Research On Load Accumulation And Collisions In Data Collection For WSNs

Along with the development of IoT, the supporting technology that Wire-less Sensor Networks(WSNs) has obtained a wide attention and become one of the central issues. In recent years, new technologies, like for instance Large-scale Wireless Sensor Networks towarding environment monitoring, small-scale Wireless Body Area Networks towarding medical and health monitoring, are receiving more and more attentions. The significant challenge of large-scale Wireless Sensor Networks is the load accumulation in data gathering process,which leads to serious energy consumption and unbalance. While the data colli-sion and channel conflict induced by intensive of nodes and mobility of human body, are the most worthy of consideration in Wireless Body Area Networks,which bring coexistence problem. According to the spacial correlation and K-sparse feature of the monitoring data, we study the data gathering scheme based on Hybrid Compressive Sensing(Hybrid-CS). While, according to the high den-sity and mobility, we study the data collision of CSMA/CA and the channel conflict of channel hopping meachanism respectively. The main contributions and innovations of this thesis are as follows:(1) An enhanced Hybrid-CS data gathering scheme in large-scale Wireless Sensor Networks.Due to the similarities among the environment monitoring data in large-scale Wireless Sensor Networks, too much redundant data waste unnecessary resources. Hybrid-CS integrates the technical advantages of compressive sens-ing and the location advantages of the leaf nodes. However, Hybrid-CS did not consider the spacial correlation in local area. This thesis provides an enhanced Hybrid-CS compressive sensing data gathering scheme, which integrates partial nodes seclection and compressive sensing to reduce the network cost further. In addition, this thesis presents a performance evaluation model for estimating the the packet cost reduction rate and energy comsumption that our scheme reaches.Simulation result shows that with partial nodes selection, spacial correlation is well exploited, which saved the network cost and optimized the Hybrid-CS.(2) Data collision analysis and estimation of duty cycled CSMA/CA schemes in dense Wireless Body Area Networks.CSMA/CA has an excellent advantage on solving coexistence problem,because CCA is employed for channel state detection. However, high density of senser nodes make data collision to be the most challenge of competitive MAC protocol, which brings a validity and reliability damage. Therefore, analyzing the collision probability of duty cycled CSMA/CA MAC scheme is significant to guarantee the effective and reliablity of data collection in WBANs. This thesis firstly studies the types of data collision at the mircolevel, including the direct collision and staggered collision caused by hidden terminals and exposed terminals respectively. Then we discussed serval direct collisions between RTS and CTS and deeply researched one special direct collision, which can not be detected and discovered before the data frame is interruptted. Along with the increase of neighboring nodes, the probability of special direct collision is in-creased upto 5%-9%. In this thesis, series of hypothesises, propositions and conclusion are provided, which are used to explain the domination of data col-lision and the reason why direct collision is released. At last, simulation results verified our conclusions and confirmed that the special direct collision will be-come another key issume in coexistence research of WBANs. Meanwhile, the simulation under mobile scenario with time-varying traffic load shows that the influences of data collision on network performance.(3) Research on coexistence capabilities of enhanced channel hopping mech-anism in mobile Wireless Body Area Network.Channel hopping mechanism is one of the most common methods for solv-ing the coexistence problem and guaranteeing the reliability of data gathering in Wireless Body Area Networks. While, the operation of channel hopping also bring many shortcomings, such as energy consumption, delay, communication interrupt etc. So minimizing the frequency of channel conflict will provide a new avenue for coexisting WBANs. This thesis provide an enhanced channel hopping mechanism, which allows multi-WBANs to be allocated in the same channel. We present the definition of coexistence capabilities and provide the analysis construct of coexistence capability, including the risk parameter which reflects the probability of channel conflict and several metrics which measures the coexistence capabilities. Taking into account queueing and non-queueing behaviors, this thesis provided service-dependent and service-independent anal-ysis model respectively, which are used to estimate coexistence capability when only consider the number of coexisting WBANs. Secondly,to investigate the number of busy channels, this thesis provides two approximation methods based on uniform distribution assumption and additivity of Possion stream. Simula-tion results verified the availability of our theory analysis and approximation methods. When channel coexistence capacity is extended to 6, enhanced chan-nel hopping mechanism will acheve great coexistence capability improvement and support three times higher interfering-WBAN arrival rate.In summary, this thesis mainly focuses on the key problems and related techniques of load accumulation and collisions in data collection in WSNs.Abundant of energy is saved through enhanced Hybird-CS data gathering tech-nique in large-scale Wireless Sensor Networks. Data collision and channel con-flict are handled by duty cycled schemes and channel hopping mechanism re-spectively in Wireless Body Area Networks. Series of algorithmes and models are provided, numerious simulation results and theoretical analysis are also il-lustrated to verify the effectiveness of our studies. This thesis will bring an im-portant theoretical and technical support for the wide applications of the Wire-less Sensor Networks.