Course Notes, Simulation of Communication Systems, Sharif, EE, Iman Gholampour, imangh@sharif.

edu , Fall 2011 Course Notes, Simulation of Communication Systems, Sharif, EE, Iman Gholampour, imangh@sharif.edu , Fall 2011

Course Notes, Simulation of Communication Systems, Sharif, EE, Iman Gholampour, imangh@sharif.edu , Fall 2011

Simulation of Communications Systems, 25178

Syllabus
1) 2) 3) 4) 5) 6) 7) 8) 9) 10) 11) 12) 13) 14) 15) Introduction to Simulation and Modeling Role of Simulation in Communications Systems Life Cycle Simulation Methodology Practical Issues in Simulation of Communication Systems Representation of Signals and Systems in Simulation Environment Modeling and Simulation of Communications Systems Elements Generation of Data Signals, Random Numbers and Processes Modeling and Simulation of Non-linearities Modeling and Simulation of Time Varying Systems Modeling & Simulation of Communication Channels (Waveform/Discrete Channels) Monte Carlo Methods Rare Events Simulation and Importance Sampling Acceleration in MC Methods Semi-Analytic Methods in Simulation of Communication Systems Advanced Simulation Techniques: Tail Extrapolation, pdf Estimators, Splitting Case Studies
1

Simulation of Communications Systems, 25178

Text Book and References
1) 2) 3) 4) 5) 6) 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. Simulating Wireless Communication Systems: Practical Models in C++, C. B. Rorabaugh, Prentice Hall, 2004, ISBN: 0-13-022268-2. Rare Event Simulation using Monte Carlo Methods, G. Rubino, B. Tuffin, John Wiley and Sons, 2009, ISBN: 978-0-470-77269-0. 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. Introduction to communication systems simulation, M. Schiff, Artech House, 2006, ISBN-10159693-002-0. Simulation of Communication Systems, Modeling, Methodology, and Techniques, M. C. Jeruchim, P. Balaban, K. S. Shanmugan, Cluwer Academic Publishers, 2nd Edition 2002, ISBN 0306-46267-2. Simulation Techniques, Models of Communications, Signals and Process, F.M. Gardner, J. D. Baker, John Wiley & Sons Inc. 1997, ISBN 0-471-51764-9. Contemporary Communication Systems Using Matlab and Simulink, J.G. Proakis, M Salehi, G. Bauch, CL-Engineering 2003, ISBN 0-534-40617-3. Telecommunications Breakdown, C. R. Johnson, Jr., W.A. Sethares, Prentice Hall, 2004, ISBN: 0131-43047-5. Algorithms for Communications Systems and their Applications, N. Benvenuto, John Wiley & Sons Inc. 2003, ISBN 0-470-84389-6. Selected papers and book chapters
2

7) 8) 9) 10) 11)

Course Notes, Simulation of Communication Systems, Sharif, EE, Iman Gholampour, imangh@sharif.edu , Fall 2011

Simulation of Communications System Systems s

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 (Telecommunications) Communications is a process of transferring information from one entity to another System: a group of independent but inter-related elements comprising a unified whole

Etymology: 14th -17th Century

Middle English Middle English Old French Greek Latin

Simulation: Similar, like Communication: make it common, share locally Information: from Informing, Giving Shape to Mind System: animal body as an organized whole, sum of the vital processes in an organism
4

Course Notes, Simulation of Communication Systems, Sharif, EE, Iman Gholampour, imangh@sharif.edu , Fall 2011 Course Notes, Simulation of Communication Systems, Sharif, EE, Iman Gholampour, imangh@sharif.edu , Fall 2011

Course Notes, Simulation of Communication Systems, Sharif, EE, Iman Gholampour, imangh@sharif.edu , Fall 2011

Communication Systems
Fidelity Complexity Spectral Efficiency Complexity Aspects
1) 2) 3) 4)

Models
First Step to Study a System Art of Modeling To Develop a Behavioral Model Model: An Abstraction of a Real System to Predict and Formulate the System Behavior
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)
Analytical Models (Measurement Models) 3) Simulation Models
1) 2)

Architecture Hostile Deployment Environment High Data Rates, High Quality Limited Bandwidth, Power, Size,

Complex Techniques for Modulation, Pulse Shaping, Source and Channel Coding, Interleaving, Equalization, Synchronization, Carrier Recovery,
5

Usually Continuous sampling/quantization Mostly Discrete

6

Course Notes, Simulation of Communication Systems, Sharif, EE, Iman Gholampour, imangh@sharif.edu , Fall 2011

Modeling Validation & Solution

1. 2. 3. 4. 5.

Simulation Models
Abstraction
Physical Entity
Many Assumptions

Modeling Validation Reexamining the Formulation of the problem Consistent Dimensionality of Math Expressions Varying the Input Checking the Output Retrospective Test Prospective Test

Accuracy

Continuous

Analytical Model
More!

Modeling Solution 1. Analytical 2. Numerical

7

Discrete

Simulation Model

Complexity
8

Course Notes, Simulation of Communication Systems, Sharif, EE, Iman Gholampour, imangh@sharif.edu , Fall 2011 Course Notes, Simulation of Communication Systems, Sharif, EE, Iman Gholampour, imangh@sharif.edu , Fall 2011

Course Notes, Simulation of Communication Systems, Sharif, EE, Iman Gholampour, imangh@sharif.edu , Fall 2011

Divide and Conquer

Input Block 1 System Output Block 2

Simulation Models
1. Static versus Dynamic Models State variables do not depend on time 2. Deterministic versus Stochastic Models State variables are fixed or non-random 3. Continuous versus Discrete Models State variables are defined in all times

Intermediate

Block 3

We might be interested in some intermediate parameters (signals/states), not all Means more abstraction
9

10

Course Notes, Simulation of Communication Systems, Sharif, EE, Iman Gholampour, imangh@sharif.edu , Fall 2011

Simulation Roles
System behavior and life cycle predictions Parametric studies What-if questions Design: trade-off studies bit-true validation Performance evaluation Measurements, test procedures Rare conditions/cases Graphical view of signals Comparisons Deployment anomaly investigation
11

Communications Systems Design

Simulation appears in many phases!
From Design to Deployment Design Trade-Off Studies Parameter Optimization Performance Evaluation Establishing Test Procedures Benchmarks End of Life Prediction Anomaly Investigation
12

Course Notes, Simulation of Communication Systems, Sharif, EE, Iman Gholampour, imangh@sharif.edu , Fall 2011 Course Notes, Simulation of Communication Systems, Sharif, EE, Iman Gholampour, imangh@sharif.edu , Fall 2011

Course Notes, Simulation of Communication Systems, Sharif, EE, Iman Gholampour, imangh@sharif.edu , Fall 2011

D&D Process Communications Systems

Statement/Analysis of user requirements and performance expectations (MRD) System engineering System design (blueprint) Implementation and testing of key components Completion of HW prototype Validation of simulation model End of life prediction
13

Different Aspects, Knowledge

DSP Communications System Theory Numerical Analysis/ Number Theory Probability Theory, Stochastic Processes Estimation Computer Science
Either used in the system or concepts help in simulation
14

Course Notes, Simulation of Communication Systems, Sharif, EE, Iman Gholampour, imangh@sharif.edu , Fall 2011

Simulation SW Packages
1. 2. 3. 4. 5.

Simulation Methodology
Simulation Steps: 1) Quantitative
Science of simulation

Model Builder Model Library Model User Interface (may be a GUI) Simulation Kernel: data driven, time driven, event driven Postprocessors

2) Qualitative
Methodology or the art of simulation - Basic Purpose of Com Systems: Process Waveforms and Symbols - Simulation of Com Systems: Generating and processing of the sampled values Fundamental Simulation Steps: 1) 2) 3) 4) Mapping the problem into a simulation model Decomposing the problem into a set of smaller ones Selecting appropriate set of techniques to solve sub-problems Combining the sub-problems solutions to solve the main one
16

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 Notes, Simulation of Communication Systems, Sharif, EE, Iman Gholampour, imangh@sharif.edu , Fall 2011 Course Notes, Simulation of Communication Systems, Sharif, EE, Iman Gholampour, imangh@sharif.edu , Fall 2011

Course Notes, Simulation of Communication Systems, Sharif, EE, Iman Gholampour, imangh@sharif.edu , Fall 2011

Problem Mapping Techniques (comments on 1)

Generic Theme
Start 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 C) Simplifications (approximations/assumptions)

Hierarchical Representation
Vertical, 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
AM Demodulator

Managing the complexity in two directions Vertical: Layers Horizontal: Partitions

Envelope Detector

Amp

Non linearity

LPF

L1 L2 L3 L4

System Sub-systems Components Physical

17

18

Course Notes, Simulation of Communication Systems, Sharif, EE, Iman Gholampour, imangh@sharif.edu , Fall 2011

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

Simplification
Too much details in the block diagram Omission of blocks with no significant impacts Approximations (linearity, time-invariance, )
Example: Quasi static cases

Conditioning
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} da
19

Combining Blocks When intermediate details are not important

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

20

Course Notes, Simulation of Communication Systems, Sharif, EE, Iman Gholampour, imangh@sharif.edu , Fall 2011 Course Notes, Simulation of Communication Systems, Sharif, EE, Iman Gholampour, imangh@sharif.edu , Fall 2011

Course Notes, Simulation of Communication Systems, Sharif, EE, Iman Gholampour, imangh@sharif.edu , Fall 2011

Modeling of Individual Blocks (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 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

Modeling of Individual Blocks

Generic Methods
Interface to Other Blocks
Consistency, Compatibility between blocks Well-defined and Well documented interfaces Probable Problems: Inconsistency in different domains of processing Signal types Block size Step size Inconsistency of parameters specification in different blocks
22

Course Notes, Simulation of Communication Systems, Sharif, EE, Iman Gholampour, imangh@sharif.edu , Fall 2011

Modeling of Individual Blocks

Generic Methods
Choosing the Sampling Rate
LP-equivalent BW x 2 Unlimited: Analog (3dbBW x 8 to16), Digital (Symbol rate x 8to16)

Selecting appropriate (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
Input waveform Block Output waveform

Block processing, one or N sample per invocation

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 Processing

Multi-rate Sampling, Buffering, Interpolation/Decimation

From Library + Setting Parameters Generation of random numbers and waveforms

Parameterization, for design optimization

external knobs, visible from outside
23

Analysis (+ analysis parameters): Inline: Estimation during Simulation Offline: Estimation after Simulation
24

Course Notes, Simulation of Communication Systems, Sharif, EE, Iman Gholampour, imangh@sharif.edu , Fall 2011 Course Notes, Simulation of Communication Systems, Sharif, EE, Iman Gholampour, imangh@sharif.edu , Fall 2011

Course Notes, Simulation of Communication Systems, Sharif, EE, Iman Gholampour, imangh@sharif.edu , Fall 2011

Combining (comments on 4)
Supporting Interconnections requires some methods
Block1 Block2 Block3

Random Process Modeling and Simulation

Generating input waveforms, noise, interference to drive the simulation models
All are random in Nature Need some fidelity measures 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

Re-sampling, format, type, space, system theory tools We also need validation: 1) Analytically if possible (Theory) 2) Based on some measurements (Experiment) 3) Based on Intuition (Engineering vision)
25

Course Notes, Simulation of Communication Systems, Sharif, EE, Iman Gholampour, imangh@sharif.edu , Fall 2011

Performance Estimation
Performance Measures
Analog: SNR, Digital BER

MATLAB / SIMULINK
Base of our examples and exercises Will learn how to use them efficiently MEX Functions Nested Functions GUI and more

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

Trade off: Accuracy and Simulation Run time Unbiased and Consistent Estimation? It is unbiased, if some conditions are met Variance reduction techniques
Bit Stream D BER Comparison Count errors
27

System

28