A Time-Frequency Domain Adaptive Impedance Matching Approach Based on Deep Neural Network

Frequent fluctuations of antenna impedance within the time-frequency domain of mobile devices require efficient and accurate impedance matching. In this work, we propose an adaptive impedance matching approach based on deep neural network (DNN). Using frequency, voltage standing wave ratio (VSWR) and peak voltages as inputs, this approach employs a DNN to directly output the matching parameters. Compared to traditional analytical methods, the proposed approach reduces the complexity of the detection circuit. That is because the proposed approach requires only the amplitude information of the radio frequency (RF) signal to correct mismatches. For a simulated planar inverted-F antenna, experimental results demonstrate that the proposed DNN-based approach achieves more efficient iteration-free matching compared to simulated annealing particle swarm optimization and genetic algorithms.