Simulation & Modeling
Full-wave EM simulation, channel modeling, and RF system co-design.
Simulation is the substrate of antenna research. Before any structure is fabricated, before any prototype is tested in the field, it must be designed, validated, and optimized in a high-fidelity computational environment. Our Hangzhou lab operates a full simulation stack covering Maxwell's equation solvers, deterministic and stochastic channel models, and system-level co-simulation with AI control algorithms in the loop.
Full-wave EM Solvers
We use commercial and open-source solvers including HFSS, CST, and openEMS for antenna characterization across Sub-6GHz, mmWave, and emerging terahertz frequency bands. Method-of-moments, finite-element, and finite-difference time-domain techniques are deployed based on geometry complexity and frequency range.
Channel Modeling
Antennas live inside channels. Our channel models combine deterministic ray-tracing for site-specific propagation studies (factory floors, stadium interiors, urban canyons) with stochastic 3GPP-conformant models for statistical analysis. AI-driven channel models, learned from measurement data, increasingly complement these traditional approaches.
System Co-Design
The unique value of our simulation stack lies in tight integration with AI control algorithms. Neural beam management policies are trained against simulated channels with realistic antenna models, closing the loop between physics and learning earlier than traditional design flows allow.
- [1]Yee, K. (1966). Numerical Solution of Initial Boundary Value Problems Involving Maxwell's Equations. IEEE Trans. Antennas and Propagation, 14(3), 302–307.
- [2]3GPP TR 38.901 (2024). Study on Channel Model for Frequencies from 0.5 to 100 GHz.
- [3]Harrington, R. F. (1993). Field Computation by Moment Methods. IEEE Press.
- [4]Liu, Y., et al. (2024). Deep Learning for Wireless Channel Modeling: A Survey. IEEE Communications Surveys & Tutorials.