摘要
针对传感器网络应用对节点同步精度要求高、通信开销小等需求,研究一种基于最优线性拟合的时间同步算法。在分析FTSP算法的基础上,通过节点分级策略,减少同步分组数据传输量,并引入概率统计学中参数估计理论改进线性回归算法,减小异常数据点对同步精度的影响,延长节点同步时间。实验结果表明:该算法能有效减少同步通信开销,实现高精度的时间同步,同步误差在μs级。
According to the requirements of high synchronization precision among nodes and low communication costs,a new synchronization algorithm based on optimal linear fit has been studied.After analysis of FTSP algorithm,node classifying strategy was adopted to reduce the transition amounts of synchronization packet data,and parameter estimating theory of probability and statistics was then introduced to improve linear regression algorithm,reduce the influence of abnormal data to synchronization precision and extend node synchronization time.Experiments showed that this new algorithm can effectively reduce the synchronization costs and realize high synchronization precision with μs level errors.
出处
《传感技术学报》
CAS
CSCD
北大核心
2010年第12期1787-1791,共5页
Chinese Journal of Sensors and Actuators
基金
国家科技支撑计划项目资助(2009BAI87B05)
创伤烧伤与复合伤国家重点实验室自主研究课题资助(SKLZZ200903)
