Research On Key Algorithms For Node Location Of Wireless Sensor Networks Based On TDOA

TDOA (Time difference of Arrival) is one of the key technology and research focus of target monitoring applications which is based on WSN (Wireless Sensor Network). In traditional sensor networks, TDOA has got many research achievements. In recent years, with the higher precision requirement, existing TDOA method become ineffective. It urgently needs us to design a series of new coverage optimization methods. Therefore, this paper puts forward a new algorithm to solve localization algorithm problem of TDOA in wireless sensor network. At last, performance analysis and simulation results are given.At first, this paper introduces the concept of wireless sensor networks and the basic theory and describes TDOA model, the fundamental of which is hyperbolic equations.Secondly, aimed at in the presence of obstacles exist in the monitored area, introducing a kind of gain modification algorithm based on kalman filtering. The algorithm rejects larger error measured value of TDOA, then and modify gain by analyzing innovations to restrict error of measurement value in the controllable range. The simulation results on MATLAB reveal that the algorithm can eliminates quickly evaluated error of TDOA in NLOS (Non Line of Sight) propagation and improve the location accuracy effectively.Once more, for the problem of iterative algorithm depended on nonlinear degree of the objective functions and accuracy of iterative initial value. This paper further presents a modified Newton iteration algorithm, the algorithm can modify Hessian matrix to make it positive definite matrix, which can be closer to the extreme value point in the direction during each search, and Newton iterative formula of the algorithm in a number of amendments to improve the convergence rate of the root. The effectiveness of the algorithm is verified by MATLAB. The simulation results show that the algorithm further improve iteration speed and precision of the iterative algorithm.At last, for the poor robustness problems of inaccurate model or state of mutations, this paper proposes a modified UKF (Unscented Kalman Filter) algorithm. The algorithm adopts hypothesis testing method to check abnormal states of system, and multiple fading factors is introduced to amend abnormal states, then modify filtering gain during iterative process to restrict all the measurement errors and iterative accumulated errors in controllable range. The performance of the algorithm is verified by MATLAB. The simulation results demonstrate that the method can positioning well, but also improve accuracy of overall positioning.