Safe Bird Migration : Evaluating a Low-Cost FMCW Doppler Radar for Offshore Wind Farm Collision Avoidance

Sammanfattning

Background: Offshore wind farms play an important role in the transition to renewable energy but can negatively impact bird populations through collisions with turbine blades. Existing vision-based bird protection systems are limited by weather conditions, visibility, and nighttime operation, motivating the investigation of alternative sensing technologies. Objectives: This thesis aims to investigate the feasibility of a low-cost Frequency-Modulated Continuous-Wave (FMCW) Radio Detection and Ranging (RADAR) system for avifauna detection at offshore wind farms, with the objective of enabling reliable short-range bird detection using commercially available off-the-shelf components. The study addresses how a low-cost FMCW radar system can be designed and utilized to achieve reliable small-target (bird) detection in offshore wind farm environments through appropriate parameter selection and signal processing methods. Methods: A prototype FMCW radar system was designed and implemented using commercially available components. Theoretical system parameters were first derived from radar models and subsequently refined through experimental optimization. The radar was evaluated through directivity, range, velocity, and target detection tests using calibration targets, drones, and birds in outdoor environments. Signal processing techniques including range-Doppler estimation and Constant False Alarm Rate (CFAR) filtering were applied to analyze radar performance. Results: Theoretical parameter estimates were found to be insufficient due to practical hardware limitations, requiring experimental optimization that established a minimum chirp duration of 500 μs for reliable operation. The radar achieved an antenna gain of 16.18 dBi, an azimuth beamwidth of 4.82°, and an elevation beamwidth of 3.90°. Dynamic targets with an estimated RCS of 0.05 m2 were detected at distances up to 165 m, exceeding the target requirement of 100 m. Bird detections were achieved at ranges up to 70 m, and micro-Doppler signatures generated by drone propellers and bird wingbeats were successfully observed. The total component cost of the prototype was approximately 3760 euros. Conclusions: The results demonstrate that a low-cost FMCW radar system based on off-the-shelf components can provide reliable short-range avifauna detection and tracking. The system exceeded the target detection range while also demonstrating the ability to capture micro-Doppler signatures that may support future target classification. Although improvements in hardware integration, signal processing, and long-term field validation are required, the proposed system shows strong potential as a complementary sensing solution for bird protection at offshore wind farms.