Research On Network Design,Routing And Localization In Smart Wireless Sensor Networks

Recent advances in technology have made the application field of smart wireless sensor networks(WSN)extend continually.Frequent growth in ubiquitous wireless connections and multimedia applications demands higher energy efficiency for WSN.Specifically,the commu-nications are in WSNs when the battery capacity of nodes in such networks is limited in the absence of energy harvesting sources.Clustered network design and energy efficient routing are the powerful tools to improve energy efficiency.Clustered WSNs are energy constrained net-works,which have a direct relationship with the network lifetime and throughput issues.Thus,there is a legitimate demand for upgrading the existing cluster designs through optimal cluster head(CH)selection to solve energy imbalance problem such as "hot spot".In addition,the fre-quent re-clustering process which results in high energy consumption due to message broadcast overhead need to be addressed.The issues above provide inspiration to design a network which ensures energy conservation by energy aware routing to minimize computational overhead in the existing protocols and ultimately,results in maximizing the overall network performance.The dissertation studies improved clustered network architecture,new energy efficient clus-tering protocol,enhanced throughput optimization algorithm,and novel three-dimensional node localization scheme.The contributions of the dissertation are as follows.(1)The existing network performance optimization schemes in WSNs need improvements to work effectively.A novel hybrid clustering network design method for large-scale WSNs is proposed.The proposed scheme uses the idea of disk partitioning to shape a large-scale network into water rippling shaped clusters and power tuning technique to design the concentric layers initiated by the base station(BS).Next,a pivotal area is defined to select the next best possible cluster head to handle re-clustering overhead.Finally,the optimal cluster size and width are de-termined based on the maximum distance between two adjacent clusters.The proposed scheme offers an energy-aware cluster head selection method,improves cluster design,and reduces com-putational overhead.Simulation results illustrate that the proposed scheme is better,in terms of energy consumption,network set-up completion time,and overall network performance.(2)Nodes longevity is a critical issue in battery operated sensor networks.A new distribut-ed energy efficient clustering protocol for both homogeneous and heterogeneous environments in WSNs is proposed.In the proposed protocol,the timer value concept is used to select the most efficient CH for every round by taking into account multiple parameters between a node and its neighbors,including residual energy,average distance of its neighbors,and node density.Specifically,the average distance of neighbors is calculated for each node,and then the timer value is estimated.Based on this timer value,each node decides itself to be the CH for a partic-ular round or not.Any node with the least timer value within a cluster is selected as CH for a particular round.Thus the CH selected using this approach is the optimal CH at that particular time.Additionally,for load management the cluster size of the CHs positioned near the BS is kept smaller in comparison with the CHs located further away.The CHs placed near the BS act as leading nodes due to data forwarding to the BS and bear an equal load that is similar to the load of CHs situated away from the BS.Simulation results indicate that the proposed pro-tocol performs better than the existing schemes with regard to longer network lifetime,network connectivity,and enhanced data delivery.(3)Randomly deployed wireless sensor nodes often generate a large amount of data in cluster-based WSNs.For the purpose of efficient data delivery,a new throughput optimization protocol that focuses on energy-aware routing and contributes throughout different network phas-es i.e.from network deployment to network communication is proposed.During the path setup phase,the next hop node is selected based on the minimum hop-count value.While in cluster formation phase,a CH announcement message is broadcasted by the selected CHs in the form of join request packet within a specific transmission range covering a maximum area.Moreover,the static power technique is used for transmission to reduce congestion and preserve transmis-sion energy.The localized behavior,minimized packet broadcast,and customized proactive routing of the proposed scheme at the node level preserves the energy of unattended deployed nodes.Simulation results demonstrate that the proposed scheme improves overall network per-formance,maximizes throughput and minimizes end-to-end delay in both low and high-density scenarios.(4)Current localization methods in WSN use effective radio signal strength indicator(RSSI)scheme for passive objects.The two main factors that impact localization are transceiver distance and deployment density.A novel three dimensional(3D)weighted centroid node localization method with RSSI ranging model strategy is proposed.The scheme tracks the position of a client for both indoor and open-air situations by utilizing a single remote sensor with negligible sub-sequent bumble.In case of 3D node localization,the proposed scheme uses four beacon-nodes and divides a pyramid-like structure into triangulation to perform three-dimensional positioning of the desired node instead of the intersection of three circles.Moreover,the path loss exponent log-normal model is used to determine RSSI.Finally,the maximum likelihood estimation of the localization is performed,and positioning errors of both the localization methods are compared.Experimental results indicate that the proposed method can achieve higher localization accuracy with fewer computations and reduction of sensor nodes.