Real-time implementation of quaternion-based human footstep detection

• Main Author: Shen, Huizhi.
• Other Authors: Andy Khong Wai Hoong
• Format: Final Year Project
• Language: English
• Published: 2013
• Subjects: DRNTU::Engineering::Electrical and electronic engineering::Electronic systems::Signal processing
• Online Access: http://hdl.handle.net/10356/54489
• Institution: Nanyang Technological University

Footstep detection, a main component in human surveillance system, requires effective algorithm to be developed to detect the presence of footstep for a desired detection range. In this project, a quaternion theory is applied to represent trajectory of particle velocity generated by footstep seismic waves and captured by tri-axial geophone. Directional statistics is used on the rotation axis, rotation angle and orientation quaternion to detect the presence of elliptical polarization and hence footsteps. The quaternion-based algorithm is implemented in real-time. Further, the rotational characteristics using quaternion is combined with kurtosis, an energy-based technique a more robust detection. Vehicle bearing estimation is another important task in most surveillance systems. Rotation axis given by quaternion provides significant information of source direction which can be considered as an indicator of location of source.