Maneuvering Target Tracking Algorithm Based On Combination Filter In Information Fusion

With the popularity and development of electronic information technology, information fusion which is a rising interdisciplinary subcategory is more and more important .Techniques to combine or fuse data are drawn from a diverse set of more traditional disciplines, including digital signal processing, statistical estimation, and control theory, artificial intelligence and classic numerical methods. Historically, information fusion methods were development primarily for military applications. However, in recent years, these methods have been applied to civilian applications.The target-tracking problem is an important aspect in the information fusion. The target -tracking problem has attracted lots of attentions for its applications in the military and civil affairs, and the maneuvering tracking in the military surveillance systems has especially become a hotspot to study in field. The target problem involves the state estimation of hybrid systems with continuous and discrete time variables. Tracking target in a accurate mode is the main objective in the design of target tracking system. Alpha-Beta-Gamma filters , Alpha-Beta filters as simplified Kalman filters have been widely used in tracking system because of advantages of low computation, quick response and high precision when target maneuvering .Firstly, we briefly introduce fundamental theory of Information Fusion Including the background, the basic definition, and the basic model and so on. At the end of this part, we come up with researching directions of this filed. Fusion of information from multiple sensors improves the accuracy of applications ranging from target tracking and battlefield surveillance to no civil applications. The JDL process model is intended to be very general and useful across multiple application areas.Secondly, starting with models of maneuvering targets, this chapter has discussed the following models: differential polynomial model, CV and CA models, time-correlated model, semi-Markov model, naval statistical model, highly maneuvering Jerk model and current statistical model. In this thesis, we study a theory and method of the target tracking problem, Estimation formulation, model algorithm, measurement models in different tracking coordinate systems, dynamic models, and evaluation of algorithm are discussed in this thesis.:Cartesian coordinates, hybrid coordinates and polar coordinates.Thirdly, various basic methods of tracking filtering and prediction have been discussed such as Two-point Extrapolator, Alpha-Beta-Gamma filter, and Alpha-Beta filter ,Kalman filter ;adaptive filtering problem of target tracking has been discussed, detection adaptive filtering and the interacting multiple model method. Filtering is the estimation of the current state of a dynamic system—the reason for the use of the word"filter"is that the process for obtaining the"best estimate"from noisy data amounts to"filtering out"the noise .The term filtering is thus used in the sense of eliminating an undesired signal, which, in this case, is the noise. Estimation is the process of inferring the value of a quantity of interest from indirect, inaccurate and uncertain observations. Theoretically the Kalman Filter is an estimator for what is called the linear quadratic problem, which is the problem of estimating the instantaneous state of a linear dynamic system perturbed by white noise-by using measurements linearly related to the state but corrupted by white noise. The resulting estimator is statistically optimal with respect to any quadratic function of estimation error.Fourthly, various basic methods of tracking filtering and prediction have been evaluated. Comparison and conclusion have been made; this paper has discussed these models and there advantages or shortcomings. The advantage of the Kalman filter is that makes the best utilization of the data and the knowledge of the system and the disturbances. The disadvantages, like for any optimal technique, are that it is possibly sensitive to modeling errors and might be computationally expensive. In view of this, it is very important to have a clear understanding of the assumptions under which an algorithm is optimal and how they relate to the real world. The limits of original Kalman filter and adaptive filter are investigated and the corresponding improved algorithm has been adopted in order to ensure high accuracy of sensor level tracking .I adopts fuzzy mathematics to structure using Alpha-Beta and Alpha-Beta-Gamma assembled filter. Additionally in order to demonstrate the effectiveness of schemes worked out here, computer simulations are done, which show good results of new approaches. The effectively of the sensor level tracking has been exhibited in the enhancement in tracking accuracy.Monte Carlo is a numerical technique that makes use of random numbers to solve a problem. Thus, for carrying out a Monte Carlo simulation, we require a sequence of numbers which are random, independent, real and uniformly distributed in the range .MATLAB simulating tool is used in this thesis to implement simulation of different phases in tracking procedures, so that we can take advantage of its excellent graphing capabilities and a programming interface that is very close to the mathematical equations used for different filtering and its applications The validity of sensor level tracking algorithm and the entire information fusion target tracking algorithm of whole system have been testified by the simulating results.