China
Africa
Explore chapters and articles related to this topic
Overview of Research Status
Published in Jingchang Nan, Mingming Gao, Nonlinear Modeling Analysis and Predistortion Algorithm Research of Radio Frequency Power Amplifiers, 2021
At present, domestic and foreign scholars are carrying out the research on RF/microwave circuit modeling based on neural networks. Professor Q. J. Zhang, as the founder and leader of neural network modeling in the RF/microwave field, proposed and validated the feasibility of neural networks applied to microwave circuit simulation and statistical design. He has proposed the use of dynamic neural networks in behavioral modeling of RF circuits and engaged in the cutting-edge research on neuro-modeling and space mapping optimization, which are two major technologies in the development history of RF/microwave computer-aided design. There are also many research groups from domestic universities that have been carrying out similar researches in this field, such as Tsinghua University, Shanghai Jiao Tong University, Xidian University, Hangzhou Dianzi University, etc. The research results have been widely published in international authoritative journals, such as IEEE Transactions on Microwave Theory and Techniques, IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems.
Designing RF and Microwave Band Pass Filters Using Coupled Resonators
Published in IETE Journal of Education, 2021
To optimize the filter, Space Mapping (original/linear) is used where Coupling Matrix (extracted using ADS as shown in Figure 8) is used as Coarse Model and 3D filter itself is the Fine Model [1]. The Space Mapping technique establishes a mathematical link (mapping) between the spaces of the design parameters of the two models: the accurate and time-intensive fine model and the less accurate but fast coarse model. The aim is to avoid the computationally expensive calculations encountered in optimizing the filter structure using the time-intensive fine model [1]. Resonators are modelled as parallel LC circuits. IR and IO couplings are modelled using ABCD parameters shown as rectangular box between the resonators. The input to the Coarse Model (i.e. ADS circuit) is the S-parameter results of the Fine Model (i.e. HFSS model). The output from the Coarse Model is the coupling elements obtained through brute force optimization. The key point to be noted here is that the optimization is performed on the Coarse Model (which is fast) compared to the optimization in Fine Model (HFSS model, which is slow). Since there are five independent parameters (resonator 1 postheight, resonator 2 postheight, iris width between resonator 1 and resonator 2, iris width between resonator 2 and resonator 3, and SMA height at input coupling) to optimize, Space Mapping requires at least 5 + 1 preliminary simulations with random variations in the parameters. The filter is optimized in 11 iterations. The final dimensions of the filter are shown in Figure 9 and results in Figure 10.
Space mapping techniques for the optimal inflow control of transmission lines
Published in Optimization Methods and Software, 2018
In a space mapping framework a fine model optimization problem is supposed to be solved. This fine-scale model is very accurate, but direct optimization methods are either very time consuming or hard to apply. We assume that a simulation of the fine model is possible, even though we intend to avoid extensive usage. A second, easier model, the coarse model, resigns accuracy for the advantage of fast simulation and optimization. The idea of a space mapping algorithm is to execute a coarse model optimization in every iteration to find a descent direction for the current approximation of the fine model optimal solution.