Research On The Key Issues Of WSN-MEs On Data Collection

Wireless sensor networks (WSNs) have emerged as an efective solution for a widerange of applications. Most of the traditional WSN architectures consist of static nodeswhich are densely deployed over a sensing area. In recent years it has become a researchfocus in WSNs by introducing mobile elements (MEs) into traditional WSNs. The mobility-enabled WSNs improve energy efciency and reduce networking cost significantly. How-ever, the mobility management also introduces significant challenges which do not arise instatic WSNs. There are diferent schemes for data collection application for WSN-MEs. Forreal-time application scenarios, how to capture the trajectory of mobile elements and builddynamic routing is a challenge. If direct contact scheme is used, the large data collectionlatency can be incurred by low movement speed of mobile nodes. It is observed that the la-tency can be efectively shortened by planning the trajectory of mobile nodes and performinglocal data aggregation. This kind of approach usually considers routing and mobility controlin data collection. Due to the low cost and the large-scale deployment in uncontrolled oreven harsh environments, it is common for sensor nodes to become faulty. The existence ofthese faulty nodes in WSNs causes the data collection many adverse efects such as inaccu-racy of the gathered information, extra energy consumption, unreliable routing. Therefore,how to identify faulty nodes and eliminate their impact and build resilient routing on thedata collection gradually attracts more and more attentions.On the basis of in-depth analysis of existing research, we study several key issues sur-rounding data collection in WSN-MEs. The main works and corresponding contributions ofthis thesis include the following three aspects.1. We present a grid-based real-time data gathering protocol in WSNs-MEs to addressthe dynamic routing problem.(1) We propose a flexible and even grid partition schemebased on transmission power, with which well scalability and reduced routing hops can bereceived.(2) The grid head election mechanism considering residual energy of networknodes is utilized to balance energy consumption.(3) We present a construction approachof initial optimal routing tree. Furthermore, in order to decrease the latency, a local routingupdating idea is applied in the grid structure. In this approach, protocol only disseminatelatest location information about mobile nodes among grid heads.(4) Simulation results show that our approach can achieve satisfactory efectiveness under diferent settings.2. In order to reduce data collection latency due to the low movement speed of mobiledata collector (MDC), we propose efcient rendezvous based data collection algorithms withrelay hop constraint.(1) We define the mobile data collection problem based on rendezvousnodes as MDC-RN, which jointly considers MDC tour and routes in aggregation trees, andprove that it is NP-Hard.(2) We develop two efcient rendezvous algorithms to solve theMDC-RN problem. The former is a heuristic algorithm which always prefer to select thenode with maximum load from the d-hop neighbors as the RN candidate. The latter catersto the characteristic of WSNs, and determines RN iteratively in distributed manner. On thebasis of selected RNs, using algorithm for traveling salesman problem to produce mobilenode tour, along which ME periodically visits these RNs and picks up the cached data.(3)The efectiveness of our schemes is validated through both theoretical analysis and extensivesimulations.3. To address the reliability due to faulty nodes, we design a diagnosis-based RN elec-tion and multi-path routing protocol for WSNs. We analyse the opportunities and challengesof fault diagnosis based on comparison model in WSNs, and so far there has been little workdone for this owing to the inherent characteristic of WSNs. Furthermore, we design and im-plement a comparison-based fault identification and multi-path routing (CFIMR) for WSNs.The features making CFIMR distinct are as follows:(1) RN election based on fault diagnosiseliminates the possibility of RN acting by faulty nodes which reduce energy consumptionand fault information transmission.(2) The constructed directed acyclic graph provides amulti-path routing approach, which increases system fault tolerance.(3) RN election andmulti-hop routing consider residual energy and transmission cost, respectively, which bal-ances energy consumption and promotes energy efciency significantly. Consequently, thecommunication overhead and network lifetime of CFIMR are desirable.