Vector Intelligent Sensor for Human Motion Detection and Navigation

D. Qiu, J. Litton, C. Yang

Abstract:	In this paper, we investigate two human motion detection mechanisms: feature-based and waveform-based. The feature-based motion detection mechanism uses pattern classification algorithms to detect motions using the trained motion models. Same motion models as well as feature thresholds will be applied to match the new received motion signature to the ones stored in a database. The objective of the waveform-based motion detection mechanism is to complement the feature-based approach, thus determining various motion types using calibrated waveforms and their statistics for transition motion states. From an information theory point of view, the waveform contains more information content to describe motion types while the extracted features represent the useful and clean information. This paper learns the tradeoff of the two information types, features and waveforms. Measurements from MEMS IMUs will be used to test and validate the feature-based and waveform-based human motion detection mechanisms.
Published in:	Proceedings of the 2011 International Technical Meeting of The Institute of Navigation January 24 - 26, 2011 Catamaran Resort Hotel San Diego, CA
Pages:	670 - 677
Cite this article:	Qiu, D., Litton, J., Yang, C., "Vector Intelligent Sensor for Human Motion Detection and Navigation," Proceedings of the 2011 International Technical Meeting of The Institute of Navigation, San Diego, CA, January 2011, pp. 670-677.
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In