Investigation on optimization of some implanted antennas with machine learning

Development in implantable electronic devices have been made a considerable progress in healthcare and biomedical technologies. Improving the healthcare quality by the possibility of continuously monitoring the physiological signals of different parts of human body, providing the support to an organ and drug delivery to name a few, in addition to the continuous miniaturization of electrical sensors, advancement in wireless communication and digital signal processing, has entered biomedical technology into the new era. The main part of the implantable devices which provides the wireless communication is the antenna. The performance of these antennas inside the body has faced with multiple challenges such as being in the lossy medium from radio propagation perspective, their size and material which should be compatible physically and chemically inside the human body. This thesis is focused on designing and optimizing the implanted antennas with taking advantage of Machine Learning techniques. To achieve this goal two different methods of machine learning which are regression and deep neural networks, known as DNNs, have been employed. Both methods are recommended as learning method for the optimization but for some challenges like overfitting and accuracy both have been considered. The reason of this decision is also discussed in the thesis.