A Rejection and Recognition Method Based on Chain Coverage Model in Radar HRRP Target Recognition

摘要

In this study, an algorithm for chain coverage model is proposed. The proposed algorithm is in accordance with existing out-of-database target rejection techniques in radar target recognition Using High Range Resolution Profiles (HRRP). This algorithm arranges the target samples in order based on the changes of elevation angles to construct a sausage-like coverage area for every two neighboring samples. Then the chain-like coverage model is constructed by connecting each area in proper order. This model depicts a complicated boundary structure of target samples distribution in a high-dimensional space. In this case, each kind of target constructs its own compact coverage area. The judgment with acceptance or rejection is that in terms of test samples whether or not within the target coverage area. If the test sample is accepted by several targets simultaneously, a generalized membership function is put forward to evaluate the strong or weak degree of acceptance samples subject to certain target and realize target identification. The experimental results demonstrate the effectiveness of the algorithm.