Based On The Method Of The Microwave Signal Recognition Of Human Motion Study

In this project, we propose a novel method for recognition of human bodymovement using signal propagation path loss. The medium used is microwave signals ata frequency of 2.4GHz and the data gathering system is based on an existing system thatwas built by Prof. Peter Hall and Dr. Yuri Nechayev in the University of Birmingham.The transmitter and receiver are both placed on human body. The propagation path lossis chosen as the input data, and two group of data was obtained while the person carriesout 18 different static postures and 'moving' postures, both in the chamber and the lab.A signal processing system has been built to recognize the original data obtained byProf. Peter Hall and Dr. Yuri Nechayev, without considering the body propagationcharacteristics. This system includes pre-processing, feature extraction, classificationand dimensionality reduction, and gives a correctness of 97.8% in chamber and 93.3%in lab. However, the experiment result and the theoretical theory of the bodypropagation characteristics show that it is very difficult to recognize the body staticpostures by using the original data gathering system. Therefore we have proposed a newmethod to gather the movement data of the human body, in which we acquired the dataof 6 head related movements that are repeated periodically, including head turning toleft/right, head forward/backward and body turns to left/right. Data selection, averaging,de-trend, and curve fitting are used to pre-process the data. Experimental work showsthat the features such as variance, density spread of the movement can be used todiscriminate the head related movements. Phase shift can be used to separate theopposite movement, for example the head turning to the left and head turning to theright. The results show the feasibility of this method. However, a lot more researchwork needs to be done in future. Particularly, more experiments are required for featureextraction. Meanwhile a good combination of the on-body propagation characteristics isnecessary.