simulation ooff century communications system...

Simulation of Simulation of Communications Systems, 25178Communications Systems, 25178

Syllabus

1) Introduction to Simulation and Modeling

2) Role of Simulation in Communications Systems Life Cycle

3) Simulation Methodology

4) Practical Issues in Simulation of Communication Systems

5) Representation of Signals and Systems in Simulation Environment

6) Modeling and Simulation of Communications Systems Elements

7) Generation of Data Signals, Random Numbers and Processes

Course N

otes, Simulation of C

ommunication System

s, Sharif,

1

7) Generation of Data Signals, Random Numbers and Processes

8) Modeling and Simulation of Non-linearities

9) Modeling and Simulation of Time Varying Systems

10) Modeling & Simulation of Communication Channels (Waveform/Discrete Channels)

11) Monte Carlo Methods

12) Rare Events Simulation and Importance Sampling Acceleration in MC Methods

13) Semi-Analytic Methods in Simulation of Communication Systems

14) Advanced Simulation Techniques: Tail Extrapolation, pdf Estimators, Splitting …

15) Case Studies

Course N


ommunication System

s, Sharif, EE, Iman G

holampour, im

[email protected] , Fall 2011

Simulation of Simulation of Communications Systems, 25178Communications Systems, 25178

Text Book and References

1) “Principles of Communication Systems Simulation with Wireless Applications”, W. H. Tranter, K. S. Shanmugan, T. S. Rappaport, K. L. Kosbar, Prentice Hall, 2004, ISBN 0-13-494790-8.

2) “Simulating Wireless Communication Systems: Practical Models in C++”, C. B. Rorabaugh, Prentice Hall, 2004, ISBN: 0-13-022268-2.

3) “Rare Event Simulation using Monte Carlo Methods”, G. Rubino, B. Tuffin, John Wiley and Sons, 2009, ISBN: 978-0-470-77269-0.

4) “Modelling the Wireless Propagation Channel: A simulation approach with Matlab”, F. P Fontan, P. M. Espineira, John Wiley and Sons, 2008, ISBN: 978-0-470-72785-0.

5) “Introduction to communication systems simulation”, M. Schiff, Artech House, 2006, ISBN-101-59693-002-0.

Course N


ommunication System

s, Sharif,

2

5) “Introduction to communication systems simulation”, M. Schiff, Artech House, 2006, ISBN-101-59693-002-0.

6) “Simulation of Communication Systems, Modeling, Methodology, and Techniques”, M. C. Jeruchim, P. Balaban, K. S. Shanmugan, Cluwer Academic Publishers, 2nd Edition 2002, ISBN 0-306-46267-2.

7) “Simulation Techniques, Models of Communications, Signals and Process”, F.M. Gardner, J. D. Baker, John Wiley & Sons Inc. 1997, ISBN 0-471-51764-9.

8) “Contemporary Communication Systems Using Matlab and Simulink”, J.G. Proakis, M Salehi, G. Bauch, CL-Engineering 2003, ISBN 0-534-40617-3.

9) “Telecommunications Breakdown”, C. R. Johnson, Jr., W.A. Sethares, Prentice Hall, 2004, ISBN: 0-131-43047-5.

10) “Algorithms for Communications Systems and their Applications”, N. Benvenuto, John Wiley & Sons Inc. 2003, ISBN 0-470-84389-6.

11) Selected papers and book chapters

Course N


ommunication System


holampour, im


SimulationSimulation of of Communications Communications SystemSystemss

�� Simulation Simulation is the act of imitating the behavior of some situation

or some process by means of something suitably analogous

In computer science, the technique of representing the real world by a computer program

Course N


ommunication System

s, Sharif,

3

�� ((Telecommunications) Communications Telecommunications) Communications is a process of is a process of transferring transferring informationinformation from one entity to anotherfrom one entity to another

�� System: System: a group of independent but inter-related elements

comprising a unified whole

Course N


ommunication System


holampour, im


Etymology: 14Etymology: 14thth --1717thth CenturyCentury

Middle English � Old French� Latin

Middle English Middle English �� GreekGreek

�� Simulation: Similar, likeSimulation: Similar, like

Course N


ommunication System

s, Sharif,

�� Communication: make it common, share locallyCommunication: make it common, share locally

�� Information: from Informing, Giving Shape to MindInformation: from Informing, Giving Shape to Mind

�� System: System: animal body as an organized whole, sum of the

vital processes in an organism

4

Course N


ommunication System


holampour, im


Communication SystemsCommunication SystemsFidelity ‹–› Complexity ‹–› Spectral Efficiency

Complexity Aspects

1) Architecture

2) Hostile Deployment Environment

3) High Data Rates, High Quality

Course N


ommunication System

s, Sharif,

3) High Data Rates, High Quality

4) Limited Bandwidth, Power, Size, …

Complex Techniques for Modulation, Pulse Shaping, Source

and Channel Coding, Interleaving, Equalization,

Synchronization, Carrier Recovery, …

5

Course N


ommunication System


holampour, im


ModelsModels

First Step to Study a System … Art of ModelingTo Develop a Behavioral Model

Model: An Abstraction of a Real System to Predict andFormulate the System BehaviorCaptures the in/out behavior of the system under specific conditions

Course N


ommunication System

s, Sharif,

6

Captures the in/out behavior of the system under specific conditions

Often Mathematical (Formulas, Relations, Logic)

Physical Systems Translate to Mathematical Systems thru Models

Accuracy versus Simplicity (Modeling Trade-off)

1) Analytical Models � Usually Continuous

2) (Measurement Models) sampling/quantization

3) Simulation Models � Mostly Discrete

Course N


ommunication System


holampour, im


Modeling Validation & SolutionModeling Validation & Solution

� Modeling Validation

1. Reexamining the Formulation of the problem

2. Consistent Dimensionality of Math Expressions

3. Varying the Input Checking the Output

Course N


ommunication System

s, Sharif,

4. Retrospective Test

5. Prospective Test

� Modeling Solution

1. Analytical

2. Numerical 7

Course N


ommunication System


holampour, im


Simulation ModelsSimulation Models

Physical Entity

Abstraction Accuracy

Course N


ommunication System

s, Sharif,

Many Assumptions

8

Analytical Model

Simulation Model

Complexity

Course N


ommunication System


holampour, im


Continuous

Discrete

More!

Divide and ConquerDivide and Conquer

System

Block1

Block2

Input

Output

Course N


ommunication System

s, Sharif,

9

2

Block3

We might be interested in some intermediate parameters

(signals/states), not all

Means more abstraction

Intermediate

Course N


ommunication System


holampour, im


Simulation ModelsSimulation Models

1. Static versus Dynamic Models

State variables do not depend on time

2. Deterministic versus Stochastic Models

Course N


ommunication System

s, Sharif,

State variables are fixed or non-random

3. Continuous versus Discrete Models

State variables are defined in all times

10

Course N


ommunication System


holampour, im


Simulation RolesSimulation Roles

� System behavior and life cycle predictions� Parametric studies� What-if questions� Design: trade-off studies � bit-true validation� Performance evaluation

Course N


ommunication System

s, Sharif,

� Performance evaluation� Measurements, test procedures� Rare conditions/cases� Graphical view of signals� Comparisons� Deployment anomaly investigation

11

Course N


ommunication System


holampour, im


Communications Systems DesignCommunications Systems Design

Simulation appears in many phases!From Design to Deployment

� Design Trade-Off Studies

� Parameter Optimization

Course N


ommunication System

s, Sharif, � Parameter Optimization

� Performance Evaluation

� Establishing Test Procedures

� Benchmarks

� End of Life Prediction

� Anomaly Investigation

12

Course N


ommunication System


holampour, im


D&D Process D&D Process Communications SystemsCommunications Systems

� Statement/Analysis of user requirements and performance expectations (MRD)

� System engineering� System design (blueprint)

Course N


ommunication System

s, Sharif, � System design (blueprint)� Implementation and testing of key components

� Completion of HW prototype� Validation of simulation model� End of life prediction

13

Course N


ommunication System


holampour, im


Different Aspects, Knowledge

� DSP� Communications System Theory� Numerical Analysis/ Number Theory� Probability Theory, Stochastic Processes

Course N


ommunication System

s, Sharif, � Probability Theory, Stochastic Processes� Estimation� Computer Science…

Either used in the system or concepts help in simulation

14

Course N


ommunication System


holampour, im


Simulation SW Packages1. Model Builder2. Model Library3. Model User Interface (may be a GUI)4. Simulation Kernel: data driven, time driven, event driven5. Postprocessors

Course N


ommunication System

s, Sharif,

Choose/Build the SIM models (1,2)Add the SIM Parameters (3)Choose Design Parameters (3)Simulation Stop/End/Completion (4)Post-Processing (5) : Display (Waveform plot, Spectral Plot, Scatter Plot,

Eye Diagram, …), Analysis, …

Simulation � Low level C/C++ � Bit-True C � HDL or ASM

15

Course N


ommunication System


holampour, im


Simulation MethodologySimulation Steps:

1) Quantitative

Science of simulation

2) Qualitative

Methodology or the art of simulation

- Basic Purpose of Com Systems: Process Waveforms and Symbols

Course N


ommunication System

s, Sharif, - Basic Purpose of Com Systems: Process Waveforms and Symbols

- Simulation of Com Systems: Generating and processing of the sampled values

Fundamental Simulation Steps:

1) Mapping the problem into a simulation model

2) Decomposing the problem into a set of smaller ones

3) Selecting appropriate set of techniques to solve sub-problems

4) Combining the sub-problems solutions to solve the main one 16

Course N


ommunication System


holampour, im


Problem Mapping Techniques Problem Mapping Techniques (comments on 1)(comments on 1)

Generic ThemeStart with “clear statement of the problem”

Include everything that you can think of in the initial block diagram

A) Hierarchical Representation

B) Partitioning and Conditioning

Course N


ommunication System

s, Sharif, B) Partitioning and Conditioning

C) Simplifications (approximations/assumptions)

Managing the complexity

in two directions

Vertical: Layers

Horizontal: Partitions

17

L1 System

L2 Sub-systems

L3 Components

L4 Physical

Course N


ommunication System


holampour, im


Hierarchical RepresentationVertical, Layers

Back annotation: create higher level models from details of lower layer model and replace…

Co-simulation: using a separate simulation to prepare higher layer model

Course N


ommunication System

s, Sharif,

18

Envelope Detector

AM Demodulator

AmpNon

linearity LPF

Course N


ommunication System


holampour, im


Partitioning and Conditioning

� Partitioning

Separating the task in the same layer of abstraction

Main complex problem � A set of interrelated/independent problems � Simulate separately and combine

� Conditioning

Fix the condition or state of a portion of the system

Course N


ommunication System

s, Sharif, Fix the condition or state of a portion of the system

Simulate the rest

Repeat for different states and conditions

(parts are simulated separately)

Main results derived by averaging

f(a,b) = f(a) f(b|a) f(a)� SIM1 f(b|a)� SIM2

E[g(A,B)] = ∫∫ g(a,b) f(a,b)da db= ∫f(a){∫g(a,b)f(b|a) db} da19

Course N


ommunication System


holampour, im


SimplificationSimplification

Too much details in the block diagram� Omission of blocks with no significant impacts� Approximations (linearity, time-invariance, …)

Example: Quasi static cases

� Combining Blocks

Course N


ommunication System

s, Sharif,

� Combining BlocksWhen intermediate details are not importantExample: Performance estimation

In AM detector: For SNR calculation no need to consider the filter circuit.

20

Course N


ommunication System


holampour, im


Modeling of Individual Blocks Modeling of Individual Blocks (comments on 2)(comments on 2)

Generic Block:{ y[k], y[k-1], …, y[k-m] } =

F{x[k-j], x[k-j-1], …, x[k-j-n]; k; p1, …pq}

Producing m samples per invocation

‘F’ independent of k ~ time invariant

Course N


ommunication System

s, Sharif, ‘F’ independent of k ~ time invariant

m>0 ~ block input/output

m=0 ~ sample by sample model

n=0 ~ memory-less

‘F’ linear or nonlinear

Band limited or unlimited

Time Domain or Frequency Domain or Transform Domain

21

Course N


ommunication System


holampour, im


Modeling of Individual Blocks…Modeling of Individual Blocks…

Generic Methods…� Interface to Other Blocks

Consistency, Compatibility between blocks

Well-defined and Well documented interfaces

Probable Problems:

Course N


ommunication System

s, Sharif, Probable Problems:

Inconsistency in different domains of processing

� Signal types

� Block size

� Step size

Inconsistency of parameters specification in different blocks

…………

22

Course N


ommunication System


holampour, im


Modeling of Individual Blocks…Modeling of Individual Blocks…

Generic Methods…� Choosing the Sampling RateLP-equivalent BW x 2

Unlimited: Analog (3dbBW x 8 to16), Digital (Symbol rate x 8to16)

� Block processing, one or “N” sample per invocation

“N>1”: Efficient when invocation overhead is large

Course N


ommunication System

s, Sharif, “N>1”: Efficient when invocation overhead is large

delay of NT, complicated when non-linearity and/or feedback, needs scheduling when different blocks have different “N”

� Variable Step-Size ProcessingMulti-rate Sampling, Buffering, Interpolation/Decimation

� Parameterization, for design optimization“external knobs”, visible from outside

23

Course N


ommunication System


holampour, im


Selecting appropriate … Selecting appropriate … (comments on 3)(comments on 3)

Selecting an appropriate set of modeling simulation and estimation techniques to solve sub problems

Can be rigorous, algorithmic, well defined

Or “Tricks of the trade”

Course N


ommunication System

s, Sharif, Input waveform Output waveform

24

Course N


ommunication System


holampour, im


Block

Generation of random numbers and waveforms

Analysis (+ analysis parameters): Inline: Estimation during SimulationOffline: Estimation after Simulation

From Library + Setting Parameters

Input waveform Output waveform

Combining … Combining … (comments on 4)(comments on 4)

Supporting Interconnections requires some methods

Re-sampling, format, type, space, system theory tools

Course N


ommunication System

s, Sharif,

Block3Block2Block1

We also need validation:

1) Analytically if possible (Theory)

2) Based on some measurements (Experiment)

3) Based on Intuition (Engineering vision)

25

Course N


ommunication System


holampour, im


Random Process Modeling and SimulationRandom Process Modeling and Simulation

Generating input waveforms, noise, interference to drive the simulation models

All are random in Nature

Need some fidelity measures

Course N


ommunication System

s, Sharif,

Methodology

� Gaussian approximation (Central limit theorem)

� Equivalent Process Representation

� Slow versus Fast Processes◦ Moderately different: Down-sampling/up-sampling

◦ Highly different: Partition and condition on the slow process

26

Course N


ommunication System


holampour, im


Performance EstimationPerformance Estimation

� Performance MeasuresAnalog: SNR, Digital BER

� Monte Carlo Techniques: Pe ≈ Ne/N, N>>1

Trade off: Accuracy and Simulation Run time

Unbiased and Consistent Estimation?

Course N


ommunication System

s, Sharif,

Unbiased and Consistent Estimation?

It is unbiased, if some conditions are met

Variance reduction techniques

27

System

D

ComparisonCount errors

Bit Stream

BER

Course N


ommunication System


holampour, im


MATLAB / SIMULINKMATLAB / SIMULINKCourse N


ommunication System

s, Sharif,

Base of our examples and exercisesWill learn how to use them efficiently …

• MEX Functions• Nested Functions• GUI• and more…

28

Course N


ommunication System


holampour, im


simulation ooff century communications system...

Documents