Mission Planning Of Multisensor Collaborative Surveillance In Sensor Networks

Wireless sensor networks(WSN)is composed of a large number of small energy-efficient sensor nodes,with limited storage,energy and computing ability.Due to the characteristics of good economical efficiency,flexibility and robustness,the wireless sensor networks has been widely used in areas such as target surveillance,environmental monitoring,smart home,precision agriculture and space exploration.And target surveillance is one of the most important applications of wireless sensor networks,which refers to the continuous observation and tracking to the taregts with electromagnetic radiation signals.Because of the limited energy and bandwidth of nodes,only when the sensor resources are managed dynamically and the nodes are given the optimal task planning,the performance of target tracking can be ensured,and the energy consumption can be reduced as much as possible,prolonging the network lifetime.In this paper,the target tracking sensor management problem in wireless sensor networks is deeply studied based on the analysis of domestic and international research status.The main research contents and results are as follows:1.Considering the requirement of decision-making timeliness and certain accuracy in the multi-target tracking sensor resource assignment problem,in order to meet the specific needs of each target,a novel optimization model is established,and a collaborative tracking sensor management algorithm ensuring the stable precision is proposed.This algorithm creatively fuses the covariance matrix of the actual filtering,and establishes a novel optimization model,which guarantees the tracking accuracy and makes the problem of multiple solutions more easily solved.On the basis of this,the basic binary particle swarm optimization(BPSO)is improved by combining the sensor allocation problem,and the convergence speed of the algorithm is accelerated by the population initialization method which is compatible with the constraint condition.By using a random parameter to guide the particle,two different speed updating strategies are selected to ensure the diversity of the population,which makes the algorithm more easily jump out of the local optimum.In addition,a retention mechanism to retain the allocation mechanism is been designed,avoiding the frequent switching of the sensors,making the tracking accuracy more stable.The simulation results show that compared with traditional algorithms the proposed one runs with a faster speed,avoids the local optimum and provides the stable accuracy required.So the proposed algorithm has a strong adaptability for the sensor management problem in actual combat scene.2.Considering the problems of the limited energy in wireless sensor networks and the high complexity and poor robustness of the centralized sensor management,the multi-allocation strategy based on auction is improved and an energy-efficient distributed sensor management algorithm is proposed.Based on the predicted target position,the sensor management optimization model suitable for distributed decision is established based on information gain after the observation and combined the constraint condition.The price of the sensor is adjusted by the improved bidirectional auction to achieve the market equilibrium and at last the target-sensor allocation is finished.Compared with the original algorithm,the notice of the sellers' price is reduced and the allocation of sensor resources on the targets can be completed with local information and a small amount of interaction among the sellers and buyers.The proposed algorithm introduces the Carrier Sensing Multiple Access(CSMA)mechanism,realizes the efficient self-organization of the nodes,reduces the sending and receiving of useless information,and improves the effectiveness of the energy.The simulation results show the algorithm can effectively deal with multi-target tracking problem,guarantee the tracking performance and save the energy to prolong the network lifetime.So the proposed algorithm has a strong adaptability for energy constrained WSN.3.A distributed target tracking sensor allocation algorithm based on potential game is proposed to solve the problem of target tracking for wireless sensor networks with limited observation distance,communication distance and energy in more general real scene.The algorithm takes the Geometric Dilution of Precision(GDOP)as a decision criterion,which is simple and independent of the filtering algorithm.A local information game model of sensor selection is established by using the idea of neighbors' cooperation to promote the optimization of the whole.It is proved that the model is an exact potential game model which ensures the existence of the pure strategy Nash equilibrium;And the algorithm makes full use of the relationship between the sensor communication and the observation radius in the wireless sensor networks and a parallel best response dynamics is proposed as the game learning algorithm.It is proved that the game can converge to the Nash equilibrium when the game participant(sensor)exchange the decision information with one-hop neighbors,and the convergence speed is faster than the original best response dynamics.Based on the CSMA mechanism in the communication networks,a fully distributed decision nodes selection mechanism which does not need the decision scheduler is designed,which is more in line with the self-organization characteristics of the wireless sensor networks.The simulation results show that the proposed algorithm has great advantages in convergence speed,tracking accuracy and energy efficiency.