Confidence Weighted Time Synchronization Of Wireless Sensor Networks For Steel Structure Damage Detection

To overcome the problems that traditional wired structure damage detection with blind zone is cumbersome and costly.In this paper,the multi-robot is used as the carrier to construct the steel structure damage detection system based on mobile wireless sensor network,which can realize detection without blind zone.However,damage detection based on wireless sensor network requires multi-robot nodes to collect vibration data synchronously,while multi-robot self-organizing networks belong to distributed system.Due to factors such as different initial clocks of each node,the error of crystal oscillator process and the interference of steel structure environment,there is a problem that the time is not synchronized between the robot nodes.In addition,steel structure damage detection requires long-term monitoring,and the power supply carried by the robot node is limited.Therefore,under the premise of ensuring time synchronization accuracy,energy-saving optimization of the wireless sensor network must be realized.To overcome the problems that time is not synchronized and high energy cost of wireless sensor networks for steel structure damage detection,based on the flooding time synchronization algorithm,the study on time synchronization error and energy saving optimization of sensor nodes is carried out.The main work and results of this study are as follows:Firstly,aiming at the problems that the interference of the steel structure environment leads to abnormal synchronization data in the wireless sensor network,based on the flooding time synchronization algorithm,the synchronization deviation in the linear regression table are estimated,and the abnormal time synchronization data is eliminated by using the confidence interval.However,the data in the confidence interval can not eliminate the abnormal data caused by environmental interference completely.Therefore,the random weighted average algorithm is introduced.After establishing a clock drift rate data table,the weighting coefficient is dynamically solved to optimize the stability of the clock drift rate.Then fit the effective data and improve the stability and fault tolerance of the time synchronization algorithm.In the ideal and abnormal situation,the FTSP algorithm and the improved FTSP algorithm are simulated,and the improved algorithm has better fault tolerance and synchronization precision than the FTSP algorithm.Secondly,the energy-saving optimization of the wireless sensor network is carried out.At the network level,aiming at the problem that the entire nodes in the network broadcast with large energy consumption,the parent node filtering mechanism is introduced to divide the nodes into active nodes and passive nodes,thereby eliminating redundant broadcast nodes,achieving energy-saving optimization and improving network performance.In addition,the optimized time synchronization algorithm is more robust to node failures.At the node level,energy dynamic management is introduced.According to the working process of the robot node,the invalid power consumption is reduced by dynamically switching the node state.Then,the Telosb node and the TinyOS operating system are used as experiment platforms to perform microsecond time synchronization testing of wireless sensor networks.The experimental results show that the synchronization accuracy based on the improved FTSP fully meets the requirements of steel structure damage detection,and it has better fault tolerance in the case of abnormal interference.Finally,In order to develop wireless sensor nodes with better performance,millisecond time synchronization testing was conducted with CC2530 node and ZStack protocol stack.It lays a foundation for further development of the wireless sensor network steel structure damage detection system based on multi-detection robots.