Research On Wireless Sensor Network Multi-channel Access Mechanism Based On Structural Condition Monitoring Under The Conditions Of Disaster

Wireless sensor network is of high robustness, accuracy, flexibility and othercharacteristics, which makes it has an irreplaceable advantage when applied to structuralfacilities monitoring. Wireless sensor network is used in structural facilities monitoring(such as offshore platform), when disaster occurs, massive data collected by wirelesssensor nodes need real-time transmission. However, the traditional wireless sensornetwork is based on Zigbee, which primarily considers reducing energy consumption.The characteristics of low transmission rate and poor real-time cannot meet thereal-time transmission requirements of massive data. So this paper mainly researchesstructural condition monitoring for real-time transmission problems of massive dataunder the conditions of disaster.For structural facilities monitoring, this paper uses cluster-based wireless sensornetwork, which sets a cluster as a unit, managing the data transmission through twostages of inner-cluster and inter-cluster transmissions. This paper aims at theinner-cluster and inter-cluster massive data real-time transmission of wireless sensornetwork transmission system for structural condition monitoring, the main researchwork carried out as follows:(1) Wireless sensor network is applied to large-scale structure monitoring, and thispaper proposes wireless sensor network transmission system for structural conditionmonitoring under the conditions of disaster. The system has five parts, which includesthe clustering, channel negotiation, inner-cluster data transmission, cluster head datafusion, and inter-cluster data transmission. This paper analyzes the problems of eachpart of the system for realizing massive data real-time transmission, and researches theinner-cluster and inter-cluster transmission problems.(2) For inner-cluster massive data transmission, this paper proposes a matchingalgorithm based on sub-region, which is based on the characteristics of wireless sensornodes, and combined the virtual MIMO technology with its matching strategy. In thisresearch environment, cluster head node configurates multiple antennas, inter-clusterordinary node configurates an antenna. To form virtual MIMO structure, the transmittersingle antenna arrays jointly constitute a receiver multi-antenna array. At this point, thecluster head node communicates with two nodes at the same time, the throughput isimproved. Aim at the conditions of disaster, this paper proposes matching algorithm based on subregion, its main idea is that reasonably divide the inner-cluster region intomultiple subregions, then select the appropriate node from each subregion to pair andtransmit. It effectively avoids that getting information is too concentrated in one orseveral subregions lead to sweeping phenomenon. The simulation results show that, in acertain range, with increase of the subregions number, the data information got bycluster head node appears homogeneous distribution trend between different subregions.(3) For inter-cluster massive data transmission, this paper considers multi-channelaccess mechanism and proposes an auto rate dynamic channel allocation protocol basedon DCA protocol, namely AR-DCA protocol (Auto Rate Dynamic Channel AssignmentProtocol). This paper solves the problems existing in the traditional DCA protocol, andimproves the problems of packet loss rate is high when the channel quality become poor,throughput is low when the channel quality become good in traditional DCA protocol.Simulation results show that, while the channel quality is good，the average networkthroughput of the AR-DCA protocol is significantly improved comparing with thetraditional DCA protocol; while the quality is poor, frame loss rate of the AR-DCAprotocol is reduced by about38.5%.