The contemporary advance of technology is primarily due to significant development of artificial intelligence. The aim of this work is to design a system, which uses the aspects of both machine learning and artificial intelligence. Such a system will be implemented in MATLAB environment and experimentally verified on a device called Flexy. Using aforementioned device, we are measuring sound and vibrational signals, which further undergo a series of algorithms to extract only useful parts of the signal. Based on measured values of physical parameters, the system creates a table of differences between the real values and the ones estimated by our system itself. Concluding from the experimental part of this work, one can say that the system works reliably and effectively. However, too slow manipulation with a potentiometer causes a signal distortion that is reflected on temporary increase of amplitude, until the power reaches steady state. Another challenge to be overcome is the background noise of the environment that could overlap the desired signal source, or several of these devices working in vicinity, which could lead to indication of wrong impulses.The outcome of this thesis is an automatized system that can estimate the parameters based on the analysis of their physical properties accurately enough. The principle can be used not only to replace expensive flowrate meters, but also as a security system against third party attacks via wireless networks. Based on the extent of deviation, a trained user should be able to detect a faulty operation of the device.