Conjointly active and passive modelings with deep neural networks as fully automated optimizations for upper-mid band 6G communications

结合主动和被动建模以及深度神经网络,实现中高频段 6G 通信的全自动优化

阅读:1

Abstract

Today wireless systems include the fifth and sixth generations (5G and 6G) technologies and are growing day by day that result in exponentially increasing data traffic. For providing a reliable and high performance radio frequency (RF) designs especially for 6G networks, amplifiers and antenna as active and passive components play important roles. In the 5G/6G communication systems, the propagation loss is considerably large and its compensation requires high output power generated from the amplifiers for guaranteeing the satisfied quality of transmitted signal. From another point of view, the installed antennas must be able to optimally manage the radiated signals and handle/compensate nonlinear performances of the RF circuitry. Hence, advanced modeling and multi-objective optimization algorithms are required for designing and optimizing high performance amplifiers and antennas in terms of output power, gain, efficiency, linearity, and bandwidth. Concurrently optimizing active and passive components is not straightforward and typically it requires additional efforts by the RF designers. To tackle this drawback, a two-step methodology is proposed: (1) configuring the initial structure of active and passive devices, and (2) sizing the configured devices. In this work, various methods are introduced for structuring the topology of circuits and then artificial intelligence, including machine learning and neural networks, is preferred among other surrogate modelling for sizing the designs. These neural networks are satisfied due to the accurate modeling responses and are able to provide an automated optimization process leads to employ multi-objective optimization methods. In this work, an automated optimization process for comprehensive design of high-performance amplifiers with antennas through bottom-up optimization (BUO) method and long short-term memory (LSTM)-based deep neural networks (DNNs) is proposed. At the output layer of DNNs, the multi-objective multi-verse optimizer (MOMVO) method is employed for optimizing various specifications of active device (i.e., amplifier), and passive device (i.e., antenna), concurrently. In the presented method, all the electromagnetic (EM) design rules are implemented which results in reducing simulation time in the harmonic balance simulation environment that also provides ready to fabricate layouts. The novelty consists of the all-inclusive style that (1) reduces the manual breaks, aka time-to-market, and (2) delivers ready-to-fabricate layouts of the device that exhibits global optimum performances, automatically. The validation of the proposed method is verified by designing and optimizing high power amplifier (HPA) with antenna in the frequency band from 9.0 GHz to 9.6 GHz, suitable for upper-mid band 6G communications.

特别声明

1、本页面内容包含部分的内容是基于公开信息的合理引用;引用内容仅为补充信息,不代表本站立场。

2、若认为本页面引用内容涉及侵权,请及时与本站联系,我们将第一时间处理。

3、其他媒体/个人如需使用本页面原创内容,需注明“来源:[生知库]”并获得授权;使用引用内容的,需自行联系原作者获得许可。

4、投稿及合作请联系:info@biocloudy.com。