摘要
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。
To solve some knotty problems in sonar target detection measures, a new compound deciding network model is proposed. This model is applied to some complex recognition problems of sonar target and relevant algorithm is advanced. The simulation results show that the new network model has a faster speed to converge, a higher recognition rate, and robustness. It can retain the brain′s sustained learning ability and has the ability to provide judgement of reject.
出处
《控制与决策》
EI
CSCD
北大核心
2000年第1期86-88,12,共4页
Control and Decision
