Dynamic Deployment And Optimal Coverage Control For Mobile Sensor Networks With Complicated Condition

Sensor network plays a very important role in the military and civilian fields,has great development and application prospects.Coverage reflects the nondestructive performance of the detection and perception of the target,region and event,which is one of the key performance of sensor network,and is also the basic prerequisite of sensor network deployment.How to effectively improve the coverage of the mission domain is one of the hot issues in sensor network research.However,the existing deployment research usually requires the guidance of accurate global information and manual participation,and the deployment environment is ideal,and the adaptability,robustness and coordination of the sensor network is weak.For the above deficiencies,this dissertation combines the practical application needs,takes the coverage as an indicator,and takes the uncertainties of environment,target and their own dynamics,and the constraints of target connectivity,energy consumption,obstacles and communications into account.Through the top-level optimization planning and the bottom coverage control strategies design,under above complex conditions,the mobile sensor network is converged to the planned state,and finally the coverage is optimized.The main work and innovations can be summarized as follows:(1)A priority coverage based on distributed collaborative search is designed to meet the uncertainty of the environment where the location with coverage value is unknown.Nodes form some groups that will not be divided,and the priority coverage is achieved through the cooperative search and self adjustment strategy.The coverage is uniform and the nodes avoid colliding with each other.On this basis,in order to improve the search efficiency in the direction of motion,a cooperative search strategy and a neighborhood follower strategy based on straight line formation are designed to realize the fast search and priority coverage.(2)That the density function which reflects the coverage value of the environment is unknown is defined as the uncertainty of the environment.The awareness coverage control and the optimal coverage with connectivity constraints are studied for the continuous region and the discrete targets under this uncertainty respectively.In the problem of awareness coverage control under environment uncertainty,an adaptive awareness coverage control strategy based on virtual leader is designed for the shortcomings of the switching control used in the related literature.First,a closed loop path that traverses the mission domain is designed.Secondly,an adaptive control strategy,making the sensors reach the satisfactory awareness degree within their own coverage range is designed.The importance of each point in the environment is modeled by the unknown density function.A neural network learning algorithm is designed to learn the unknown parameters of the density function online,and finally the satisfactory awareness coverage of the mission domain will be realized.In addition,a distributed graph coverage based on uncertain information game theory to the optimal discrete objects coverage with connectivity constraints under environment uncertainty is proposed.A utility function with both environmental exploration and coverage is designed,which is proved to be a potential game.The EM method is used to estimate the density function parameters of the coverage value.When the estimation error is large,the individual explore the environment to collect the sample nodes mainly,when the estimated error is reduced,the individual is dominated by optimal coverage.The game theory is proved to converge to the Nash Equilibrium in probability.(3)That the position of target to be covered is not accurate is defined as the uncertainty of the mission domain.Aiming at the deployment for the sensor network with heterogeneous dynamics and this uncertainty,a distributed convex region aggregation strategy for mobile sensor network with heterogeneous dynamic is proposed.The mission domain is set as the intersection of individual local perception.The network includes mobile sensors with first order and second order dynamics.First,the distributed control strategy is designed,so that the individual can converge to a point in the mission domain,only depending on the local perception and the location information of the neighbors within their own communication scope.On this basis,the artificial potential field is introduced,and the sensors are deployed dispersedly in a bounded range near the mission domain.Finally,analyzing the influence of parameters on the aggregation.(4)Two optimal deployment with the energy constraints and the model uncertainty for a class of Euler-Lagrangian system that meet the actual engineering needs are studied.A dynamic optimization method with energy limitation and energy balance is designed.Firstly,the Voronoi segmentation of the mission domain is divided based on the residual energy of the sensors.The mission domain is allocated dynamically so that the energy consumption of the sensors tends to be balanced,and then the energy control strategy is designed so that the sensor energy is not lower than the specified lower bound.Aiming at the uncertainty of its own dynamic,the dynamic deployment strategy for a class of second order Euler-Lagrangian system with uncertain parameters is studied.The control strategy is designed by Backstepping method so that the system can track the desired trajectory.The stability is proved by theoretical analysis,and the only equilibrium point of the system is the centroid of the Voronoi segmentation.Finally,it is proved that the method can make the sensors converge to the optimal deployment position in the mission domain with obstacle,and the coverage is maximized.(5)In order to solve the problem of communication limitation and energy saving,a novel control strategy based on Event-trigriger is designed.Firstly,the mission domain is divided by Voronoi,and the control strategy based on Event-trigriger is designed for the first order dynamics.Unlike the existing literature,the control strategy combined with sliding mode control does not need real-time communication with neighbors to construct Voronoi division,but a design of an adaptive estimation parameters to approximate the centroid of their own Voronoi cell at the last trigger time.It is proved that if the mission domain and the sensor speed are bounded,the sensors can achieve the optimal deployment,and will not appear Zeno phenomenon.