Simulation of Communication Systems - Winlab模拟通信系统的WINLAB.pptVIP

Simulation of Communication Systems Wireless Systems Instructional Design Simulation “hierarchy” Waveform Level Simulations Usually used when analytical evaluation of performance is difficult (nonlinearities, ISI caused by bandlimiting filters) Typically: Generate sampled values of the input waveforms (process) Process them through system models and generate output Estimate the performance by comparing inputs and outputs Methodology Ideally model is a perfect replica of the real system – hard to do Instead we introduce approximations to reduce complexity or run-time: Modeling level – simplification of the specific functions Performance evaluation level – estimation of performance measures Methodology (cont.) Modeling: System Modeling - highest level of description; complexity reduction Device Modeling – block or subsystem (e.g. transfer function – on every clock cycle: input-transfer-output”) Random Process Modeling: Source random process (imitated with pseudo random number generator – RNG) Time-variant random channel Equivalent random process (ERP) Methodology (cont.) Monte Carlo simulation – as the name implies relates to game of chance Input signals are assumed to be random processes Objective is to find statistical properties of Methodology (cont.) Methodology (cont.) For our example: where If only some input processes are simulated explicitly – partial MC (quasianalytical simulation) Random number generation is essential for MC simulations Requires RNG generation methods from a wide variety of distributions and with arbitrary autocorrelation (PSD). RNG Important properties: Algebraic Structure (uncorrelated samples) Period Statistical Distribution Uniform RNG Congruental or the power residue method RNG (cont) where M0 large (prime) integer - modulus 0aM - multiplier c {0,1} – increment 0 = X(0) = M-1 – seed Generates random sequence of integers 0= X(k) = M-1 Uniform RNG RNG (cont) “Few good men” (for 32 bit machines)