Swarm-Based RF Measurement: Using Controlled drone network
Author
Schulz, Christoffer
Term
4. semester
Education
Publication year
2026
Abstract
This project explores whether a small drone swarm can find the strongest point of a directional antenna’s signal faster than traditional single-drone raster scans. Three RMTT drones fly in a triangular formation, communicate via ESP-NOW, and are tracked by the Vicon motion capture system at Aalborg University. The method is gradient-based, meaning the drones use small differences in signal strength (RSSI) to orient toward the peak. A gradient estimation algorithm was validated in MATLAB: the formation inferred the uphill direction from relative signal strengths and converged in 27 iterations. Two strategies to escape local maxima were tested: a jerk-based restart and a formation expansion. The expansion was more deterministic and efficient. For the physical swarm, a lead controller was designed from experimentally identified dynamics and maintained a stable formation. Indoor RSSI measurements were notably noisy, but a heatmap from 21 measurement points confirmed that the antenna peak was detectable from position-tagged RSSI data. Overall, the results show the approach is feasible: the algorithm, controller, and measurement platform each support progress toward a fully integrated, swarm-based RF measurement system.
Dette projekt undersøger, om en lille dronesværm kan finde det stærkeste punkt i signalet fra en retningsantenne hurtigere end traditionelle linje-for-linje scanninger med én drone. Tre RMTT-droner flyver i trekantformation, kommunikerer via ESP-NOW, og deres positioner spores med Vicon motion capture-systemet på Aalborg Universitet. Tilgangen er gradient-baseret, dvs. dronerne bruger små forskelle i signalstyrke (RSSI) til at pege mod det sted, hvor signalet er stærkest. En gradientestimeringsalgoritme blev valideret i MATLAB: formationen kunne udlede bevægelsesretningen ud fra relative signalstyrker og konvergerede på 27 iterationer. To strategier til at undslippe lokale maksima blev afprøvet: et jerk-baseret restart og en udvidelse af formationen. Udvidelsen viste sig mere deterministisk og effektiv. Til den fysiske sværm blev der designet en lead-controller baseret på eksperimentelt identificeret dynamik, som viste stabil formationsholdning. RSSI-målinger indendørs var markant støjprægede, men et heatmap fra 21 målepunkter bekræftede, at antennens top kunne identificeres fra positionsmærkede RSSI-data. Samlet viser projektet, at tilgangen er gennemførlig: algoritmen, controlleren og måleplatformen peger tilsammen mod et fuldt integreret, sværmbaseret RF-målesystem.
[This apstract has been rewritten with the help of AI based on the project's original abstract]