Research On Regional Information Estimation Using Wireless Sensor Network

With the development and progress of society,the distribution of key information(such as temperature,chemical concentration,and radiation intensity)in the target region has attracted extensive attention from researchers.However,the distribution is complex,unknown and timevarying,and it is impossible to obtain the distribution of the information in the entire region by using single sensor directly.Therefore,sensor network based information distribution estimation method arise.For the problem of the information distribution estimation in engineering application,the information distribution estimation algorithms is proposed to achieve the efficient monitoring of the target region.The main research work and contributions are as follows:(1)For the estimation problem of the dynamic information distribution in the time-varying region,a dynamic sampling model is established.Together with Surface fitting technique and the Kalman filter algorithm,a novel estimation algorithm is proposed for the time-varying information distribution.Taking the estimation errors as objective function,the position control law is designed for the mobile sensors to reduce the redundancy of sampling information,and to further improve the estimation accuracy.Simulations verify the effectiveness of proposed method.(2)For the estimation problem of information distribution in time-varying region under communication delay,a receding horizon theory based information distribution estimation method is proposed.The optimal sampling positions of the mobile sensors is obtained with the objective function optimization.Further,the stability and feasibility of the proposed method and control law are analyzed,and the upper bound values of the delays is obtained.(3)For the estimation problem of information distribution in narrow and long region(such as roads,rivers,etc.),a dynamic and static hybrid sensor network is constructed.The abovementioned information distribution estimation method is improved.On this basis,the optimal sampling positions of various sensors in the dynamic and static hybrid sensor network is calculated to further improve the efficiency and accuracy of the proposed method of the narrow and long region.Finally,the applicability of the two different methods for narrow and long region is compared and analyzed by numerical simulation.