The Research On Situation Assessment In Multi-sensor Information Fusion

As an important problem in modern scientific research field, study of multi-sensor information fusion technique aims to how to combine information from multiple sensors and other associated sources so that the resulting estimation and inference is, in terms of precision improvement and uncertainty reduction, better than it would be possible if any of the sources were used individually. In fact, information fusion is a new information processing method for a system which contains many the same type sensors or various type sensors. Among the whole information fusion technique, situation assessment is the key technique for high level processing of information fusion. And the essence of situation assessment is to make the situation elements picked up be detected, recognized and estimated. The research theme of the thesis does concentrate on situation assessment theories and inference models. The contents of the thesis are outlined as follows:1. The thesis first studies the function for fulfilling situation assessment. A three-level function processing model, which consists of current situation perception, current situation recognition and future situation projection, is set up and each level is analyzed in detail. Based on analyzing inference frame, the inference algorithms for situation assessment are discussed. For situation assessment is a gradually solving process that classifies real-time information on different abstract level, a multi-level and multi-hierarchical model is thought.2. Considering that the two inference methods described in the second half of the thesis are both based on the node models, the hierarchy of target classification is discussed in detail and then a method based on fuzzy equivalence relation is adopted for implementing target classification. The thesis gives a synthetic algorithm for fulfilling increased group formation by using the nearest-neighbor method and field knowledge, as the basis of setting up the inference models in chapter 4 and 5.3. The uncertain causal inference methods for situation assessment are studied. Two information fusion methods on D-S evidential theory and Beyesian network repectively for situation assessment system are put forward. In these studies, the important point is put on the evidence inference and information propagation in hierarchy hypothesis. And a new information propagation method, which base on Beyesian network, is gived In the end, the pros and cons of D-S evidential theory and Beyesian network in situation assessment field are discussed, and the author's thought about situation assessment is presented.