Revealing the Hidden Blind Spots in UAV Communications

William Bjorndahl

PhD Candidate, Southern Methodist University

William Bjorndahl, a Ph.D. researcher at Southern Methodist University working under the supervision of Dr. Joseph Camp, is advancing how wireless communication is modeled for Unmanned Aerial Vehicles (UAVs). Modern UAVs rely on increasingly complex antenna systems, yet researchers often approximate their radiation patterns as simple spheres or static shapes. In real-world flight, however, drone tilt, banking, and airframe obstruction introduce dynamic “nulls,” or dead zones, in signal coverage that are difficult to predict.

Accurately modeling these effects typically requires collecting massive amounts of measurement data, making the process both time-consuming and expensive for each new drone platform. To address this challenge, Bjorndahl and his team developed SPARK (Sparse Parametric Antenna Representation using Kernels), a novel approach that reconstructs highly accurate 3D antenna radiation patterns using only sparse measurements. Instead of relying on thousands of data points, SPARK leverages kernel-based representations to capture complex antenna behavior efficiently, enabling scalable and realistic modeling of UAV communication systems.

AERPAW played a central role in enabling and validating this work by providing both controlled and real-world experimental data. High-fidelity 3D antenna radiation patterns measured in an anechoic chamber through the AERPAW testbed were used to evaluate the accuracy of the SPARK model, while additional datasets from AERPAW supported ray tracing over UAV flight trajectories to generate corresponding real-world measurements and simulation data. By combining these resources, the team was able to bridge the gap between laboratory measurements and dynamic flight conditions, demonstrating that accurate antenna models can be achieved with significantly reduced data collection. This work represents an important step toward more realistic and efficient modeling of UAV wireless systems, reducing barriers to experimentation and accelerating innovation in next-generation aerial communication networks.

The research is reflected in several forthcoming publications, including “BeamMix: 3D Gaussian Mixture-of-Experts for Element-Space Wireless Channel Modeling” (ACM MobiSys 2026), “SPARK: Sparse Parametric Antenna Representation using Kernels” (IEEE INFOCOM 2026), and “PropSplat: Map-free RF Field Reconstruction via 3D Gaussian Propagation Splatting” (IEEE DySPAN 2026).