A Toolkit for Efficient Co-Simulation and Inverse-Design of Broadband Electronically Small Antennas

Presently, there are several challenges to realizing broadband ESAs in the real-world. The first and most critical of which is the limited bandwidth of passive ESA systems. The well-known Chu limit defines the inherent tradeoff between physical size and bandwidth one must Circuits make when considering antenna system miniaturization. This bandwidth limitation can be the Antenna ESA's inherent and highly dispersive reactance and/or its potentially low value of input resistance. However, the achievable real-world performance of impedance matched ESAs is highly dependent on the non-Foster circuit design which is, in turn, generally limited by conventional design approaches and the inability to perform a holistic inverse-design (or 'inverse co-design') procedure of the complete antenna and circuit system. Thus, the conventional approach is to isolate the antenna and circuit comprising the ESA as much as possible in order to reduce the simulation complexity (see Fig. 1). However, this choice often has a negative impact on system SWaP (size, weight, and power) and achievable antenna bandwidth enhancement. Therefore, considerable improvements to antenna and circuit co-simulation techniques, which can significantly reduce the computational cost (i.e., time and memory) of modeling broadband ESAs, are urgently needed in order to make inverse co-design a tractable part of the ESA development process. Moreover, pairing new solvers with emerging design synthesis techniques such as topology optimization can lead to previously unexplored antenna geometries whose dispersive behaviors can be tailored in conjunction with active circuitry to realize significant improvements in impedance matching bandwidth. Meanwhile, Deep Learning (DL) and Deep Neural Networks (DNNs) have the potential to accelerate stable NIC design by learning how to properly weight the values of the components within the non-Foster circuit. When combined with Multi-Objective Optimization (MOO) techniques, these developments will enable a new capability for engineers to efficiently design and optimize broadband ESAs and study their inherent gain, bandwidth, and SWaP tradeoffs. PSU-CEARL's expertise in these areas [1]–[6] uniquely positions us to address the extreme challenges mentioned above by developing a suite of modeling tools for new and disruptive broadband ESAs. The following technical writeup presents a more in-depth discussion of the present challenges in the modeling and optimization of broadband ESAs while providing insights to our proposed solutions. The proposed effort is anticipated to cost $100k over a 12-month period of performance.