This Small Business Innovation Research (SBIR) Phase II project seeks to develop and launch a software tool that applies advanced AI algorithms to antenna design and optimization. Manual antenna design and optimization methods are time- and labor-intensive, limit complexity, and require significant expertise. Genetic algorithm (GA) optimization has demonstrated success at quickly finding effective antenna design solutions not ordinarily found through engineering intuition. To harness the power of these search algorithms currently, an engineer must be an expert in both GAs and Electromagnetics. In Phase I, feasibility of a highly-automated design approach where useful antennas can be generated without requiring significant guidance was demonstrated: the user simply inputs design requirements (e.g., RF performance, dimensions, etc.), and an automated optimization produces compliant designs.
If successful, this technology promises to improve the performance and economics of future antenna applications for commercial and government customers. The world is in the midst of an explosion in the number of new wireless, mobile, and RF systems - all of which rely on one or more antennas. Yet antenna design has changed little in the past two decades, with large up-front costs and slow, inefficient trial-and-error methods. X5 Systems is attempting to bring to market a next generation way to design antennas: one that is faster, better, cheaper. The commercial potential of the proposed software system encompasses application areas of interest to companies in mobile and wireless, RFID, and consumer electronics, as well as government agencies - especially applications that have exacting performance, schedule, and cost requirements.
The world is in the midst of an explosion in the number of new wireless, mobile, and RF systems – all of which rely on one or more antennas. Yet the antenna design process has changed little in the past two decades, with large up-front engineering costs and slow, inefficient trial-and-error methods. Through this Small Business Innovation Research (SBIR) Phase II project, X5 Systems brings to market a next-generation way to design antennas: one that is faster, better, cheaper – critical to mobile and wireless devices where small size, longer battery-life, and time-to-market considerations can make or break companies operating in these multi-billion dollar markets. Manual antenna design and optimization methods are time- and labor-intensive, limit complexity, and require significant expertise and investment in software tools and computers. Stochastic optimization algorithms and other Artificial Intelligence (AI) techniques have been extremely successful at quickly finding effective antenna design solutions not ordinarily found through engineering intuition. X5 Systems’ team’s work at MIT, Stanford, Carnegie Mellon, Air Force Research Laboratory, and NASA has been groundbreaking in this area, resulting in the seminal patent in the field and AI-designed antennas deployed on three NASA spacecraft in 2006. However, until now, to harness the power of these algorithms, an engineer has had to be an expert in both AI and Electromagnetics. In Phase I of this effort, the feasibility of a highly-automated design approach where useful antennas can be generated without requiring significant guidance was demonstrated: the user simply inputs design requirements (e.g., RF performance, size limits, etc.), and an automated optimization system produces compliant designs. In this Phase II effort, the tool was developed and matured to the point where it was able to be released for commercial use. The system developed in this effort is radically different from any other on the market, with an improved value proposition – it asks the user for his/her requirements and proceeds to automatically design one or more antennas for them. Instead of the current labor- and budget-intensive method of engineering in which designs can take months to develop, this software completes designs in minutes or hours, and can deliver counter-intuitive state-of-the-art antennas that match or exceed performance of human-designed antennas. This tool reduces project risk, allows what-if trade studies, and provides higher-quality, lower-cost designs in a field that is very cost and performance sensitive. This technology promises to greatly improve the performance and economics of future antenna applications for commercial and government customers. The commercial potential of the proposed software system encompasses many application areas of interest to companies in mobile and wireless, RFID, and consumer electronics, as well as government agencies – especially applications that have exacting performance, schedule, and cost requirements.
