Research On Energy Efficiency Of VMIMO-based Routing For Wireless Sensor Networks

With the development of information technology and mobile Internet, the rela-tions of people to people, people to objects, objects to objects become closer gradually. Wireless sensor networks (WSNs), as a key technology of Internet of things (IOTs), have attracted wide attention in both academia and industry. WSNs are a new form of network, and built a bridge of information exchanges for people and objects. Now there are many applications with WSNs, such as in the health care, environmental mon-itoring, military security and other fields. However, before the widespread deployment of sensor networks, there are still some key technical issues need to be solved. Par-ticularly, the energy efficiency problem limits the deployment scale, and restricts the network lifetime seriously.The sensor nodes are usually small with a simple structure, and have limited en-ergy. As the charging and battery replacement are often subject to various environ-mental factors, the main way to prolong the network lifetime is to reduce the energy consumption of nodes, and improve the energy efficiency. Among the many energy-saving schemes, the energy-efficiency routing based on virtual multiple input multiple output (vMIMO) is a hot topic. This thesis analyses relevant research results, and find that they always builds up the vMIMO-based routing on the fixed structure such as clusters, and the MIMIO mode is omitted in most cased. As a result, they cannot fully explore the advantage of spatial diversity. In this thesis, we study a general and simple vMIMO model in which no fixed structure is required, and any communication mode of vMIMO is allowed for sake of the energy efficiency. Based on it, this thesis defines two problems of energy efficiency, and studies corresponding routing algorithm design:(1) The energy-minimum vMIMO-based routing problem. It concerns the overall energy level of each node in the route. The routing solutions need to find a path from the source node to the destination node, which has minimal energy consumption. This thesis proposes a distributed energy-minimum vMIMO-based algorithm (DEMVA) to solve this problem. The algorithm constructs the virtual cooperative graph (VCG) to de-scribe the vMIMO communications between nodes, so that it can compute the objective route based on Bellman-Ford. Then the thesis gives the detailed analysis of time com-plexity and message complexity according to the node and link numbers in VCG. The simulation results illustrate that DEMVA has best performance than other algorithms, which can fully exploit the cooperation among nodes to decrease the energy consump-tion. For example, it can save energy about30%compared with the non-cooperative routing on average. (2) The lifetime-optimal vMIMO-based routing problem. When routing solutions only focus on the overall energy consumption, and ignore the residual energy level of nodes, as time goes on, the network will encounter routing interrupt, topology changes and so on. That will affect the actual energy consumption. This thesis tries to ex-plore the energy efficiency from routing lifetime, and proposes a distributed lifetime-oriented vMIMO-based algorithm (DLOVA). Firstly it computes the lifetime-optimal routing without vMIMO communications, which is modified from shortest path algo-rithm. Second, the cooperative relations between nodes will be exploited, so that the routing lifetime will be extended gradually after multiple iterations. From the theoret-ical analysis, this algorithm can achieve the approximate performance ratio of4. The simulations show the DLOVA can work well in many situations. For example, DLOVA can prolong the lifetime about20.2%in dense topologies compared with the cooperative routing algorithm on average.