Random signals – “noises” – aren’t necessarily hindrances to be eliminated, they can carry information about the examined system. They can also play a constructive role – optimal functioning of some systems are made only possible by appropriate noise application. In the dissertation results are presented in areas which are examples of utilising noises in a constructive role or as an information source.

The subject of the first half of the thesis is the analysis of the Kirchhoff-Law-Johnson-Noise (KLJN) secure key exchange protocol. First, the necessary and sufficient conditions of noise properties for unconditional security are deducted using the tools of mathematical statistics only, giving a mathematical proof for the system’s perfect security. Next, the generalization of the protocol is presented, allowing the two communicating parties to use different hardware, i.e. resistors with different values. This result not only makes the practical application of the protocol much easier, but resulted in the reinterpretation of the classical physical description of the original KLJN protocol’s security . Finally the supplement of the generalized protocol is presented, in which the components previously bringing non-ideality and information leakage into the system became a part of the unconditionally secure ideal system, which is evidently a big step forward for the protocol’s practical applications.

Thereafter a new field of application for using fluctuations as an information source is shown. The presented results about analyzing kayak paddlers’ motion signals pointed out that the quality of the paddling is correlated to the fluctuation of the period and stroke impulse, which characterise the period of the motion. Thus the temporal indicators characterizing the period fluctuations and the spectral indicators based on the raw motion signals’ signal-to-noise ratio could contain extra information. The latter method of spectral variability analysis could be useful for other periodic signals as well.