Research On The Technology Of Airborne SAR And ESM Information Fusion

Since 80's of 20th century, as two important airbone sensors with gradually matrue technology, Synthetic Aperture Radar (SAR) and Electronic Support Measure (ESM) are playing more and more important role in modern battlefield, and now, airbone multi-sensor data fusion is a hot spot of research. The paper is intended to solve the data fusion problem when fusing the airbone SAR and ESM information, including the study of guiding probability of using an ESM sensor guiding a SAR, target tracking problem facing maneuvering and non-maneuvering targets and Automatical Target Recognizing (ATR) problem using SAR and ESM imformation. All studies in this paper are base on single Unmanned Aerial Vehicle (UAV) with SAR and ESM, and the notional targets are the vessels.Firstly the paper introduces the basic knowledge about the state estimation in ESM target tracking. The classic filtering method in target tracking is Kalman Filtering, the paper shows its basic theory and its algorithm process. The traditional filtering method of non-linear filtering is Extend Kalman Filtering (EKF), and the paper shows both the values and the limitations of EKF.Then an analytical expression of the successful guiding probability of using an ESM sensor guiding a SAR is derived. With the analytical expression, the successful guiding probability can be easily obtained. The guiding probability influenced by tracking route is also studied, after which the principles of the tracking route for fast guiding are given. Then better tracking algorithm, Unscented Kalman Filtering (UKF), is introduced, it gives the basic algorithm process.of UKF, and better guiding performance when using it is proven.Heading direction can be extracted from the SAR images, which could be a good maneuvering indicator. Using this information, the paper gives a new way for maneuvering and non-maneuvering target tracking, and better performances of this way are shown, especially when combining with IMM method for maneuvering target tracking, the advances are obvious.The paper solved target recognizing problems using neural newtwork group, in which subnetwork can be constructed according to the characteristic of the sensors'information, respectively. The neural newtwork based on fuzzy rules is used in SAR subnetwork, while the in ESM subnetwork, neural network based on fuzzy decision is adopted. For the outputs of the ESM, fuzzy set theory is used to achieve the transformation from the radiant point to the object type. In the end, the every output is regarded as evidence. And based on the DS evidence theory, the decision fusion is completed. Finally a numerical simulation demonstrates the efficiency of the proposed method.