Data-Driven Sparse Sensor Placement Optimization on Wings for Flight-By-Feel: Bioinspired Approach and Application.

Area of Science:

• Aerospace Engineering
• Biomimetics
• Sensor Technology

Background:

• Flight-by-feel (FBF) uses wing sensors for flight control, inspired by biological systems.
• Artificial FBF systems face Size, Weight, and Power (SWaP) constraints, especially for small aircraft.
• Optimizing sensor placement is crucial due to nonlinear airflow and local minima in conventional methods.

Purpose of the Study:

• To develop and evaluate an optimization approach for sparse sensor placement on aircraft wings.
• To determine the optimal number and location of sensors for effective flight state detection.
• To address the limitations of conventional optimization techniques in complex airflow fields.

Main Methods:

• The Sparse Sensor Placement Optimization for Prediction (SSPOP) algorithm was employed.
• SSPOP utilizes singular value decomposition and linear discriminant analysis to identify information-rich sensor locations.
• The algorithm was evaluated on a 3D delta wing model with variable length artificial hair sensors.

Main Results:

• The SSPOP-derived sensor placement ranked in the top one percent for angle of attack prediction accuracy.
• SSPOP proved more robust and nearest to optimal than greedy search and gradient-based methods in over 90% of models.
• This study is the first to apply SSPOP to a 3D model and incorporate variable length sensors for diverse velocity sensitivity.

Conclusions:

• SSPOP is a highly effective algorithm for optimizing sparse sensor placement in complex 3D airflow.
• The findings pave the way for experimental validation of artificial hair-cell airflow sensor placement.
• This research represents a significant advancement towards achieving biomimetic flight-by-feel control systems.