Head Pose Estimation and Tracking for Moving Target Active Noise Control Using Convolutional Neural Network

  • Author(s) / Creator(s)
  • This paper introduces a novel approach towards an implementation of head pose estimation techniques for the application of noise control. An existing multi-stage process, which uses convolutional neural network (CNNs), for head pose estimation is modified, implemented, and validated for moving target active noise control. With evaluation and development of application focused methodologies, this paper highlights the modifications necessary alongside the challenges involving real- time tracking. Utilizing such a method to improve noise comfort within aircraft cabins through an integrated real-time tracking system allows for the capability to create a zone of quiet bubble around a passenger’s head during their flight time. Through various stand-alone and integrated system tests, the head tracking algorithm proposed shows desirable accuracies, exhibited through low total average % errors. Additionally, the integrated system also shows positive results with low tracking errors alongside effective real-time capabilities. The overall implementation provides a convincing solution for head- tracking system to be integrated with active noise control for moving targets within aircraft cabins.

    Part of Proceedings of the Canadian Society for Mechanical Engineering International Congress 2022.

  • Date created
  • Subjects / Keywords
  • Type of Item
    Article (Published)
  • DOI
  • License
    Attribution-NonCommercial 4.0 International