Previous [ 1] [ 2] [ 3] [ 4] [ 5] [ 6] [ 7] [ 8] [ 9] [ 10] [ 11] [ 12] [ 13] [ 14] [ 15] [ 16]


Journal of Information Science and Engineering, Vol. 32 No. 4, pp. 947-968 (July 2016)


Collaborative Sensing for Heterogeneous Sensor Networks*


KEJIANG XIAO1, RUI WANG2,+, CAI CAI3, SHAOHUA ZENG1 AND ZAHID MAHMOOD4
1Information and Communication Company of State Grid Hunan Electric Power Company
Changsha, 410007 China
2Institute of Computing Technology
Chinese Academy of Sciences
Beijing, 100190 China
3Department of Basic Curriculums
Changsha Environmental Protection College
Changsha, 410004 China
4School of Computer and Communication Engineering
University of Sciences and Technology Beijing
Beijing, 100083 China
E-mail: 985089629@qq.com; wangrui@ustb.edu.cn

Collaboration between low-quality sensor and high-quality sensor can achieve tradeoff between accuracy and energy efficiency in heterogeneous sensor networks (HSNs). Generally, HSNs are deeply integrated with dynamic physical environments. The monitored target's dynamics are the most important and common factors of the dynamic environments. Some important parameters (such as active opportunity, sampling frequency and sampling time) fail to adapt to the changes, which undermines the collaboration's performance when the state of the monitored target changes. Even the system performance is not up to user requirements or large amounts of energy are consumed. To solve this problem, we propose an adaptive collaboration scheme named EasiAC by the collaboration between magnetic and camera sensors. First, for the dynamics of the monitored target, EasiAC utilizes the magnetic sensors to predict the targetíŽs state via Bayesian filtering based method. Second, to achieve good performance of such collaboration, EasiAC adjusts the camera sensorsíŽ active opportunity, optimal sampling frequency and sampling time dynamically according to the estimated results from the magnetic sensors. Finally, a boosting based algorithm named BbTC is proposed to make classification for the target to achieve high accuracy (average classification accuracy is more than 98%). We evaluate EasiAC method through extensive simulations and real road environment experiments. The results demonstrate that EasiAC needs less energy consumption (saving 97% energy) than traditional solutions, while maintaining the performance at acceptable level (average image integration ratio is 90%) in the presence of target's dynamics.

Keywords: Bayesian filtering, dynamic environments, collaboration, sampling frequency, target classification, wireless sensor networks

Full Text (ą■Ąň└╔) Retrieve PDF document (201607_08.pdf)

Received March 21, 2015; revised November 3, 2015; accepted December 21, 2015.
Communicated byJiann-Liang Chen.
* This paper is supported by National Natural Science Foundation of China (NSFC) under Grant No. 61379134.
+ Corresponding author: Rui Wang (wangrui@ustb.edu.cn)