论文部分内容阅读
由于单传感器辐射源识别的局限性,在低信噪比条件下仅提高单侦测平台的识别能力无法满足实际需求,为此提出基于协作表示Boosting的辐射源多传感器融合识别算法.利用多传感器数据信息的冗余性和互补性,对多处理支路采用时频分析提取特征,并由协作表示分类器求得残差.根据Boosting在训练阶段的权重组合得到最小分类残差,实现多传感器决策域的融合识别.仿真实验结果验证了所提出方法有效性,并且在低信噪比情况下噪声鲁棒性更优异,易于实现.
Due to the limitation of single-sensor radiation source identification, it is unable to meet the actual demand only by improving the recognition ability of single-detection platform under low signal-to-noise ratio conditions. To solve this problem, a multi-sensor fusion recognition algorithm based on cooperative Boosting is proposed. Data information redundancy and complementarity, multi-processing branch using time-frequency analysis to extract features, and collaborative representation of the classifier to obtain the residual.According to the Boosting training phase weight combination of minimum classification residuals, to achieve multi-sensor The fusion recognition of decision-making domain.The simulation results verify the effectiveness of the proposed method, and the noise robustness is more excellent and easy to implement under low SNR.