Channel & Bandwidth allocation
Rate adaptation & Tx power control
25 Feb 2012
sdr & cr: implementaion initiatives
36 of 60
Centre for Development of Advanced Computing
from SDR to CR: concerns
Algorithm complexities – of course
A truly wideband radio front end
Tx side RF Power amps
Rx side – wide band sensing
AD/DA conversion bottle-necks
Noise, sensitivity, interference
protection, SFDR
25 Feb 2012
sdr & cr: implementaion initiatives
37 of 60
Centre for Development of Advanced Computing
agenda
introduction
SDR architecture
SDR waveforms
From SDR to CR
Spectrum Sensing Engine – design &
implementation
application scenarios
conclusions
25 Feb 2012
sdr & cr: implementaion initiatives
38 of 60
Centre for Development of Advanced Computing
Primary Signal Details
Primary: Terrestrial analog TV txn in India
System: CCIR system B,G PAL
Bands:
• Band II VHF: 174 to 225 MHz
• Band IV UHF: 470 to 582 MHz
Channel BW: 8 MHz
25 Feb 2012
sdr & cr: implementaion initiatives
39 of 60
Centre for Development of Advanced Computing
Requirements
Incumbent Detection Threshold:
 -94 dBm (measured at peak of sync)
Channel Detection Time:
 <=2 sec per channel
Detection Performance:
 Probability of Detection >=90% at False
Alarm rate of <= 10%
Guided by IEEE 802.22 WRAN WG interim recommendations
25 Feb 2012
sdr & cr: implementaion initiatives
40 of 60
Centre for Development of Advanced Computing
Platform - Lyrtech SFF SDR
25 Feb 2012
sdr & cr: implementaion initiatives
41 of 60
Centre for Development of Advanced Computing
25 Feb 2012
sdr & cr: implementaion initiatives
42 of 60
Centre for Development of Advanced Computing
From SS Algorithm to SS Engine
(from “recipe” to “dish”)
Study of algorithm
Study of hardware architecture
Designing software architecture
Optimal partitioning of algorithm
Model based design / C-program development
Fixed point considerations

dynamic range, bit growth, over/under-run,
truncation error
Defining and realizing interfaces
Debugging, testing and optimization
25 Feb 2012
sdr & cr: implementaion initiatives
43 of 60
Centre for Development of Advanced Computing
Detection Scheme
Pre-processing/Feature extraction stage
 Extracts the spectrum around the picture carrier
Energy Detection stage
 Computes the energy around the pictures carrier
25 Feb 2012
sdr & cr: implementaion initiatives
44 of 60
Centre for Development of Advanced Computing
25 Feb 2012
sdr & cr: implementaion initiatives
45 of 60
Centre for Development of Advanced Computing
Pre-Processing
Decimating filter
stages
Digital IF
BW: 8MHz
125 Msps
14 bits
5
5
5
2
to sensing
algorithm
BW: 100kHz
500 Ksps
32 bits
DDS @
30MHz
25 Feb 2012
sdr & cr: implementaion initiatives
46 of 60
Centre for Development of Advanced Computing
Implementation
Implemented in the FPGA part (Virtex 4)
Design using Simulink / System generator

Model based design approach
Fixed point implementation
Word-length selection



25 Feb 2012
bit growth
truncation error
resource utilization
sdr & cr: implementaion initiatives
47 of 60
Centre for Development of Advanced Computing
Energy Detection
N samples x B
buffers
Gives the
adavntage of
averaging
Reduces the FFT
implementation
complexity
25 Feb 2012
sdr & cr: implementaion initiatives
48 of 60
Centre for Development of Advanced Computing
Implementation
Implemented in the DSP (TMS320C64X)
Code developed in C language
Debugged using Code Composer Studio & XDS560 ICE
A fixed point implementation
Word-lengths selection
25 Feb 2012
sdr & cr: implementaion initiatives
49 of 60
Centre for Development of Advanced Computing
User Interface
25 Feb 2012
sdr & cr: implementaion initiatives
50 of 60
Centre for Development of Advanced Computing
25 Feb 2012
sdr & cr: implementaion initiatives
51 of 60
Centre for Development of Advanced Computing
Lab Setup
Centre for Development of Advanced Computing
Test Settings
Picture carrier: 67.25MHz
Channel Bandwidth: 8MHz
Video pattern: White
Sensing duration: 20.4 ms
Sampling rate: 125Msps
Ensemble size: 1e5
SNR values: -30dB, -27dB, -24dB
25 Feb 2012
sdr & cr: implementaion initiatives
53 of 60
Centre for Development of Advanced Computing
PAPR for ATV
Courtesy: Martyn J. Horspool, Analog-to-digital Upgradeable
Transmitters For the Worldwide Market, Harris Corporation
25 Feb 2012
sdr & cr: implementaion initiatives
54 of 60
Centre for Development of Advanced Computing
Results
25 Feb 2012
sdr & cr: implementaion initiatives
55 of 60
Centre for Development of Advanced Computing
Results
System meets the false alarm/miss
detection performance at -27 dB SNR
Highly encouraging result
Enough margin to accommodate large scale
fading
Caveat: This is only a lab measurement
25 Feb 2012
sdr & cr: implementaion initiatives
56 of 60
Centre for Development of Advanced Computing
agenda
introduction
SDR architecture
SDR waveforms
from SDR to CR
Spectrum Sensing Engine
application scenarios
conclusions
25 Feb 2012
sdr & cr: implementaion initiatives
57 of 60
Centre for Development of Advanced Computing
application areas
Military
PMR (public safety, police, paramilitary)
Disaster management
Commercial Cellular (Base Stations)
Rural broadband access
IEEE 802.22 system adaptaion
Tele-Medicine
25 Feb 2012
sdr & cr: implementaion initiatives
58 of 60
Centre for Development of Advanced Computing
acknowledgements
Simon Zachariah
Beena K. T.
S. Sagar
Chandra R. Murthy
Shine K. P.
Satheesh Kumar S.
25 Feb 2012
sdr & cr: implementaion initiatives
59 of 60
Centre for Development of Advanced Computing
Questions ??
…thank you