Reverse recognition of body postures using on-body radio channel characteristics

The problem of posture detection is of considerable significance for assisted living (AL). In most cases, radio channel models for wireless body area network (WBANs) are fixed when a specific body posture is considered. To the best of the authors' knowledge, little work has been done on the reverse body posture information extraction using WBAN radio channel characteristics. This study aims to classify human postures from on-body narrowband wireless channel information. It is demonstrated that by applying the random forest (RF) classification technique, the action of the human body can be detected. The classification error is perfectly acceptable for RF algorithm. Two propagation environments were compared and the results indicate that the classification error is less in the anechoic chamber (21.39%). In summary, this study provides a novel approach to detect human body postures by using body-centric wireless channel information, and will be beneficial for AL.