A Research On Universal Algorithms For Generating Continuous Random Variable And Correlated Stochastic Series

Random number and correlated stochastic series have great value in simulation area. To design a gaussian random number generator used for channel simulation and coder performance test, high-speed universal random number and correlated stochastic series generators are studied.Continuous random variable generating algorithms are studied in chapter 2. Existing universal algorithms are introduced. Three algorithms are proposed. The first one exactly generates random numbers of desired distribution with fairly high speed, and requires acceptable hardware resources. The second approximately generates desired numbers with adjustable approximation precise, with higher speed than the first, and requiring few hardware resources. The third has better approximation performance than the second but is slower, requiring a little more resources. A gaussian noise generator implemented with FPGA is also presented.Correlated stochastic series generating algorithms are studied in chapter 3. Existing universal algorithms are introduced. Three algorithms are proposed. The first is a probabilistic model; its output process has the same one-dimensional distribution as input process and has correlation structure as normal AR and MA model; its speed is quite high; its hardware implementation is quite simple. The second uses colored gaussian noise as input, generates process with similar correlation function as input, requires a little more resources. The third improves the second and achieves better performance at the cost of more complexity.The proposed algorithms together meet with a good many applications.