Research On Algorithms Of Sensor Management For Target Tracking

With the rapid development of science and technology, a lot of powerful sensors emerge. The application of computers and advanced signal processing methods makes sensor management possible. At the same time, the requirements of modern war and increasing complexity of data fusion system demand higher performance of sensor management methods. As an important part of date fusion system, sensor management has attracted researchers'great attention. In this thesis, on the basis of the theory of sensor management and existing work, further research is done. The main contributions are as follows:1. Aiming at high computational complexity of Information-Theory sensor management algorithm, an approximately optimizing algorithm is introduced. Based on an analysis of communicating limitation, two forms are proposed. The shortage of existing covariance control methods is pointed out and the substitute is presented using information entropy.2. Aiming at form I of communicating limitation, a sensor management method which can track the most targets is presented. This method uses"covariance control"to determine whether target meets desired precision or not, and the loss of measurements is considered. Through simulations and analysis, it adds extra restrictions to exclude unreasonable resolutions.3. Aiming at form II of communicating limitation, a sensor management method is presented, in which the communication capability is considered. This method uses fewest sensors to maintain the desired tracking precision. During the sensor allocating processing, it uses the idea of an approximate algorithm.4. Shortages of the existing research model of sensor management are analyzed. The restriction of sensor management algorithm facing low fusion level in data fusion system is pointed out and the further research direction is presented.