Data Gathering Method For Optimizing The Lifetime Of Unreliable Wireless Sensor Network

With the advances of the technology in Wireless Sensor Network (WSN), WSN has been widely used in more and more areas. A basic function of WSN is to collect data. Obviously, only when the whole network is in effective working condition, can the data collection be conducted. Once the network fails, it can no longer provide effective service. Therefore, optimizing network lifetime is a key issue for WSN. Links in a WSN are prone to lose packet, causing packet lost and retransmission, which costs large amount of energy and shortens the lifetime of the WSN. Besides, there exists a tradeoff between energy saving and delay satisfaction. Real-time applications of WSN, such as fire monitoring, gas or radiation leakage tracking, etc.. require that the data is collected within a delay constraint, i.e., the expended time for delivering a data packet from any sensor node to the Sink must be no more than a preset delay constraint. In a word, how to balance network lifetime and delay has become a challenge in the research area of prolonging the lifetime of the WSN within delay constraint.This work aims to propose a data gathering scheme that can simultaneously satisfy the following two goals:(1) Given delay constraint, the sink in an unreliable WSN gathers as much data as possible;(2) The scheme is energy-efficient to prolong the lifetime of the network. The main idea is to construct a data gathering tree, and prolong the lifetime of the WSN by reducing the node’s energy consumption on retransmission and balancing the energy consumption between nodes under delay constraint.In this paper, the lifetime optimization problem (LOP) is developed. A Tree-based Energy and Delay Aware Scheme (TEDAS) is proposed, which is able to maximize the lifetime of WSN while delay bound is satisfied. As TEDAS may get trapped at a local optimum rather than the global one, Genetic Algorithm (GA-based Energy and Delay Aware Scheme, GEDAS) is further proposed to improve the gathering tree. Bloom filter is applied to solve the routing loop problem, which consumes much energy making nodes depleted of energy. The contributions of this work are as follows:(1) To solve the problem of unreliable links, we apply ETX (Expected Transmission Count) to measure the energy consumption of nodes. To achieve the real-time requirement and prolonged network lifetime, we proposed a method to measure the delay of packets. According to the node residual energy, link condition and delay constraint, LOP (Lifetime Optimization Problem) is formed to maximize the lifetime of WSN.(2) TEDAS (Tree-based Energy and Delay Aware Scheme) is proposed to solve LOP. TEDAS construct lifetime maximized data gathering under delay constraint. TEDAS first construct MEST (Minimum ETX Spanning Tree) and then gradually improve the lifetime of WSN by prolonging the lifetime of bottleneck nodes repeatedly.(3) TEDAS can gradually improve the lifetime of WSN by pruning and grafting a sub-tree to a target node repeatedly. However, the main shortcoming in the TEDAS lies in that it may get trapped at a local optimum rather than the global one because it is greedy-like heuristics. Noting that GA is able to achieve the global optimum at the expense of increasing the computational effort, we present a GA-based scheme, i.e., the GEDAS (Genetic algorithm-based Energy and Delay Aware Scheme), to remedy the drawback existing in the TEDAS. To solve the LOP, solution representation, selection, crossover and mutation operation of GEDAS are specified and immune mechanism is proposed to guarantee the delay of the tree.(4) In the process of data collection tree research, loop is found to be inevitable, which is a main reason for short network life. Bloom Filter is proposed to discover loop.(5) Simulation is executed in the condition of no delay constraint and different delay constraint to verify the performance of TEDAS and GEDAS. The effect of network density and network size on lifetime is also analyzed. The results show that the proposed TEDAS outperforms some existing schemes such as MEST, RT (Random Tree), SPTS (Shortest Path Tree with Semi-matching), in terms of network lifetime and the volume of valid data. Besides, GEDAS overcomes the shortcomings of TEDAS, further improves the lifetime performance.