Zhu Zhenmin

Heart rate measurement via smart phone acceleration sensor

The invention of various types of sensors, enhancement of computing ability of the smart terminal and booming development of wearable and portable device gave an practical way to monitor the elders physical condition, especially people with heart disease or blood pressure disease. This paper describes a method to get the apical pulse signal via acceleration sensor and proposes a new way estimating the heart rate, besides, a signal processing procedures are given to get and tracking the signal. Considering in detail, we firstly analyzed the generation mechanism, and found suitable positions to deploy sensors, then we came up with an idea that separating heart rate signal and respiratory via wavelet analysis, further a real-time heart tracking algorithm is designed and local prediction via phase-space method is used to predict missing heart rate. After that, we discussed about the embedding dimension and the delay time of different ages of people. Finally, we carried out an experiment with commercial device, and proposed a evaluation method based signal quality, then analyzed the tracking performance. This low cost technology can be used in real time monitor on wearable devices especially emergency condition and long-term health support.

UDIO: An Ubiquitous Device Inter-operation System Based on IGRS

To address the problem of service composition in pervasive computing, and to provide the IGRS ubiquitous devices the abilities of inter-operation and cooperation, we design the UDIO, a middleware of dynamic service composition based on IGRS protocol. In this paper, the system architecture of UIDO, as well as a matrix-based path-finding algorithm for service composition, are introduced which prove to be available and efficient in our IGRS smart home.

Research of mobile applications-oriented peer-to-peer context-aware mechanism

Mobility and distribution are the key characteristics of pervasive computing environment. It is an inevitable problem for mobile applications how to find available context in a distributed context-aware system. This paper proposes a context-aware mechanism to enable mobile applications to be bound up with different agents adapting to the changing location of mobile applications. The mechanism relies on a distributed peer-to-peer architecture. Simulation experiments demonstrate that the context-aware mechanism gets advantage over CTK(Context Toolkit) in load balance and service response time.

A new DVS prediction algorithm for ubiquitous devices based on Cycle Mode Matching

To take advantage of the full potential of ubiquitous devices, we will need systems which minimize power consumption. In this paper, a new time-interval-based Dynamic Voltage Scaling (DVS) prediction algorithm, Cycle Mode Matching (CMM) is proposed through reasonable analysis of the prediction parts of main current DVS algorithms. The simulation results show that the accuracy of DVS prediction is vastly improved especially for ubiquitous devices without the perceptible impact of system performance.