Research On The Architecture And The Related Algorithms Of Sensor Networks Based On MR-MC

Wireless sensor network has been paid high attention in the academic and military areas as its low-power, low-cost, distribution and self-organization characteristics. With the development of wireless communication, computer technology and large-scale IC technology, the cost of sensor node has been reduced greatly, and this not only lessens the bad impact of price on the development of sensor network, but also provides a new method for researcher. Recently, with the arising of Internet of Things, sensor network has been widely used in the area of digital manufacturing and industry control, and meanwhile this will put forward higher requirement on its performance.Sensor network is high related with its application background, the existing research results can not be used for the manufacturing sensor network; single radio and single channel sensor network can not solve the problem of hidden terminal and exposed terminal radically, and with the increasing of sensor nodes, the competition on wireless channel will be aggravated; in addition, manufacturing network has high demand on network real-time performance and reliability, and this is a big challenge for sensor network. So research on the MR-MC (Multi-Radio Multi-Channel) sensor network architecture and the related algorithms is of great academic and application value.This thesis researches deeply on the key technologies of sensor network which has high demand on network real-time performance and reliability from three aspects: network architecture, the related algorithms of datalink and network layer. The main work is shown as follows:Firstly, research on the MR-MC cellular and hierarchical network architecture. As the problem of intense competition on wireless channel and poor system processing ability in SR-SC (Single-Radio Single-Channel) sensor network, a cellular and hierarchical network architecture basd on MR-MC has been proposed. By introducing MR-MC technology and subnetting, the competition on wireless channel and communication interference between sensor nodes have been reduced from physical and logistic layers, and the network capacity and the real-time performance of sensor network, as well as the utilization ratio of wireless channel have been improved.Secondly, research on the wireless channel access and control method. Because CSMA/CA protocol has uncertainty in the aspect of wireless channel access delay, a virtual token wireless MAC protocol based on consultation mechanism has been proposed. By making use of the continuity of sensor node address in subnet, virtual token is used to access and control wireless channel, and the number of network management messages brought by token maintenance and the energy consumption of sensor node have been reduced effectively, the effective utilization ratio of wireless channel and the real-time performance of sensor network have been improved.Thirdly, research on the multi-channel and network bandwidth allocation algorithm. As the characteristic of static allocation of channel resource in sensor network, a multi-channel allocation algorithm based on minimum interference cost has been proposed. Channel reuse has improved the network capacity, the choose of interference radius has reduced the interference between channel, the use of alarm mechanism in channel pool has decreased the risk of failure; as the fixed bandwidth allocation method in token ring network, a network bandwidth allocation algorithm based on priority and load-aware has been proposed. The order of virtual token passing in subnet and the number of messages sensor node can send when holding token will be adjusted dynamically according to the traffic load of single node and whole network and the priority of messages, and make the bandwidth allocation more reasonable and meet the bandwidth requirement of high and heavy node.Fourthly, research on the sensor node location algorithm. Because the existing sensor node location algorithms are mainly focus on 2D space and they have poor accuracy, a parameter matching 3D accurate location algorithm based on RSSI has been proposed. The parameter matching of signal attenuation model between every pair of unknown node and beacon node has reduced the effect on location accuracy as individual difference and environment factor; by making use of theoretical and empirical model to compensate the attenuation of wireless signal in obstacle, the location accuracy has been improved further; the change of role between unknown node and beacon node can reduce the effect brought by the unbalanced distribution of beacon nodes; change the transmitting power of sensor node and the time interval of positioning can effectively reduce the energy consumption of sensor node.Fifth, research on the sensor network routing protocol. As the limitation of resources in sensor node and the advantages of geographical routing algorithm, by making use of the coordinate information got from 3D location algorithm, a 3D geographical routing algorithm based on coordinate information has been proposed, which combines the ideal minimum energy model and the hop number model, and balanced the relationship between single hop distance and energy consumption; choose the sensor node in their cross area which is nearest to the space straight line between the source node and the destination node for routing, and this has shorten the distance of data transmission; when encountering "empty" during the routing, randomly choose the minus neighbouring node as the next hop node, and this can reduce the possibility of encountering "empty" again, while always transmitting along forward direction has avoided the happening of circle phenomenon.