Unit-II-SDR ARCHITECTURE

THE SOFTWARE RADIO FUNCTIONAL ARCHITECTURE

A. The Software Radio Functional Model

[Link] Set

 By using Multiband technology more than one RF band of communications

channel can be accessible

 The RF channel then is generalized to the channel set

 This set includes RF channels, but radio nodes like PCS base stations and
portable military radios also interconnected to fiber and cable; therefore
these are also included in the channel set.

[Link] encoder

 The channel encoder of a multiband radio includes RF/channel access, IF

processing, and modem.

 The RF/channel access includes wideband antennas, and the multi-element

arrays of smart antennas

 This segment also provides multiple signal paths and RF conversion that
span multiple RF bands
 IF processing may include filtering, further frequency translation,
space/time diversity processing, beam forming, and related functions.

 Multimode radios generate multiple air interface waveforms (modes)

defined principally in the modem

 These waveforms may be in different bands and may span multiple bands.

 A software-defined personality includes RF band, channel set (e.g., control

and traffic channels), air interface waveforms, and related functions.

[Link] security (INFOSEC)

 Many applications do not require information security (INFOSEC)

.
 Authentication reduces fraud.

 INFOSEC is therefore included in the functional model although the function

may be null for many applications.

 Transmission security (TRANSEC) hides the fact of a communications event

(e.g., by spread spectrum techniques )

 INFOSEC is therefore included in the functional model although the function

may be null for many applications.

 The source coder/decoder pair now includes the data, facsimile, video, and
multimedia sources.
 Some sources will be physically remote from the radio node, connected via
the synchronous digital hierarchy (SDH) , a local area network(LAN) .

 These functions may be implemented in multithreaded multiprocessor

software organized by a joint control function

[Link] Control

 Joint control ensures system stability, error recovery, timely data flow, and
isochronous streaming of voice and video.

 As radios become more advanced, joint control becomes more Complex

 Beam forming today requires dedicated processors

 In order to reduce wireless interference and improve the reliability of Wi-Fi

signal, the 5th generation Wi-Fi standard 802.11ac adapts an eective
technology, named beamforming.

 Beamforming can transmit data towards directions of given clients, instead

of radiating data in all directions

 Dynamic selection of band, mode, and diversity as a function of QoS causing

conflicts for processing resources
 In addition, processing resources are lost through equipment failures
 Joint control integrates fault modes, and support functions on processing
resources
 It include ASICs, FPGAs, DSPs, and general-purpose computers to yield a
reliable telecommunications object
 In a software radio, the user can upload a variety of new air interface
personalities
 The required resources (e.g., RF access, digitized bandwidth, memory, and
processing
 capacity) must not exceed those available on the radio platform.

 Some mechanism for evolution (development) support is therefore

necessary to define the waveform personalities, to download them (e.g.,
over the air) and to ensure that each new personality is safe before being
activated
 The evolution-support function therefore must include a software factory
 In addition, the evolution of the radio platform (the analog and digital
hardware of the radio node) must also be supported.
 This may be accomplished via the design of advanced hardware modules in
an integrated evolution support environment, or by the acquisition of
commercial off-the-shelf (COTS) hardware modules, or both

[Link] Allocation of the Software Radio Functional

Model
Functional Component Allocated Functions
1 Source Coding and Audio, data, video, and fax interfaces
Decoding
2 Service and Network Multiplexing; setup and control; data
Support services, internetworking

3 Information Security Transmission security, authentication,

privacy, data integrity

4 Channel Coding and Baseband modem, timing recovery,

Decoding: Modem equalization,
channel waveforms,.
processing
5 IF Processing Beamforming, diversity combining,
characterization
of all IF channels
6 RF Access Antenna, diversity, RF Conversion

7 Channel Set(s) Simultaneity, multiband propagation.

8 Multiple Personalities Multiband, multimode, agile services,

interoperable with legacy3
modes
9 Evolution Support Define and manage personalities
10 Joint Control Joint source/channel coding, dynamic QoS vs.
load control, processing resource
management
B. Functional Interfaces

 After identifying the functions to be accomplished in a software radio, one

must define the interface points among the functional components

 The IF waveform includes most aspects of the air interface, but the signals
have been filtered and converted to an IF that facilitates processing.

 In addition, IF processing may include A/D and D/A conversion.

 Baseband waveforms are almost always digital streams

 They may also be sampled replicas of analog signals, such as digitized FM
waveforms.

 The modem delivers what may be called decoded channel bits to the
INFOSEC function if one is present

 The modem may transform analog IF signals directly to channel bits

 INFOSEC then transforms these protected bits into clear bits (“red” bits)
.
 These bits may be manipulated through a protocol stack in order to yield
source bits or network bits.

 Network bits conform to a network protocol, while source bits are

appropriate for a source decoder

 The interface to local sources of voice, music, video, etc. includes an analog
Transducer
 Access to remote sources is accomplished via the network interface.

 In addition to these signal-processing interfaces, there are control

interfaces mediated by the user or network

 Personalities are downloaded to the radio via the software object interface.

 The simplest mechanism for maintaining radio software after deployment is

the downloading of a complete binary image of the radio.

 These interfaces may be thought of as the “horizontal” interfaces of the

 software radio, since they are concatenated to form signal and control
flows among sources and channels.

 In traditional radio engineering, the definition of such interfaces facilitated

the design and development of the radio.

 Variations in these definitions from one design team to another did not
matter, provided the component suppliersand the systems integrator all
agreed.

 The idea of plug-and-play hardware and software modules has become

popular in personal computing.
 The wireless industry seeks to benefit from the adaptation of plug-and-play
technology to the software radio.

 Plug-and-play requires industry wide agreement on architecture.

[Link]
 Since industry wide agreement on anything can be challenging, one should
begin with a definition of architecture

 Architecture is defined as “a fundamental underlying design of computer

hardware software, or both’

The IEEE prescribes that architecture consists of components and

interfaces.

1. Functions, Components, and Design Rules

 None of the many possible definitions of architecture suit the purposes of

defining architecture for the software radio.

 Suitable definition of architecture for SDR is a comprehensive, consistent

set of functions, components and design rules accordingto which radio
communications systems may be organized, designed, constructed,
deployed.

 This notion of architecture more clearly addresses partitioning for plug-and-

play, and the reuse of functional components.

 By including functions and design rules, architecture supports component

reuse

 A useful architecture partitions functions and components such that

 (a) functions are assigned to components clearly and
(b) physical interfaces among components correspond to logical interfaces
among functions.

 The design rules must ensure that when the hardware and software
components are mated, the resulting system doing the allotted functions

 Accommodating such diverse needs leads to complex radio systems that

must be further partitioned in order manage this complexity.
[Link]-and-Play

 If an architecture supports plug-and-play, then the design rules have been

crafted so that hardware and software modules from different suppliers will
work together when plugged into an existing system
 Hardware modules will plug-and-play if the physical interfaces and logical
structure of the functions supplied by that module are compatible with the
physical interfaces, logical structure of the host hardware platform.

 Software modules will plug-and-play if there is a comprehensive but simple

interface to the host environment, and
 if the module offers the information that it needs in order to employ it as a
resource.
 Software radio architecture, then, defines the partitioning of functions into
groups, which may subsequently be allocated to components.

Basic SDR
The Hardware Architecture of an SDR

 The basic SDR must include the radio front-end, the modem, the
cryptographic security function, and the application function.
 In addition, some radios will also include support for network devices
connected to the radio
 Allowing the radio to provide network services and to be remotely
controlled over the local Ethernet.

Traditional digital receiver signal processing block diagram

 Traditional transmitter signal processing block diagram

 Some radios will also provide for control of external radio frequency (RF)
analog functions such as antenna management, power amplifiers, or
special-purpose filters.
 The hardware and software architectures should allow RF external
features to be added
 The RF front-end (RFFE) consists of the following functions to support
the receive mode:
 Antenna-matching unit, low-noise amplifier, filters, local oscillators, and
analog-to-digital (A/D) converters (ADCs) to capture the desired signal
and suppress undesired signals to a practical extent.
 This maximizes the dynamic range of the ADC available to capture the
desired signal.
 To support the transmit mode, the RFFE will include digital-to-analog
(D/A) converters (DACs), local oscillators, filters, power amplifiers, and
antenna-matching circuits.
 In transmit mode, the important property of these circuits is to
synthesize the RF signal without introducing noise and spurious
emissions at any other frequencies that might interfere with other users
in the spectrum.
 The modem processes the received signal or synthesizes the transmitted
signal, or both for a full duplex radio.
 In the receive process the modem will shift the carrier frequency of the
desired signal to a specific frequency
 The digital filter provides a high level of suppression of interfering signals
.
 The modem then time-aligns and de-spreads the signal as required, and
refilters the signal to the information bandwidth
 The modem may include an equalizer to correct for channel multipath
artifacts.
 The modem will compare the received symbols with the possible
received symbols and make a best possible estimate of which symbols
were transmitted.
 Of course, if there is a weak signal or strong interference, some symbols
may be received in error.
 If the waveform includes FEC coding, the modem will decode the
received sequence of
 encoded symbols by using the structured redundancy introduced in the
coding process to detect and correct the encoded symbols that were
received in error.
 The process the modem performs for transmit is the inverse of that for
receive
 The modem takes bits of information to be transmitted, groups the
information into packets and adds a structured redundancy to provide
for error correction at the receiver,
 Groups bits to be formed into symbols selects a wave shape to represent
each symbol.
 synthesizes each wave shape and filters each wave shape to keep it
within its desired bandwidth.
 It may spread the signal to a much wider bandwidth by multiplying the
symbol by a wideband waveform
 The final waveform is filtered to match the desired transmit signal
bandwidth.
 If the waveform includes a time-slotted structure such as time division
multiple access (TDMA) waveforms
 the radio will wait for the appropriate time while placing samples that
represent the waveform into an output
 first in, first out (FIFO) buffer ready to be applied to the DAC
 The modem must also control the power amplifier and the local
oscillators to produce the desired carrier frequency and must control the
antenna-matching unit to minimize voltage standing wave radio (VSWR).
 The modem may also control the external RF elements, including
transmit versus receive mode, carrier frequency, and smart antenna
control.
 The cryptographic security function must encrypt any information to be
transmitted.
 Because the encryption processes are unique to each application, these
cannot be generalized.
 The Digital Encryption Standard (DES) and the Advanced Encryption
Standard (AES) from the US National Institute of Standards and
Technology (NIST) provide example cryptographic processes.
 In addition to providing the user with privacy for voice communication,
cryptography also plays a major role in assuring that the billing is to an
authenticated user terminal
 In the future, it will also be used to authenticate transactions of
delivering software and purchasing services.
 The application processor will typically implement a vocoder, a video
coder, and/or a data coder, as well as selected web browser functions
 (A vocoder (/ˈvoʊkoʊdər/, short for voice encoder) is a category of voice
codecthat analyzes and synthesizes the human voice signal for audio
data compression, multiplexing, voice encryption, voice transformation,
etc)
 In each case, the objective is to use knowledge of the properties of the
digitized representation of the information to compress the data rate to
an acceptable level for transmission.
 Voice, video, and data coding typically utilize knowledge of the
redundancy in the source signal (speech or image) to compress the data
rate.
 Compression factors typically in excess of 10:1 are achieved in voice
coding, and up to 100:1 in video coding.
 Data coding has a variety of redundancies within the message, or
between the message and common messages sent in that radio system.
 Data compression ranges from 10 to 50 percent, depending on how
much redundancy can be identified in the original information data
stream
 Typically, speech and video applications run on a DSP processor
 Text and web browsing typically run on a GPP.
 As speech recognition technology continues
 To improve its accuracy, we can expect that the keyboard and display
will be augmented by speech input and output functionality.
 On cognitive radios with adequate processors, it may be possible to run
speech recognition and synthesis on the cognitive radio,

Computational Processing Resources in an SDR

 The design of an SDR must expect the computational resources needed
to implement its most complex application.
 The computational resources may consist of GPPs, DSPs, FPGAs, and
occasionally will include other chips that extend the computational
capacity.
 Generally, the SDR vendor will avoid inclusion of dedicated-purpose
non-programmable chips because the flexibility to support waveforms
and applications is limited,
 Currently, an example GPP selected by many SDR developers is the
PowerPC.
 The PowerPC is available from several vendors.
 This class of processor supports a very wide variety of addressing modes,
floating point and integer computation, and a large memory space,
usually including multiple levels of on-chip and off-chip cache memory.
 These processors currently perform more than 1 billion mathematical
operations per second (mops).
 PowerPC (a backronym for Performance Optimization With Enhanced
RISC – Performance Computing, sometimes abbreviated as PPC) is
a RISCinstruction set architecture created by the 1991 Apple–IBM–
Motorola alliance, known as AIM.
 Reduced instruction set computing, or RISC (pronounced 'risk'), is a CPU
design strategy based on the insight that a simplified instruction set .
 GPPs in this class usually pipeline the arithmetic functions and decision
logic functions several levels deep in order to achieve these speeds.
 They also frequently execute many instructions in parallel, typically
performing the effective address computations in parallel with
arithmetic computation, logic evaluations, and branch decisions.
 Most important to the waveform modulation and demodulation
processes is the speed at which these processors can perform real or
complex multiply.
 The waveform signal processing represents more than 90 percent of the
total computational load in most waveforms,
 The protocols to participate in the networks frequently represent 90
percent of the lines of code.
 Therefore, SDR architecture include DSP-type hardware multiply
accumulate functions,
 (the multiply–accumulate operation is a common step that computes
the product of two numbers and adds that product to an accumulator.)

 so that the signal processes can be performed at high speed.

 GPP-type processors for the protocol stack processing.
 Usually the implication of this specialization is that the device has a
somewhat unusual memory architecture,
 Usually partitioned so that it can fetch two operands simultaneously
 and also be able to fetch the next software instruction in parallel with
the operand fetches
 Currently, DSPs can perform fractional mathematics (integer) multiply
accumulate instructions at rates of 1 GHz.
 and floating point multiply accumulates at 600 MHz.
 DSPs are also available with many parallel multiply accumulate engines,
reporting rates of more than 8 Gmops.
 The other major feature of the DSP is that it has far fewer and less
sophisticated addressing modes.
 As a result, the DSP is much more efficient at signal processing
 but less capable to accommodate the software associated with the
network protocols.
 FPGAs have recently become capable of providing tremendous amounts
of multiply accumulate operations on a single chip,
 Beat DSPs by more than an order of magnitude.
 By defining the on-chip interconnect of many gates, more than 100
multiply accumulators can be arranged to perform multiply accumulate
processes at frequencies of more than 200 MHz
 In addition to the DSP, FPGAs can also provide the timing logic to
synthesize clocks, baud rate, chip rate, time slot, and frame timing, thus
leading to a reasonably compact waveform implementation.
 By expressing all of the signal processing as a set of register transfer
operations and multiply accumulate engines, very complex waveforms
can be implemented in one chip.
 complex signal processes that are not efficiently
 Implemented on a DSP, such as Cordic operations, log magnitude
operations, and difference magnitude operations,
 All can be implemented by the specialized hardware in FPGAs.
 FPGA implementations tend to be higher power and more costly than
DSP chips.
 All three of these computational resources demand significant off-chip
memory.
 For example, a GPP may have more than 128 Mbytes of off-chip
instruction memory to support a protocols for today’s telephony
standards.
 Today’s SDRs provide a reasonable mix of these computational
alternatives to implement wide variety of desirable applications.

The Software Architecture of an SDR

 The objective is to place waveforms and applications onto a software-
based radio platform in a standardized way.
 These waveforms and applications are installed, used, and replaced by
other applications as required to achieve the user’s objectives.
 To make the waveform and application interfaces standardized, the
hardware platform present a set of highly standardized interfaces.
 This way, vendors can develop their waveforms independent of the
knowledge of the underlying hardware.
 Similarly, hardware developers can develop a radio with standardized
interfaces,
 which can subsequently be expected to run a wide variety of waveforms
from standardized libraries.
 In computer programming, an Application Programming Interface (API)
is a set of routine definitions, protocols, and tools for building software
and applications. A good API makes it easier to develop a program by
providing all the building blocks, which are then put together by the
programmer. An API may be for a web-based system, operating system,
database system, computer hardware, or software library.
 Web information system, or web-based information system, is
an information system that uses Internet web technologies to deliver
information and services, to user
 This way, the waveform development proceeds by assuming a
standardized set of interfaces (APIs) for the radio hardware, crossing
 interfaces to the unique underlying hardware through a set of common
drivers.
 In addition, the method by which a waveform is installed into a radio,
activated , deactivated, and de-installed, and the way in which radios
use the standard interfaces must be standardized.
 so that waveforms are reasonably portable to more than one hardware
platform implementation
 The use of published interfaces and industry standards in SDR
implementations will shift development process, away from tightly
coupled hardware software solutions.
 To achieve this, the SDR radio is decomposed into a stack of hardware
and software functions, with open standard interfaces.
 The stack starts with the hardware and the one or more data buses that
move information among the various processors.
 On top of the hardware, several standardized layers of software are
installed.
 This includes the boot loader, the operating system (OS); the board
support package (BSP, which consists of input/output drivers that know
how to control each interface);
 and a layer called the hardware abstraction layer (HAL). The HAL
provides a method for GPPs to communicate with DSPs and FPGA
processors.
 The US Government has defined a standardized software architecture,
known as the Software Communication Architecture (SCA),
 which has also been adopted by defense contractors of many other
countries worldwide.
 The SCA is a core framework to provide a standardized process for
identifying the available computational resources of the radio.
 matching those resources to the required resources for an application.
 (The Portable Operating System Interface (POSIX) is a family
of standards specified by the IEEE Computer Society for maintaining
compatibility between operating systems.)
 The SCA is built upon a standard set of OS features called POSIX.
 which also has standardized APIs to perform OS functions such as file
management and computational thread/task scheduling
  In computing, scheduling is the method by which work is assigned to
resources that complete the work.
 The work may be virtual computation elements such
as threads, processes or data flows,
 which are in turn scheduled onto hardware resources such
as processors.
 a thread of execution is the smallest sequence of programmed
instructions that can be managed independently by a scheduler,
 which is typically a part of the operating system.
 The SCA also specifies a Common Object Request Broker Architecture
(CORBA) middleware,
 which is used to provide a standardized method for software objects to
communicate with each other, regardless of which processor they have
been installed on
 The SCA also provides a standardized method of defining the
requirements for each application, by eXtensible Markup Language
(XML).
 The XML helps to determine how to distribute and install the software
objects.
 A language is considered object-oriented if it is object-based and also
has the capability of polymorphism
 In computer science, an object can be a variable, a data structure, or
a function.
 These applications will have many reasons to interact with the Internet
as well as many local networks;
 therefore, it is also common to provide a collection of standardized radio
services, network services, and security services.

Top-Level component Interface