论文部分内容阅读
针对复杂目标识别中的一些难点问题，提出一种新的综合决策网络结构。声纳目标识别的仿真实验表明，这种新的网络模型具有较快的收敛速度，更高的识别率和更强的鲁棒性，并保持了人脑可持续学习的能力。
Aiming at some difficult problems in complex target recognition, a new integrated decision-making network structure is proposed. The simulation experiments of sonar target recognition show that the new network model has faster convergence rate, higher recognition rate and stronger robustness, and maintains the human brain’s ability of sustainable learning.