Two Papers Submitted to ICUAS 2026

We are excited to announce the submission of two papers to the 2026 International Conference on Unmanned Aircraft Systems (ICUAS):

Semantically-Informed Predictive Control for Autonomous UAV Connectivity in Urban Environments

Evangelos Vlachos and Dimitrios Tsourounis

mmWave UAV links suffer in urban environments because radar cannot tell concrete from vegetation — both attenuate the signal, but only concrete truly blocks it. Our framework fuses 60 GHz radar with a YOLO-based semantic classifier to build an Effective Capacity field (C_raw × reliability) and plan through it via a two-layer controller:

1. SI-RRT — an RRT backbone that rejects candidate waypoints in multipath traps and low-reliability zones before they enter the tree.
2. Gradient Planner — deforms the backbone by ascending ∇C_eff, simultaneously signal-seeking and reliability-seeking through the product-rule decomposition.

Monte Carlo results over 100 independent urban canyon environments show +65% effective throughput vs. signal-strength baselines, 0.96 link reliability, and 97.3% of an oracle with perfect global channel knowledge.

Hardware-Aware SE(3) Control Barrier Functions for Counter-UAS Interceptors with Directed Energy Payloads

Evangelos Vlachos, Panayiotis Kolios, and Christos Skliros

This paper addresses the safety-critical control problem of keeping a directed-energy phased-array payload’s boresight aimed at a target UAS during high-speed interception. We develop SE(3) Control Barrier Functions that enforce a field-of-view constraint on the interceptor’s attitude while respecting the physical limits of the onboard hardware, solved efficiently via ADMM at high update rates.

Both works are supported by the EU Horizon Europe project EUSOME (Grant Agreement No. 101187121).