Research on a Sound Source Localization Method for UAV Detection Based on Improved Empirical Mode Decomposition

Area of Science:

• Acoustics and Signal Processing
• Aerospace Engineering
• Robotics and Control Systems

Background:

• Traditional radar and visual tracking methods for unmanned aerial vehicles (UAVs) face limitations in certain operational environments.
• Accurate localization is crucial for UAV traffic management, security, and operational safety.
• Developing alternative localization techniques is essential for expanding UAV operational capabilities.

Purpose of the Study:

• To propose and validate an innovative acoustic source localization method for UAVs.
• To enhance the accuracy and reliability of UAV localization, particularly when radar or visual methods are infeasible.
• To provide an efficient and real-time solution for detecting and locating small UAVs.

Main Methods:

• Applied smoothing filtering and Robust Empirical Mode Decomposition (REMD) to UAV flight signals.
• Utilized an adaptive frequency window, optimized by Grey Wolf Optimizer (GWO), to extract relevant Intrinsic Mode Function (IMF) components.
• Employed the Chan-Taylor localization algorithm with weighted least squares, using calculated sensor time differences for target positioning.

Main Results:

• The proposed acoustic localization method demonstrated robustness and high performance in simulations and real-world tests.
• Localization errors were consistently below 5% within a 15 m × 15 m measurement area.
• The method proved effective for real-time detection and location determination of small UAVs.

Conclusions:

• The developed acoustic localization technique offers an efficient and accurate solution for UAV detection.
• The integration of REMD, adaptive frequency windows, and advanced localization algorithms enhances localization precision.
• This method provides a viable alternative for UAV localization in challenging environments where conventional methods fail.