Research On Probabilistic Target Coverage In Wireless Sensor Networks

Wireless sensor networks(WSNs)are widely applied in various monitoring scenarios,such as forest fire warning and air quality monitoring.The coverage problem,which is basic and important in WSNs,can be subdivided into target coverage,barrier coverage and area coverage problems,according to different coverage types.Target coverage is discussed within this thesis.By placing sensors in a specific area,the target can be monitored in real time,meanwhile the data generated by the sensors is transmitted to the sink(base station).The battery capacity of ordinary sensors is very limited,and the use of redundant sensors can extend the life of the network.Energy harvesting sensors are a new type of sensor that can capture external energy and convert it into battery power.Reasonable deployment of energy harvesting sensors has the opportunity to balance the energy budget,hence can achieve permanent coverage.According to different sensor models and optimization goals,the main work of this dissertation is as follows:(1)This dissertation proposes a probabilistic directional sensor network model for the target coverage problem,and the goal is to maximize the network lifetime under target coverage situation.Unlike traditional WSNs,in a probabilistic directional sensor network,the sensor direction is changeable,and a probabilistic monitoring model is used.First,the definition of ?-directional coverage is given,based on which the problem of maximum network lifetime of ?-directional coverage,and with its NP-hardness proved later.Then,define the effective gain and overflow gain,and calculate the target weight by the potential accumulated gain.Subsequently,two heuristic algorithms are designed and analyzed:1)The energy and effective gain based greedy selection algorithm,chooses the sensor with the most residual energy from all sensors and turn on the direction with the highest gain utility,until all targets reached gain threshold.2)The target weight based priority selection algorithm determines the target weight through the potential accumulation gain,according to which chooses the sensor with the highest utility and corresponding direction.Simulation experiment results show that the priority selection algorithm based on target weights performs better than other algorithms.(2)In order to keep the energy harvesting network running forever,this dissertation first stated the problem of ?-minimal node permanent connected coverage.The key to the problem is to minimize the total number of sensors placed at the candidate point,together with the following conditions satisfied:1.All targets meet the minimum coverage probability requirement ?.2.The sensors monitoring the target should be able to communicate with the sink.3.The energy consumption per unit time is not greater than the total energy obtained in each location.In order to solve the stated problem,we first prove that it is NP-hard problem,and then give a formal definition mathematically.Further,two heuristic strategies are proposed:1)An energy-based two-stage selection algorithm,which by constructing a connected graph and weights,finds the monitoring set and iterates the path according to the shortest path algorithm.2)A multi-round path iteration algorithm,which in each round finds the path with the largest unit gain including the monitoring point,and iterates until all target gains reach the threshold.Finally,the simulation results prove that the multi-round path iteration algorithm needs fewer sensors placed.