Gain Characteristics Estimation Of Heteromorphic Rfid Antennas Using Neuro-Space Mapping

Existing gain estimation methods for radio frequency (RF) antennas often rely on rigorous and expensive experimental facilities or are only implemented for classic structure. They perform limited application scopes. To address the challenges, this study provides an accessible method for gain estimation of heteromorphic RF identification (RFID) antennas. There are three main innovations in the proposed method. An estimation framework is proposed based on neuro-space mapping technique which effectively reduces the time consumption and avoids laborsome measurement processes. A diverse extraction integration strategy is designed for training data acquisition, to balance the estimation accuracy and the training data size. A new adaptive particle swarm optimiser embedded with scale elaboration strategy is developed, which tackles the approximation problem from the gain estimation model to the gain from high-fidelity simulations. The proposed method is tested by four types of RF antennas. Simulations results demonstrate the method possesses high accuracy and strong applicability.