Localization and Object Recognition Using Inductive Resonance Sensors

Industrijski induktivni senzori za mjerenje prisutnosti metalnih objekata (engl. proximity sensors) daju izlazni signal čija amplituda ovisi o udaljenosti točno definiranog metalnog objekta. Tema ovog diplomskog rada je razvoj sustava koji omogućuje prepoznavanje i praćenje trajektorije gibanja nepoznatog metalnog objekta te raspoznavanje vrste metala od kojeg se objekt sastoji. Sustav se sastoji od tri titrajna kruga smještena u prostoru u kojem se kreće metalni objekt i mijenja im karakteristike. Korištenjem programskog paketa Matlab razvijeni su algoritmi za amplitudnu i frekvencijsku demodulaciju signala izmjerenih na titrajnim krugovima, algoritam koji na temelju amplitude i frekvencije lokalizira objekt u prostoru detekcije i pridružuje trajektoriju gibanja u vremenu, te algoritam koji prepoznaje vrstu metala. Nakon razvoja u Matlabu, algoritmi obrade signala, prepoznavanja i lokalizacije objekta pripremljeni su za obradu u realnom vremenu na ugradbenom računalnom sustavu ARM Cortex M arhitekture.Industrial inductive sensors for measuring the presence of metal objects (english proximity sensors) give an output signal whose amplitude depends on the distance of a precisely defined metallic object. The topic of this thesis is the development of a system that enables the identification and monitoring of the trajectory of the motion of an unknown metal object and the recognition of the type of metal from which the object is composed. The system consists of three oscillating circuits located in the space in which the metal object moves and changes its characteristics. By using the Matlab software package, algorithms for amplitude and frequency demodulation of signals measured on oscillating circuits, an algorithm that localizes the object in the detection area based on the amplitude and frequency and joins the trajectory of motion in time, and an algorithm that recognizes the type of metal have been developed. After development in Matlab, algorithms for signal processing, recognition and localization of the object are prepared for the real-time processing on the embedded computer system ARM Cortex M architecture