Aerodynamic parameters from distributed heterogeneous CNT hair sensors with a feedforward neural network

Kaman Thapa Magar, Gregory W. Reich, Corey Kondash, Keith Slinker, Alexander M. Pankonien, Jeffery W. Baur, Brian Smyers

Research output: Contribution to journalArticlepeer-review


Distributed arrays of artificial hair sensors have bio-like sensing capabilities to obtain spatial and temporal surface flow information which is an important aspect of an effective fly-by-feel system. The spatiotemporal surface flow measurement enables further exploration of additional flow features such as flow stagnation, separation, and reattachment points. Due to their inherent robustness and fault tolerant capability, distributed arrays of hair sensors are well equipped to assess the aerodynamic and flow states in adverse conditions. In this paper, a local flow measurement from an array of artificial hair sensors in a wind tunnel experiment is used with a feedforward artificial neural network to predict aerodynamic parameters such as lift coefficient, moment coefficient, free-stream velocity, and angle of attack on an airfoil. We find the prediction error within 6% and 10% for lift and moment coefficients. The error for free-stream velocity and angle of attack were within 0.12 mph and 0.37 degrees. Knowledge of these parameters are key to finding the real time forces and moments which paves the way for effective control design to increase flight agility, stability, and maneuverability.

Original languageEnglish (US)
Article number066006
JournalBioinspiration and Biomimetics
Issue number6
StatePublished - Nov 10 2016
Externally publishedYes


  • aerodynamic parameter prediction
  • artificial hair sensors
  • fly by feel
  • neural network

ASJC Scopus subject areas

  • Biotechnology
  • Biophysics
  • Biochemistry
  • Molecular Medicine
  • Engineering (miscellaneous)


Dive into the research topics of 'Aerodynamic parameters from distributed heterogeneous CNT hair sensors with a feedforward neural network'. Together they form a unique fingerprint.

Cite this