1 Energy Efficient Big Data Gathering using Local data Collector in Wireless Sensor Networks T.Sujithra, R.Venkatesan Department of Computer Science and Engineering, PSG College of Technology, Coimbatore, India. Life time of the sensor and immense data gathering are challenging issues in wireless sensor network. In this paper, we propose two novel protocols called Energy Efficient Big Data Gathering using Local data Collector (EEBDG-LC) and Energy Efficient Big Data Gathering using Local data Collector with Threshold (EEBDG-LCWT). Idea of EEBDG-LC is to gather the data from the environment in the absence of cluster head in which data is collected via a node equipped with energy harvesting source called local data collector. For simplifying the routing process, local data collector is categorized based on the Euclidean distance between the local data collector and the Base Station (BS). Local data collectors which are close to BS directly transmit the data to BS. Local data collectors that are far away from BS transmit the data to BS with the help of Robotic Agent (RA). K–Means algorithm is used for positioning local data collectors in the local sensing area. As part of improvement, we propose one more protocol EEBDG-LCWT, It works based on the threshold. The main idea is to reduce the data traffic occurred inside the local sensing area and increase the lifetime of the sensor network by keeping the nodes in the idle mode based on the threshold value of the sensor node. In this approach, data gathering is carried out as in EEBDG-LC. Simulation results are taken for EEBDG-LC and EEBDG-LCWT and these results are compared with Toward Energy Efficient Big Data (TEEBD) Gathering in Densely Distributed Sensor Networks Protocol. Index Terms - local data collector, big data gathering, robotic agent, energy harvesting source, wireless sensor networks, criticality. I. INTRODUCTION Wireless sensor network is a group of senor nodes that are organized in an ad-hoc fashion. Sensor node comprises of small size memory, limited energy source, communication equipment, sensors and actuators and microprocessor for computation [1]. Amount of data and energy of the sensor node are dependent of each other. If the amount of data increases, automatically it reduces the energy level of the sensor node. Apart from data gatehring, Energy is also consumed for cluster process setup and reclustering. In case of smaller networking area, this overhead is minimal. On the other hand, it becomes a challenge when the geographical area is large. Distance is one of the factors for faster energy depletion. Hence proper placement of base station and cluster head position is very important while forming network. Scheduling the states of the transceiver plays a major role in saving energy level of the sensor node. Transceiver has three states namely idle, sleep, transmit/receive. In the idle state, it consumes only minimal amount of energy compared to active mode. In sleep mode, it completely shuts down its transceiver and there is no energy loss. In transmit / receive state, it consumes the energy needed to transmit or receive the data. By efficiently scheduling these states energy consumption can be reduced [2]. Data is routed to BS either directly or via the intermediate nodes depends on the distance from the source to BS. If the source and BS is close to each other, source directly transmits its data to BS. Otherwise, it transmits its data via intermediate nodes called multi-hop communication [3]. Our paper proposes two novel protocols namely, EEBDG-LC and EEBDG-LCWT to reduce the energy consumption occurred due to clustering process by introducing a node equipped with energy harvesting source called local data collector. It performs like cluster head. As the local data collector is static fixed at the centre position of the local sensing area and energy independent, it results in reduction of energy consumption occurred due to multi-hop routing and clustering process over head. Here K-Means algorithm [4] is used to position the local data collector in local sensing area. Threshold value is used for scheduling the states of the transceiver. It limits the nodes that are communicating with the local data collector.For simplifying the routing mechanism, one robotic agent is introduced additionally. It collects the data from the local data collector which is far away from the base station. This results in improved life time and less traffic.Rest of the work is presented as follows i) Section 2 discusses about the related work of this paper ii) Section 3 defines the problem statement and network model iii) Section 4 discuss about the proposed protocols of EEBDG- LC and EEBDG-LCWT iv) Section 5 compares the proposed approaches results with TEEBD v) Section 6 gives conclusion. II. RELATED WORK The paper Toward Energy Efficient Big Data (TEEBD) Gathering in Densely Distributed Sensor Networks proposed by Daisuke Takaishi [5] deals with big data handling in wireless sensor network. In this paper, Entire sensing area is partioned into equal sized regions. Expectation Maximization (EM) algorithm is used for cluster formation. Nodes are grouped based on the degree of depedence called responsibility. K-Means algorithm is used to find the centroid of the cluster. Data gathering is carried out at the center of the cluster by the mobile sink. Single mobile sink is used for gathering the data from all of its cluster members. Due to this, burden of the mobile sink is increased. It is not suitable for criticality based application as the delay of data gathering process is high. Travelling Salesman Problem (TSP) is used to find the optimum traversal path for the mobile sink. In TEEBD, data gathering delay, mobile sink life time and failure of the mobile sink are not taken into account. Manjeshwar and Agrawal presented a routing protocol for enhanced efficiency in wireless sensor networks (TEEN) [6]. Threshold value plays a key role in controlling data transmission. Two types of thresholds are used namely the soft and the hard threshold. Data is transferred only when sensing value reaches the hard