PROJECT & FUNDING SCHEMES

This project aims to develop an AI agent that addresses the shortage of experienced antenna engineers in the industry. The core innovation is an antenna expert model based on a retrieval-augmented generation (RAG) framework, which codifies prior design expertise (including Wi-Fi, 5G, UWB, RFID and LoRa) and relevant technical literature.
This enables a general-purpose large language model (LLM) to deliver expert-level
antenna designs.  The R&D methodology encompasses an automated workflow that integrates CAD and antenna simulation software, the development of the LLM with the antenna expert model, and a closed-loop design process. The AI agent analyzes antenna performance metrics.  If the performance falls short of specifications, the AI agent commands the CAD software to modify the antenna geometry based on the antenna expert model and then runs the simulation. This creates an iterative design that continuously refines the antenna design until it meets the target performance.
The system dramatically accelerates the design cycle—reducing it from months to days— by performing thousands of automated iterations. It also enables the creation of complex antenna geometries that comply with stringent constraints. Moreover, it empowers junior engineers to produce expert-level designs, effectively mitigating the industry's talent gap and strengthening competitiveness in the high-speed wireless communication market.