This paper presents a universal random signal generator capable of generating differently distributed random signals such as the uniform distribution and the normal distribution with a single apparatus.
This generator is composed of two essential parts: The first is shift register whose state represented by a vector a is varied irregularly by transmitting the sequence of binary random numbers. The second is the weighted adder which carries out an inner product of the vector a and a weight vector w.
It is theoretically suggested that the probability density and the frequency characteristic of the resulting random signal are related to the weight vector and the probability that ‘1’ of the binary random numbers will happen. And the several weight vectors to generate the random signal having the typical probability density are obtained.
The experiments show that the theoretical suggestion is sound, and that the generator whose weights are properly adjusted generates the random signal having a desired probability density.