We are interested in tracking human postures by deploying accelerometers on a human body. One fundamental issue in such scenarios is how to calculate the gravity, no matter when human body parts are moving or not. Assuming multiple accelerometers being deployed on a rigid part of a human body, a recent work proposes a data fusion method to estimate the gravity on that rigid part. However, how to find the optimal deployment of sensors that minimizes the estimation error of the gravity is still an open problem. In this paper, we formulate the deployment optimization problem. Since the problem is hard, we conduct experiments to observe what the optimal deployments should be.

We consider a body-area sensor network (BSN) consisting of multiple small, wearable sensor nodes deployed on a human body to track body motions. Concerning that human bodies are relatively small and wireless packets are subject to more serious contention and collision, this paper addresses the data compression problem in a BSN. We observe that, when body parts move, although sensor nodes in vicinity may compete strongly with each other, the transmitted data usually exist some levels of redundancy and even strong temporal and spatial correlations. Unlike traditional data compression approaches for large-scale and multihop sensor networks, our scheme is specifically designed for BSNs, where nodes are likely fully connected and overhearing among sensor nodes is possible. In our scheme, an offline phase is conducted in advance to learn the temporal and spatial correlations of sensing data. Then, a partial ordering of sensor nodes is determined to represent their transmission priorities so as to facilitate data compression during the online phase. We present algorithms to determine such partial ordering and discuss the design of the underlying MAC protocol to support our compression model. An experimental case study in Pilates exercises for patient rehabilitation is reported. The results show that our schemes reduce more than 70 percent of overall transmitted data compared with previous approaches.