基于无线传感器网络的奶牛运动行为实时监测系统

A Real-time Monitoring System for Cow Behavior based on Wireless Sensor Network

针对我国中小型奶牛养殖场所面临的管理方法落后、疾病监测自动化程度低、实时判定奶牛突发行为困难等问题,开发了一种基于无线传感器网络的奶牛运动行为实时监测系统。系统设计采用3层架构模式,实现奶牛基础信息与运动信息的采集、存储、管理、展示功能,其中奶牛行为判别采用一种改进的模糊聚类算法。试验表明,系统稳定性较强,可对数据进行有效的组织管理;判别算法能够可靠的将试验数据分为进食、平躺、站立、平躺动作、站立动作、快走、慢走7类运动行为,平均准确度达到83.82%,相较于K-means聚类算法,判别的精密度与精细度方面均有显著提升,该系统在奶牛运动行为实时监测领域具有广阔的应用前景。

In China's small and medium-sized dairy farms,the currently management system is still remained in a comparatively backward state.The degree of automation in disease detection is low.It is also hard to predict cows'emergent behavior in real time.In order to solve these problems,a real-time monitoring system for cows'motor behavior was developed based on wireless sensor network.This system includes three architecture patterns.It can collect,store,manage and display information for cows'basic and movement messages.The fuzzy clustering algorithm was used to distinguish cows'behavior.The test shows that this system is stable and can organize the data effectively.Discriminate algorithm can divide the experimental data into seven categories for cows'movement behavior,such as feeding,lying,standing,lying down,standing up,active walking and normal walking.The average accuracy has achieved 83.82%.Compared with K-means clustering algorithm,the precision of the discriminant and finesse was increased significantly.This system has broad application prospects in the field of cow behavior real-time monitoring.